Lack of European AI patents: detailed analysis of Europe’s technology gap and digital colonization, and why it’s too late to reverse the trend

This one has the length of a paper rather than an article.

1. The A-or-I question about AI patents: asset or indicator?
2. AI in the years ahead: will the bubble burst, will AGI be attained, or will we land in the middle?
3. Overlapping field: robotics (including self-driving cars)
4. Perfect storm: dual-gap disaster and dual-savior syndrome will have severe economic effects in 5-10 years
5. Europe’s relative poverty: lack of natural resources, large corporations, fast-growing and highly profitable businesses, investment money
6. Europoor’s colonization and brain drain
7. The most valuable IP protects network effects and lock-in
8. What prevented the creation of trillion-dollar companies with strong network effects in Europe?
9. Focus on profits, not just sales: tech has been eating conventional industries’ lunch for decades, and there is no end in sight
10. From big-eat-small to fast-eat-slow to scale-or-fail: what Europe would have had to change 20+ years ago and why it’s too late now
11. What did the U.S. and China do better? What are the UAE and Saudi Arabia doing now?
12. Could Europe have a bright future even without a world-class tech sector (and without having large tech companies)?
13. Most likely scenario: U.S. virtuous cycle and EU vicious cycle continue or accelerate
14. How you can prepare and protect yourself

1. The A-or-I question about AI patents: asset or indicator?

In the IP community we should not make the mistake of thinking that patents always equate innovation, or that only the availability of patents enables innovation. Even the role patents play in venture investment decisions is fundamentally overestimated in the digital sector, where traction and momentum are the new IP. Whether patents matter depends on the field and the patent (or patent portfolio). But regardless of causation, there is almost always a correlation between a high number of patent grants and a high level of innovative activity in the same field.

The actual value of AI patents as a means of generating licensing income and/or preserving product differentiation is untested. To the extent that any patent litigation relating to AI has happened, it was/is only indirectly related to AI, targeting hardware infrastructure⁶ (which can be used for AI, but also serves other purposes) or content authentication⁷ (which is not an AI feature, but necessitated by the availability of AI-generated deepfakes).

Patent-eligibility will have to be determined when AI-specific patents are actually litigated. There is a consensus that inventions which make large language models (LLMs) more efficient are patent-eligible. For everything at the application level, be it hardware or be it software, “do it with AI” will likely not suffice.

It is too early to tell whether AI-specific patents will become highly valuable assets. Other forms of IP (trade secrets, copyrights) and, above all, traction in the marketplace are presently, and may still be in a few years, hugely more important. But whatever the per se value of those patents will be, patent ownership statistics are consistent with other indications of Europe being far behind. That fact, in and of itself, spells doom for the have-not continent.

2. AI in the years ahead: will the bubble burst, will AGI be attained, or will we land in the middle?

Warnings against an AI bubble are not new. Recently, the chorus of experts predicting a major market correction has grown louder,⁸ and to some extent this has already happened. Investors are increasingly concerned over circular transactions. But others have identified differences between the present situation and past bubbles, such as actual capital expenditures.⁹

There is no question that Generative AI is already useful in a number of areas. Software developers are more productive. For many use cases, AI-generated graphics (even company and product logos, depending on one’s requirements) are sufficient. AI chatbots provide tech support up to a certain level. Tedious and time-consuming tasks such as generating mockups can be performed by today’s Generative AI solutions.

But Generative AI continues to make mistakes because it lacks judgment. It is, at an abstract level, a highly powerful autocomplete algorithm that can’t be trusted.¹⁰ Its use is advisable only when the one in charge is sufficiently knowledgeable to be able to identify errors. That is the way we use it at ip fray: it often saves time, but we never rely on it.

Depending on the task, a hallucination rate of 1% may be negligible (deepfakes serve their purpose unless hallucination defies the laws of nature). If you’re a lawyer and 1% of your citations are hallucinations, you risk being sanctioned, ultimately disbarred. Testing AI systems against exams (even more so when they were fed with the same or materially similar exams) delivers impressive results, but even getting 99% right is not enough when the remaining 1% could have disastrous consequences.

Thus far, AI providers are not turning a profit. They are fighting for market share, and it is unclear whether there is, for example, a potential of one or two billion people each spending $20 per month (business-to-consumer, B2C). Business-to-business (B2B) revenue opportunities should be even greater.

These are the key questions that will be answered in the years ahead and determine whether Big Tech’s capital expenditures (capex) and other investments in AI were prudent or overblown:

1. To what extent can Generative AI based on current hardware (GPU) and software (LLM) architectures be incrementally improved (better training, feedback mechanisms etc.) in order to bring the hallucination problem (which has indeed gotten lawyers into trouble)¹¹ under control?
2. When will the incremental functionality and reliability gains from larger data centers fail to offset the incremental costs? This is called diminishing returns.
3. In the event that Artificial General Intelligence (AGI), by whatever reasonable definition, is attained in the near term, will the infrastructure that is currently being built be the appropriate one for AGI and Artificial Superintelligence or will it take a hardware paradigm shift (such as to resemble more closely the way the biological brain functions)?
4. If AGI is not attained in the near term, will business and private users derive enough value from Generative AI services to allow AI providers to recuperate all (or the largest part) of their enormous investments?
5. Are there other risks such as Model Autophagy Disorder (MAD), also called model collapse, due to AI being increasingly trained on AI-generated content?¹²

The answers will come from the United States and China. Europe would have had to stop, a long time ago, being the Europe it used to be to be relevant now.

3. Overlapping field: robotics (including self-driving cars)

Full self-driving (FSD) is no longer science fiction. Its adoption is on a hockey-stick (exponential) curve:

“Rider-only” miles are those without a human supervisor. Even in a German region, Tesla has just launched an FSD-based (albeit supervised for now) public shuttle service:

In the German district of Bitburg-Pruem, we’re offering rides w/ FSD Supervised to citizens with no other means of transport, free of charge https://t.co/xlW4sz292g

— Tesla Europe & Middle East (@teslaeurope) December 26, 2025

Simple-task robots are far along and “dark factories” a reality¹³ in the U.S. and China. There could be delays before Tesla’s Optimus (not on offer to external customers) and China’s Unitree (theoretically available to early adopters, but sold out) are widely available. But it won’t take long.

With AGI, robots could do anything, including the creation of self-replicating spaceships as OpenAI CEO Sam Altman said. For now, there is only a certain overlap between AI in the sense of ChatGPT and robotics: they use different ways of modeling the world, but Transformer technologies enable them to act and not only analyze.

Structurally, the implications of data-generating AI and robotics for economy and society are very similar. For individual members of the workforce, the short-term to mid-term threat from one will be greater than from the other. Some argue that “desk jobs” are at a greater risk of replacement than manual labor, but over time the skills of robots will improve. Mass unemployment will hit drivers sooner than lawyers. As long as Generative AI is unreliable, it actually makes smarter people more productive, thereby not replacing all of them, though increased productivity also means that a smaller number of people can get a given job done. Further down the road there will be a need for extraordinarily smart humans who can supervise and direct superintelligent machines.

4. Perfect storm: dual-gap disaster and dual-savior syndrome will have severe economic effects in 5-10 years

A “perfect storm” is a particularly violent storm arising from a rare combination of adverse meteorological factors. For Europe, that one has already begun, and the consequences will be brutal.

4.1 Dual-gap disaster

Technology gap: As explained further below, there is no plausible scenario in which Europe can close the technology gap. It can’t catch up with the U.S. and China. The problem is that the value of the technology it has will not even be enough to ensure a higher standard of living than Latin America and parts of Africa will have in an AI/robotics world. As also explained further below, Europe missed the fast-eat-slow window of opportunity in tech.

Pensions gap: As economist Dr. Daniel Stelter explained,¹⁴ the actual level of indebtedness of a country is not reflected by simple cash-based accounting. A far more accurate picture would be provided if governments, just like companies, had to accrue future liabilities, such as pension payments, on balance sheets. Debt as a percentage of Gross Domestic Product (GDP) can be grossly misleading. For Germany, that number is 60% (though increasing soon), and for France and the United States, it is approximately 120%. But one just needs to visit the EC’s website to learn that pension entitlements in France amount to approximately 400% of GDP, none of which is funded, and in Germany to a similar percentage, the smallest part of which is funded:

Obviously, such an additional percentage is not the same as traditional public debt. There are no interest payments. European countries are already subsidizing their pension systems with taxpayers’ money, but that is why the pensions gap is sure to put enormous pressure on public budgets from a 2027 inflection point until at least the mid 2040s. In the 1960s, Germany had six contributors per retiree. France had four, and Germany got there in the 1970s. Right now the ratio is already below 2:1, and getting close to 1.5:1 (or less) by the 2040s.

Demographics is not voodoo or a weather forecast. It is as predictable¹⁵ as it gets. Pensions in the private sector work entirely differently in the U.S. with 401(k) retirement savings. Demographics would be gradually more favorable in the U.S. anyway.

The problem was already reliably foreseeable more than 50 years ago. It is one of the failures of the democratic systems of the affected countries that politicians first wanted to wait and see, then acted negligently (1970s), later with gross negligence (from the 1980s on) or even willfully (by deciding not to take measures that they knew were necessary but would have adversely affected their chances of reelection).¹⁶ Democracies of that kind are bound to go bankrupt¹⁷ because one generation of politicians after the other will just pass the problems on to the next (in the U.S. as well, but with far better chances of fixing the problem¹⁸), making things worse until the system collapses. That applies to conventional public debt, and to hidden debt like pensions.

The pensions gap must actually be understood as part of a wider social security gap. There are other hidden long-term obligations, such as the future costs of unselective mass immigration, which for Germany alone has been estimated by a financial expert to amount to trillions of euros.¹⁹

The two gaps (technology and pensions) are unrelated, but coincidental. Digital progress could have come earlier or later. The demographic issue is attributable not to the internet or smartphones, but to contraception and societal trends. Apart from timing, the two gaps share only one common element: political measures would have had to be taken decades ago. It was trivial to identify the pensions gap and the fix would have been obvious, but Europe’s political leaders and aides at the time were presumably not even intellectually equipped to figure out the need to get Europe ready for the Digital Era and what that would have taken.

4.2 Dual-savior syndrome

Unaware or ignorant of the magnitude of Europe’s challenges and problems, Western European societies overestimated what they could afford. They acted as if they were rich and guaranteed to remain rich forever. Two terribly costly mistakes of the “bleeding heart” kind were made:

• destructive energy policies that deprive Europe of access to cheap, reliable and abundant energy while global carbon emissions went up regardless, and
• unselective mass immigration into European social welfare systems more so than labor markets.

Neither of the two caused the dual-gap problem, but both cost factors exacerbate the overall situation, accelerate Europe’s decline, make it even harder to invest in the future, and increase the risk of social unrest and generally rising crime rates in times of economic crisis. As explained further below, countries that do not own AI must bring other assets to the table so they can use AI cost-effectively and engage in trade with those who own it. A high population density is unhelpful, especially any unproductive increase, and access to cheap, reliable and abundant energy is a key success factor for those who use AI.

4.2.1 Energy costs and availability

RH Nuttall published the following bar chart comparing industrial electricity prices by country as reported by national energy agencies:

The leading AI countries (U.S. and China) are in the green area, as is India, which also beats Europe on AI patents. Europe is in the red.

A chart published by the Financial Times shows that the actual market situation is even worse, with the UK and Germany being most insane:

AI data centers consume huge amounts of electricity, and they need it reliably, regardless of weather conditions. Former Google CEO Eric Schmidt told U.S. Congress in April 2025:²⁰

“What we need from you… is we need the energy in all forms, renewable, non-renewable, whatever. It needs to be there, and it needs to be there quickly.”

Big Tech billionaires used to sing from the “climate” hymn sheet. Not anymore. Their untold wealth ensures that they are definitely more interested in saving the planet than in making more money that would just be a number to them (no impact on standard of living). But none of them stands up now and calls on the tech industry to exercise restraint for climate reasons. There are two explanations (which are not mutually exclusive):

1. They never truly believed in a tipping point. Some may even believe that AI will be in a position to fix the problem in the future.
2. They want to win the AI race against China. They realize that virtue-signaling won’t help. For example, the EU has massively reduced its emissions, but on a global basis they increased by 30% between 2005 and 2024. The rest of the world did not follow the European “lead” at all. Ultimately, this even led to manufacturing processes being moved to countries with lower standards. There is no rational justification for this self-destructive behavior.

The policy makers who made certain decisions 20 years ago cannot be blamed for not having foreseen the electricity needs of AI. But if they had generally acted in Europe’s best interest, they would have understood that regardless of how technology develops, access to cheap and abundant energy is a key success factor. They would at least have had to strike a balance between climate and economic policy considerations. That would have made it a non-option to switch off nuclear reactors. In fact, Europe would have had to double down on nuclear energy, including the related research. Later, the economic effects of the Ukraine war should have made it politically impossible to carry out the nuclear shutdown (the net effect of which is an increase in carbon emissions) in the midst of an inflation driven largely by energy costs.

Now Europe’s misguided energy policies harm its chances in the AI era in three ways:

• Running AI services and applications in Europe is costlier than elsewhere.
• Uncertainties surrounding the availability of energy dissuade companies from setting up or growing data centers in Europe. Grid bottlenecks have resulted in “postponements” (which may effectively be cancellations) of data center projects in places like Frankfurt.
• The costs of deindustrialization further impair Europe’s ability to invest in AI even just at the user level.

4.2.2 Utilitarian vs. humanitarian immigration

Four approaches to immigration have worked for different countries at different times:

1. Old U.S. style: zero social welfare. When people came to the U.S., they had to take care of themselves from the first day. Nowadays that applies to Latin America.
2. Demand-based: allowing people to enter in order to fill jobs where they were urgently needed. This can involve sending them back home when they are jobless.
3. Criteria-based: points-based systems and similar rules.
4. Closed society (minimal immigration): Japan has other challenges, but various indicators, among them the performance of students in math tests, show the benefits of a homogeneous society. China is another great example.

What Europe has been engaging in for more than a half-century is adverse selection. The lure of the world’s most generous social welfare net combined with a lack of criteria and enforcement resulted in an economically disadvantageous mix. To be clear, there are great people from any country of origin who (or whose families) came to Europe. They are in the innovation economy (e.g., the founders of BioNTech), they are in the IP community, and they are active in many other fields. We all have such colleagues and friends. But on the bottom line, the costs are staggering. For example, approximately half of Germany’s more than five million recipients of “civic money” are foreigners, and certain groups of naturalized citizens account for a large part of the other half. On top of that, there are many immigrants who hold jobs, but are not really gainfully employed. As a result, they are also net recipients. For example, Germany has only approximately 16 million members of the active workforce who are net payers into the system. The others are either employed by the government or net recipients.

It would have been possible to invite helpful²¹ people while keeping out and throwing out the wrong ones. But European countries have prioritized what they considered to be “historic” and “humanitarian” obligations over their economic interests. Conflating refugees and immigrant workers is, however, a fatal error. Work-related immigration must be economically beneficial, which sometimes means it must be temporary. Humanitarian help must always be temporary (until a war ends) and provided where it is most cost-efficient.²²

AI should already have been a major reason for Europe to restrict immigration in 2015 and the following years.²³ It is difficult enough to determine who will be net-positive in the current situation. But with AI (including robotics) rendering more and more human workers unnecessary, unskilled and low-skilled immigrants will be even less employable. All over Europe, immigrants form the vast majority of taxi drivers, Uber drivers, and delivery service drivers. Self-driving Teslas and drones can replace them in the near term.

4.3 Other factors

The dual gap and the costs of the dual-savior syndrome account for the largest part of Europe’s problem in this era. Other major problems exist as well.

4.3.1 EU and € design flaws

The politicians who united Europe after World War II accomplished something heroic and valuable. Their basic vision of ever closer integration was smart: they wanted to start with a relatively low common denominator and envisioned that more and more responsibilities would be transferred to the European level. At some point, extricating from the common European system would no longer be feasible, or at least not without major friction. That state of affairs, they thought, might ultimately lead to a United States of Europe.

But the problem is that Europe has repeatedly put the cart before the horse because it lacked leaders like the the Founding Fathers of the United States, who understood human nature and institutional dynamics extremely well. The eurocrats

• created a free trade zone that at some point had no more mechanism to take action against Ireland’s low corporate taxes, which do not represent tax competition but are simply enabled by the fact that Ireland grants non-European companies duty-free access to a market that is roughly 100 times as large as its own,²⁴ and
• created the only financial system in the world in which the debtors, by majority vote, can set the amounts and the terms of their loans, which take different forms such as the European Central Bank (ECB) buying up government debt from the market, current accounts in the form of the Target-2 interbanking system, or all sorts of EU common debt, a contravention of the original agreement on the € for which extraordinary circumstances such as a pandemic or a war serve as a pretext.

2025 was a good year for the € exchange rate because of the U.S. government’s monetary policies, which achieved their intended effect (+4.3% GDP growth in the fourth quarter and record highs of the U.S. stock market in dollar value). But structurally the € is doomed. France has similar debt-to-GDP and deficit-to-GDP ratios as the U.S., but without a strong economy and any valuable assets. Italy and Spain have recently experienced pseudo-growth (essentially attributable to COVID funds from the EU). France and Italy each have debt of approximately ten times the level at which Greece needed a bailout. There have already been rumors of an International Monetary Fund (IMF) bailout for France (or the UK), but the IMF doesn’t have the resources this would take. France is politically unwilling and unable to enact major reforms and has no growth opportunity. One of those countries will take the € down, and Germany is no longer able to stabilize the eurozone.

The € has a “moral hazard” (which does not have an ethical connotation)²⁵ problem. It creates an incentive for irresponsible behavior beyond the incentives that exist in any democracy.

The activist attitude and lack of independence of the European Court of Justice (ECJ) is another problem. While the Justices of the U.S. Supreme Court are also nominated and confirmed by politicians, they are appointed for life. ECJ judges need a renewal after each five-year term or a new job at home. With a very few exceptions such as data retention rules, they rubberstamp political decisions, at best with a slap on the wrist.

4.3.2 Education system

Egalitarianism and unselective mass immigration have dumbed down Europe’s education system.

Not a single one of the world’s top 10 universities is based in the EU. Two UK universities and one Swiss university rank higher in a credible ranking than any EU-based university.

International comparative math tests, particularly TIMSS²⁶, show that kids in Western EU member states underperform. Their average scores do not look too bad at first sight. The problems are that those countries do not have enough top performers (for example, a performance level reached by 50% of Singaporean kids is attained by only 5% of German kids, and even fewer in Italy, Spain or France), and that their scores are going down. For France those results have been so embarrassing that the country has even abstained from participating. In terms of the percentage of top performers, Germany is at a level with Türkiye, and ten years ago France was at a level with Oman (which may by now be stronger than France).

European professors and teachers increasingly take to social media to talk about how the standards in subjects like mathematics have been lowered in recent decades.²⁷ A striking example is that in the most populous German state, final high-school exams in physics partly amount to writing an essay rather than solving specific problems by applying the laws of physics.²⁸

5. Europe’s relative poverty: lack of natural resources, large corporations, fast-growing and highly profitable businesses, investment money

5.1 Natural resource value

The following chart is from Worldostats:

If you can’t find a European country in the above chart, it’s because there isn’t any. If you subtract Norway (which is not an EU member state), even Europe as a whole would not make it into the top 10.

On a per-capita basis, Europe is in very bad shape:

Country/Region	Total Resource Value (Est.)	Per-Capita Value (Est.)	Primary Resources
Russia	$75T	$521K	Natural gas, oil, coal, timber, gold.
United States	$45T	$130K	Coal, timber, natural gas, gold, copper.
Brazil	$22T	$103K	Iron ore, timber, gold, oil, uranium.
Argentina	~$3-5T	$70K – $110K	Lithium, fertile land (soil), shale gas/oil.
EU	~$5-10T	$11K – $22K	Timber, coal, copper, natural gas.

5.2 Agriculture

If one focuses on monetary value, it looks like the EU is a huge exporter of agricultural products. But that is largely due to alcoholic beverages (such as expensive French champagne) while other major exporters sell raw materials (which the EU must import). There are many food products that Europe has to import year-round, and others that European consumers need or at least want year-round, and which Europe must import during parts of the year.

Here’s a ranking of the top net exporters of agricultural products by dollar value:

Rank	Country/Entity	Net Export Value (Trade Surplus)	Why they lead?
1	Brazil	+$140–$150 Billion	Massive surplus in soy, sugar, and beef; low import needs for basics.
2	European Union	+$60–$70 Billion*	Massive high-value exports (Wine / Cheese) outweigh heavy raw imports.
3	Argentina	+$45–$55 Billion	Very small population relative to huge grain / oilseed production.
4	Australia	+$35–$45 Billion	Huge landmass; exports ~70% of everything it grows.
5	United States	+$10–$20 Billion	Huge exports, but also a massive consumer / importer of high-end goods.

Selling wine, cheese and champagne while importing soy, rice and bananas is a good deal as long as the economy is strong. But it means that Europe must compete even for basic food products with other buyers on the world market. When the purchasing power of those competing buyers increases faster than Europe’s, and even more so when some major markets are increasingly interested in the same products (for example, Asians buying more meat and dairy products), the cost of living goes up. In the IP community, we are well off and can still afford it. But many others struggle.

Europe’s need to import raw materials is partly due to its high population density:

Rank	Country / Entity	Total Arable Land (Mln Ha)	Population (Approx.)	Hectares per Inhabitant
1	Australia	31.3	27 Million	1.16 ha
2	Canada	38.3	41 Million	0.93 ha
3	Argentina	32.8	47 Million	0.70 ha
4	United States	157.7	342 Million	0.46 ha
5	Brazil	58.3	215 Million	0.27 ha
6	European Union	98.1	448 Million	0.22 ha

The above table even understates the problem:

• With the sole exception of Canada, all other countries on the list have better climatic conditions for agriculture. They can harvest some foods more frequently, and there are raw materials that only limited parts of the EU (in the south; if any) can grow at all.
• Deforestation is undesirable, but it is a fact that certain countries could make even more land arable.
• In Latin America there is a lot of potential for increasing the efficiency of the local agriculture.

5.3 Europe’s large companies are (in relative terms) shrinking

Europe does not have a single company among the world’s 20 most valuable ones, and only one (ASML) in the top 25:

Rank	Company	Country	Market Cap (Est.)
1	NVIDIA	USA	$4.64 Trillion
2	Apple	USA	$4.05 Trillion
3	Alphabet (Google)	USA	$3.78 Trillion
4	Microsoft	USA	$3.62 Trillion
5	Amazon	USA	$2.49 Trillion
6	Meta Platforms	USA	$1.68 Trillion
7	Broadcom	USA	$1.67 Trillion
8	Tesla	USA	$1.58 Trillion
9	TSMC	Taiwan, PoC	$1.57 Trillion
10	Saudi Aramco	Saudi Arabia	$1.52 Trillion
11	Berkshire Hathaway	USA	$1.09 Trillion
12	Eli Lilly	USA	$1.02 Trillion
13	JPMorgan Chase	USA	$795 Billion
14	Walmart	USA	$890 Billion
15	Samsung Electronics	South Korea	$763 Billion
16	Tencent	China	$714 Billion
17	Visa	USA	$658 Billion
18	Oracle	USA	$579 Billion
19	Mastercard	USA	$520 Billion
20	ExxonMobil	USA	$503 Billion
21	Johnson & Johnson	USA	$500 Billion
22	Palantir	USA	$463 Billion
23	Bank of America	USA	$431 Billion
24	ASML	Netherlands	$419 Billion
25	AbbVie	USA	$407 Billion

European companies are not that valuable because they lack the right combination of profitability and growth potential. Banks may not grow fast, but if they are wildly profitable, they are valuable.²⁹ Large tech companies have both — and network effects, which will be discussed further below.

The following chart shows how Europe’s share of total global stock market value is in decline (from 30% in 2005 to 16% in 2024):

As a result of not being valuable enough, European companies are

• more likely to be acquisition targets than acquirers,
• more limited in their ability to offer stock or stock options to executives and staff,
• in a weaker position to mobilize gigantic amounts of money for investment in infrastructure such as data centers, and
• ever more inclined to list on U.S., not European, stock exchanges.

And again, the reason is that they are not profitable and fast-growing enough. That is the real issue.

The U.S. now has several companies every single one of which has a market value at a level with (or above) the entire French or German stock market.

The following chart is also interesting as it lists the top 50 investment assets in the world based on their total capitalization, and Europe plays no significant role (ranks 31, 34, and 42):

5.4 Europe’s shrinking share of global GDP

The following chart shows that Europe’s share of global GDP has been shrinking steadily since approximately 2010, while the share of the U.S. has held up pretty well despite the rise of China and India:

This shows that the U.S. economy is fit for the Digital Era, while Europe is not.

5.5 Europe fails to create new companies that scale

Combined with the problem that Europe’s large companies keep losing relevance, this means that Europe’s large companies are shrinking while its small ones aren’t growing.

Many people in Europe do not understand a fundamental economic truth:

In innovative times, the only way for an economy to have significant growth is for new companies to replace old ones. Even in the U.S., time-honored companies like IBM, General Electric, General Motors and Eastman Kodak are nowhere near as important as they used to be. That is normal because old companies generally try to milk their cash cows while new entrants want to have a disruptive effect.

In the U.S., the economy is constantly rejuvenating. The eight most valuable U.S. companies were founded after the last major European digital startup, which is SAP (founded in 1972). In the 53 years since SAP was founded, all that Europe has created is basically Spotify and Celonis.

The following charts show from different angles, but always with the same result, how far behind Europe is in the innovation economy:

5.6 Europe’s capital markets are too weak

European companies like Arm increasingly decide to go public in the U.S. or, if they have previously done so in Europe, to delist from European capital markets.

Venture investment is the lifeblood of startups. Scaling up fast is the name of the game if a company wants to convert a technological lead into lasting value. The larger a given round is, the better chances the company has to reach the next milestones and raise even more money, at an even higher valuation, next time. I have run and sold a startup, I have advised the CEO of a startup that was sold for $1B, I have helped venture investors with due diligence on a couple of occasions. I know how this works, but most people in Europe have no idea.

The following chart is a few years old and the situation has exacerbated with recent billion-dollar rounds in U.S. AI companies, but it already shows how the gap has been widening between the funds that U.S. companies raise in an A round (the first round after the seed round) versus European ones:

In 2025, the gap between U.S. and European venture capital has widened significantly, primarily driven by a “winner-takes-all” concentration in AI. While Europe has shown some resilience in sectors like HealthTech and DefenseTech, the U.S. continues to dominate in total capital deployed and the size of individual checks. The U.S. share of global venture capital reached 75% in Q1 2025, a sharp increase from previous years, largely due to massive “megadeals” in Silicon Valley. A single U.S. deal in 2025 (OpenAI’s record-breaking $40 billion round) accounted for more capital than the entire European VC ecosystem raised in all of 2024.

Metric	United States (Q1 2025)	Europe (Q1 2025)
Total Funding	~$91.5 Billion	~$12.6 – $18 Billion
Global Share	71% – 75%	11% – 16%
Dominant Sector	AI (GenAI alone = 97% of global value)	HealthTech & Biotech; later in 2025, some major DefenseTech deals

The fields in which the U.S. leads are the ones in which far larger companies are created than the ones on which Europe focuses. Given that even in an all-other-things-being equal scenario U.S. startups have higher valuations at all stages (including exit)³⁰ and better (also faster) access to funding than European ones, more and more European founders go to the U.S., either before they start or as soon as possible thereafter. As shown further below, French people living in the U.S. created hugely more market value than all French tech entrepreneurs who stayed in France.

When European startups get acquired, the buyers (at least if the deals are larger) are typically American or Asian. That is because Europe doesn’t have enough companies the size of SAP (of which Europe has one and the U.S. has many), and none in the NVIDIA-Apple-Microsoft-Google league, who could buy them.

5.7 Europe fails to create enough new-to-pipeline drugs

The following chart shows that Europe’s innovation problem is not contained to the digital sector:

6. Europoor’s colonization and brain drain

6.1 Digital colonization I: U.S. and Chinese tech revenues from Europe are in the hundreds of billions per year

When the Trump Administration complains about a trade deficit with Europe, its concerns relate only to physical goods that are formally imported into the other market. That excludes two important categories:

6.1.1 Intellectual property licenses and services

In 2023–2024, the U.S. maintained a services trade surplus of roughly $109 billion with the EU. A significant portion of this surplus is specifically driven by “Charges for the use of intellectual property.” The dominance of U.S. tech and media remains the primary driver:

Category	Typical U.S. Performance vs. Europe
Software & Cloud	The U.S. holds a near-monopoly on high-value enterprise software licensing (SaaS) and operating systems used in Europe.
Audiovisual (Movies/Music)	U.S. entertainment exports (streaming rights for Netflix, Disney+, Hollywood films) account for a multi-billion dollar annual surplus.
R&D / Industrial Patents	High in the Pharmaceutical and Semiconductor sectors. For example, the U.S. pharmaceutical industry represents about 25% of all global patent licensing revenues.
Franchise/Trademarks	Branding for U.S. chains (food, retail) contributes hundreds of millions in “royalties” from European franchisees.

Chinese companies increasingly generate licensing and services revenues from Europe as well.

6.1.2 U.S. companies’ revenue generated by EU-based (mostly Irish) subsidiaries

Revenues generated by EU-based subsidiaries of U.S. companies do not count as trade. Here are some estimated revenues of a few Irish Big Tech subsidiaries:

Company	Estimated Annual Revenue in Ireland	Notes
Google (Google Ireland Ltd)	~$75–80 Billion	Handles most of Google’s advertising revenue for the EMEA region.
Apple (AOI & Subsidiaries)	~$120+ Billion	Routes global sales (excluding Americas) through Irish-resident entities.
Microsoft (Ireland Operations)	~$60–70 Billion	Significant cloud and software licensing revenue for Europe.
Meta (Meta Platforms Ireland)	~$35–40 Billion	Primary hub for Facebook/Instagram ad revenue outside North America.

Amazon’s intra-EU revenue generation (via Luxembourg) is massive, but one must differentiate between retail and Amazon Web Services (AWS; cloud). Based on 2024 results and Q1–Q3 2025 performance, Amazon’s total revenue from the European region is estimated at $160 billion – $180 billion. Most of that is retail revenue, only a certain percentage of which constitutes a gross contribution to Amazon’s profits. But Amazon’s cloud business is very profitable. AWS remains the market leader in Europe with a 30% market share and estimated EU revenues of ~$25 billion – $30 billion.

6.2 Digital colonization II: “Buying EU companies with money earned in the EU”

It is true that U.S. tech companies use some of their profits generated in the EU to buy up European startups. While this dynamic contributes to Europe’s problem of not having enough young companies that scale, there is no national agenda or conspiracy in place. Those companies sell where they can sell, and they buy what they want to buy regardless of where they find it. The U.S. corporate tax system may help if they use low-tax Irish profits for such acquisitions as opposed to first repatriating profits, which triggers U.S. tax obligations.

6.3 Digital colonization III: labor arbitrage, extended workbench

Europeans who work in AI typically work for American rather than European companies, even if they work in Europe.

For U.S. companies, but even for Israeli startups,³¹ Europe is an interesting low-wage labor market to recruit talent. Politicians and economically illiterate people welcome the hirings made by the likes of Apple and Google in Munich, for instance. But the net effect is that the IP belongs to non-EU companies.

That is just the same as in colonial times when local workers mined gold for the colonial masters, just better-paid in absolute terms³² and with nice perks such as free healthy food.

Serving as an extended workbench for economies with larger players and stronger startup ecosystems is nice for an emerging economy. It means that income tax and social security charges are paid, the money trickles down as those people spend it, and some may learn skills that enable them later to create their own startups. For a country that is catching up, such as Vietnam, that may sound intriguing (for now, and even for countries like that, not for all perpetuity). For Europe, it’s yet another indication of the abysmal failure of its egalitarian, risk-averse society and its political system.

6.4 Brain drain

This is closely related to the phenomenon described in the previous section. Many European software developers, engineers and other high-skilled professionals hope to land a U.S. job where they earn 3-5 times more and pay lower taxes. In fact, Wal-mart truckers in the U.S. earn about as much as young doctors in Europe, and car wash and supermarket outlet managers earn $125K+ per year, more than most programmers in Europe.

An increasingly popular model is that Europeans go to Latin America or the Caribbean to perform remote work for U.S. tech companies from the same or a similar time one. This may become even more popular due to the new H-1B visa rules.

Defenders of Europe’s indefensible failure are always quick on social media to point to higher costs of living, particularly healthcare and education, in the United States. In some scenarios, that point is valid, but it is between stupid and irrational to consider that argument determinative in connection with people who have an after-tax salary of maybe $40K-$80K in Europe and take $200K-400K home in the U.S. by virtue of far higher salaries in those fields and lower tax rates. At that level, spending an extra $2K/month on healthcare and maybe a similar additional amount on food is irrelevant. The reality is that such tech workers lead a luxury life in the U.S. while they struggle to make ends meet in Europe.

The idea that healthcare is “free” is wrong. Someone ultimately has to pay, be it through taxes, other charges, or personally.

Life expectancy is pretty much the same in the U.S. at that income level. And at lower levels, obesity accounts for a significant part of the gap.

Western Europeans overestimate the strength of their public school systems (as shown in Section 4.3.2 above). The decline of public Western European schools and their inability to keep up with trends (for example, Romanian curricula are more advanced than French and German ones with respect to teaching students computer science) leads ever more Western European families, even middle-income families who can barely afford it, to enroll their children in private schools. In certain Western European cities or districts, violence by migrant children is another factor (and well-documented even in mainstream media).³³

Security arguments require a differentiated look. For example, the U.S. homicide rate is several times the European one, but large Western European cities have become statistically very unsafe compared to Eastern European ones. It remains to be the seen how different countries and cities will leverage AI to increase safety. Monaco and Dubai show that surveillance technology combined with strict law enforcement works. In the U.S., Republicans want strong law enforcement while Democrats refuse to crack down on crime. But Republicans also want to defend certain freedoms, which could limit the use of AI for security purposes.

At the moment, someone who acts carefully is not (or at least not much) less safe in the U.S. than in large Western European cities. For example, gated communities are widely available in the U.S., but not in Europe, and in the U.S., most violent crime is concentrated in certain areas, while in Western Europe knife and car attacks can happen anywhere.

When it comes to personal career choices, studies show that people make rather rational decisions. They look at such criteria as net disposable income.³⁴ It turned out (as I expected) that all that talk about researchers leaving the U.S. because of President Trump’s policies was misguided. The ones who left are historians and archeologists, maybe also gender researchers or sociologists (the latter being a field of “science” that gets a lot of attention only in Europe), but no well-known AI or quantum researcher is known to have left the U.S. in 2025. It’s one thing for a researcher or developer to disagree with the consequences of an election outcome. It’s another to take a paycut by 70% or 80% (and if you’re a researcher, a major budget cut) for that conviction. Also, researchers and developers prefer to work on exciting and famous projects.

6.5 Financial colonization: while Americans own most of their companies’ shares, Europeans don’t

Here’s one of the favorite out-of-context pie charts of pro-EU activists on social media:

Low profits, low market capitalizations, high taxes and self-destructive policies are a recipe for relative poverty: Europe is on the lunch menu. It’s being eaten alive.

Yes, Europeans own almost 50% of all foreign-owned U.S. stock. But roughly 80% of U.S. stocks are owned by Americans. That means European ownership of the total is about 10%.

By contrast, Europeans don’t own high percentages of their own publicly-traded companies’ stock. For example, Germans own only about 40% of the shares of DAX (German stock market index) companies.

When Europeans invest in America, they don’t acquire control. They do it because there aren’t enough opportunities in Europe.

6.6 Physical colonization: residential and commercial real estate; shops, hotels, restaurant chains

It is a well-known fact that foreign ownership (investment funds and private investors) of European real estate keeps increasing. To some extent that involves money laundering by foreign criminals, which worsens the European housing crisis because that category of buyers is not price-sensitive.

One just has to look at inner cities to see the increasing presence of American hotel chains, restaurant chains, and retail chains. For example, while French Decathlon, British JD Sports and the European Intersport network of independent stores are still around, Foot Locker and Lululemon are growing fast.

7. The most valuable IP protects network effects and lock-in

IP valuation is easiest when there is a licensing revenue stream in place. If a patent portfolio generates $1B per year and expiring patents are replaced with similarly valuable new ones, its value is well north of $10B. But even cellular SEP licensing revenues of approximately $20B are dwarfed by the profits of major platform companies like Apple.

As explained before, Europe’s problem is that it does not create enough companies that are highly profitable and growing fast. There is a third aspect, however: a “moat” that protects the incumbent’s position in the form of network effects or lock-in.

SEPs (other than the standards themselves) don’t have barriers to entry. Pharma doesn’t either: in the end, there’s a product that generates profits, but if a generic or other alternative becomes available, customers can switch.

Lock-in means high switching costs. Someone who has spent many hundreds of dollars on apps for a particular operating system will be dissuaded (often decisively) from switching to another mobile operating system where all those purchases have to be repeated. Such effects are gigantic when large organizations (corporations or public administrations) “standardize” on certain information technology systems. The IBM mainframe is one example.

Network effects create a barrier to entry and enable increasing returns because a product becomes more valuable with a growing number of users. The network effects protecting some Big Tech companies are so massive than even other Big Tech companies lack the resources to overcome them. When Microsoft discontinued Windows Phone development (a decision over which Microsoft CEO Satya Nadella has more recently expressed regrets), it explained that the reason they thought they couldn’t make Windows Phone succeed was the lack of apps. App developers created for the Android-iOS duopoly. Microsoft developed some apps itself, even paid some developers to make apps for Windows Phone, and engaged in “evangelism” to get more to write apps. But in vain. The other systems had countless apps, and most users would always find something, such as an app from their regional public transport service, that wasn’t available for Windows Phone.

The first major website with strong network effects was eBay: any other auction site faced the problem that buyers liked eBay because there are many sellers (therefore, a high likelihood of finding something), and sellers liked it because the presence of many buyers ensures high prices.

Tech markets with network effects are typically winner-takes-all markets.

Network effects can be one-sided (for example, LinkedIn’s core social network functionality) or two-sided (eBay: buyers/sellers; iOS: app developers/users; credit cards: vendors/customers). Android is arguably even three-sided (hardware makers, app developers, users).

To protect network effects, IP merely has to prevent clones. That means copyright may be enough.

Europe lacks IP that is related to network effects. SAP has some,³⁵ as does Spotify. That’s it for companies with a value in the 12 figures.

8. What prevented the creation of trillion-dollar companies with strong network effects in Europe?

This is not just about seeking explanations for what happened, but key to understanding a subsequent section on why Europe cannot realistically solve this problem now.

Since SAP (in 1972), Europe has not created a single new company (as opposed to joint ventures of old companies, such as ASML) that is worth at least approximately 10% of a Big Tech company. Other than SAP, there’s Spotify. That’s it. When that is the outcome after such a long period of time, the reasons are deep-rooted and can’t just be fixed with minor policy changes or exhortations to hold out.

8.1 Risk aversion

Many European startups pursue only niche opportunities. Some have no long-term strategy other than being acquired by a U.S. company. And the few who could make it big often sell out before they get there.

8.1.1 Risk aversion I: cultural

The single largest reason Europe is weak in the most valuable category of IP is that it takes enormous courage and usually significant resources to build platform businesses. The risk of failure is substantial when you need to acquire a critical mass of users, often of two different kinds where insufficient acceptance by just one of two groups will dissuade the other from using the product. The failure rate of startups aiming to build a platform is higher than that of other startups, but the potential rewards are hugely greater: the winners get increasing returns and a moat around their castle.

Risk aversion is deeply rooted. It has to do with the stigmatization of failure, particularly bankruptcy, but also the fear of openly pursuing a bold vision despite being considered crazy by others.

Jeff Bezos explained that if you make 10 investments and know that on 9 of them you will lose your money, but on 1 of them you will get a 100-fold return, you should take the deal every single time. In Europe, venture investors opt for niche opportunities and generally prefer to invest when a company has already overcome the most important risks. That means venture capital is not truly “risk capital” in Europe.

A preference for low-risk/low-return ventures is also typical of many European entrepreneurs. As a result, European startups are typically business-to-business (B2B) companies, but some of the greatest success stories are, or at least started, in business-to-consumer (B2C): Apple, Google, Amazon, Meta, Netflix, Tesla; actually even NVIDIA, whose graphics cards were mostly (not even exclusively) sold to OEMs (Original Equipment Manufacturers) but made for gamers as the end users.

The American “go big or go home” and “no risk, no gain” mentality is different from the European investment approach, which is somewhere between a bank and U.S. venture capital (and some would argue: closer to a bank). It is also more American than European to consider the second place to be the first of the losers. In winner-takes-all markets, that mentality makes all the difference.

8.1.2 Risk aversion II: tax system

The tax systems of countries like Germany contribute to the problem of risk aversion. High tax rates on very successful investments dissuade investors and fund managers from taking high risks in the beginning. For entrepreneurs, this is less of a consideration: if they have a bold idea, they want to implement it, but they need someone to fund it.

8.1.3 Risk aversion III: social safety net

Every upside has a downside. While Europe’s social safety net has undeniable benefits, it also breeds and favors low-risk bets. By contrast, Americans who want “security” for themselves and their families must get rich. They can’t count on the government.

This creates a generally more entrepreneurial climate. It also leads many startup employees to accept sub-market-level salaries in exchange for stock options (though liquidation preferences for investors often diminish the actual opportunity).

By contrast, many European employees who could found or join startups prefer to spend more time at large companies not least with a view to future pensions, where the duration of employment is a key factor.

8.1.4 Risk aversion IV: bird-in-the-hand attitude toward exits

The bolder U.S. approach leads more companies to decline offers to be acquired and instead pursue their independent growth. Companies like Google could have sold out at earlier stages, and sometimes came close, but didn’t. Promising European companies like DeepMind (which is now at the heart of Google’s AI stack) and Skype (which could have become as big as WhatsApp) sold out for a billion-dollar amount, forgoing the chance to become trillion-dollar companies themselves.

Such decisions are made by shareholders, and the relative influence of investors and founders depends on shareholder agreements and voting rights. If investors want to make a company bigger and founders wish to cash in, they can solve the problem if they want: they can buy enough of the founders’ shares for a liquidity moment that founders want. For investors, declining a takeover offer is a similar question as an initial investment. First they decide to spend a few million; maybe they spend tens of millions in subsequent rounds. Faced with an offer that guarantees a 10x or 20x return (in absolute terms, tens if not hundreds of millions), they risk losing it all (if the company goes down the tubes) when pursuing the dream of creating the next Meta or Google.

American investors think bolder and bigger. American entrepreneurs often are already wealthy when they start³⁶ and, if not, they can more easily achieve financial independence by selling some shares to investors (before or by means of an IPO) instead of selling out 100%. For lack of positive precedent, Europe-based entrepreneurs have less courage that they can build one of the world’s largest corporations in the tech industry.

8.1.5 Risk aversion V: Long-term losses

Amazon would not be what it is today if Jeff Bezos had not been able to convince enough shareholders that it was prudent to buy market share through losses. The same would most likely have been impossible in Europe. But it was the right strategy for Amazon’s retail business, and its scale later enabled Amazon to create AWS.

8.2 Ecosystem-level network effects favor the U.S. and particularly Silicon Valley

The previous section (Section 7) explained company-level (or platform-level) network effects. There are also ecosystem-level network effects, which explain why Silicon Valley accounts for about half of U.S. venture activity (by value) and a multiple of everything that Europe has to offer in that regard.

The traditionally stronger U.S. digital industry has a larger and stronger talent pool. Not only do U.S. tech companies create fewer jobs in Europe than at home, but it’s also a different focus: builders and strategists tend to be in the U.S. while foreign subsidiaries are just sales and service organizations. By contrast, when European companies set up dual headquarters, they hire some of the most important executives and engineers in the U.S. and/or their C-suite executives spend a large amount of time in the U.S. (SAP, Spotify, Celonis), far more so than the other way round.

Venture investors are also recruited from that talent pool. The prestige and financial opportunity offered in the U.S. venture industry attracts wealthy entrepreneurs and former C-level Big Tech executives. The biographies of European venture investors are typically much less impressive (many are simply finance people who never built or ran anything).

Startups find better board members in the U.S., who then raise the profile of a company, make valuable introductions to business partners, help with recruitment, and share valuable knowledge. And above all, U.S. venture funds have more money, generate more money, and the gap keeps widening.

Universities are also a factor, but in the most important fields of technology at this point, AI and quantum computing, the best researchers get hired away by Big Tech (or create their own startups).

Company-level network effects are extremely hard to overcome. But if it happens, it has so far always been accomplished by a U.S. company, with only one exception: China’s TikTok conquered Western markets in which U.S. companies were (and remain) big. Put differently, it’s unbelievably hard and risky to overcome company-level network effects, but practically impossible to vercome both company-level and ecosystem-level network effects at the same time.

8.3 Brain drain (again)

Some of the most entrepreneurial Europeans and some of the most talented European executives and engineers do not create, or are unavailable to, European startups because they found greener pastures elsewhere.

For example, French tech entrepreneurs have created far more value in the U.S. than in France itself. The contrast is particularly stark with a view to decacorns (companies worth $10B or more). France presently has one (Mistral), and a couple of large unicorns (Doctolib and Contentsquare) that are at least worth several billion dollars. But look at all the decacorns that French entrepreneurs created in the U.S.:

Company	Founder(s)	Headquarters	2025 Valuation / Market Cap
Snowflake	B. Dageville & T. Cruanes	Bozeman, MT	~$76B
Datadog	O. Pomel & A. Lê-Quôc	New York, NY	~$50B
eBay	Pierre Omidyar	San Jose, CA	~$45B
Moderna	Stéphane Bancel (CEO/Founding Team)	Cambridge, MA	~$25B – $30B

eBay founder Pierre Omidyar is a bit of an outlier here as he was born in France, but his (Iranian) family moved to the U.S. when he was only six years old. But even without him, the list shows that to make it big, French entrepreneurs have to go to America.³⁷

Similarly, the founder of Shopify (a company worth approximately $220B; Canadian, to be precise), Tobi Lütke, Sun Microsystems co-founder Andreas von Bechtolsheim and LinkedIn co-founder Konstantin Guericke are Germans. Palantir founder Alex Karp is American, but obtained a doctorate from a German university.

8.4 Fragmented market

The EU prides itself on its Single Market, yet concedes its limitations by constantly saying it needs to “perfect” it. The reality is that the EU market is neither a single market with respect to regulations (besides EU rules, numerous national ones apply) nor in commercial terms.

Multilingualism is wonderful in some ways (every single day I use at least three, typically four different languages), but it also creates a practical barrier. There are important differences between New York and California, but the U.S. is a much more homogeneous market than Europe.

Even from a consumer perspective, one can see that there are barriers. For example, there are many online retailers, but also brands that sell their own products, which are based in one EU member state but only serve that one, or a small number of others, even if they have competitive products that are not culturally specific. Languages, cultural and regulatory reasons dissuade them from taking orders from other EU member states.

There are also concerns over effective law enforcement. For example, banks in Central European countries often recommend to their customers to set a much lower limit in their online banking systems for wire transfers within the EU than within the same country. They don’t trust that some other countries will combat fraud.

With advances in machine translation, the language barrier may appear to be easier to overcome, but machine translation works better for complete sentences than for isolated terms that appear on a website or in an app.

For lack of a common language, there is no European-level media landscape allowing companies to reach all consumers. Customers often develop different preferences in each country, preventing the rise of a single European leader. Then a leading U.S. company comes and takes Europe by storm.

8.5 Cumbersome labor laws, corporate laws, other regulations

8.5.1 Labor laws

Recently, more attention has been drawn to the fact that its overly protective labor laws disadvantage Europe in the Digital Era. Europe’s restrictive Employment Protection Legislation (EPL) impairs the ability to hire and fire.

When rapid change occurs, U.S. companies can hire thousands of people in one area (for example, data centers) knowing that, if necessary, they can easily fire some or all of them again. From an employee perspective, it’s obviously preferable to be protected, but it limits the agility of companies (especially large ones) in critical situations or when opportunities arise in adjacent markets.

The three key issues are the high cost of failure, the inhibition of creative destruction, and barriers to radical innovation.

8.5.1.1 The High “Cost of Failure”

In the tech industry, where radical or “disruptive” innovation often involves high-risk ventures and high failure rates, the cost of shutting down a project or restructuring a team is a critical factor.

Research from Bocconi University (2025) found that average restructuring costs in the U.S. are equivalent to 7 months of salary per employee. In contrast, they are 31 months in Germany, 38 months in France, and 52 months in Italy. The same study found a strong negative correlation between these restructuring costs and business R&D intensity. Essentially, because it is so expensive to “fail” and let go of staff, European companies are less likely to invest in high-risk, high-reward “frontier” technologies.

8.5.1.2 The inhibition of “creative destruction”

Growth in the tech sector relies on creative destruction—the process where innovative new companies replace older, less efficient ones.

Even WTO head Pascal Lamy and European Commission research chief Marc Lemaître have explicitly blamed strict employment laws for holding back this process.³⁸ They argue that when resources (human capital) are “locked” into less productive firms due to high exit costs, they cannot flow to the new startups that drive growth.

8.5.1.3 Barriers to radical innovation, focus on incremental innovation

Labor regulations actually influence which technologies a country develops.

Because European firing costs and minimum wages are high, companies tend to focus on “labor-saving” technologies for low-skilled work (like automation in manufacturing) rather than “skill-biased” radical innovation in high-tech sectors.

Unlike the “move fast and break things” culture of Silicon Valley, European tech firms face “red tape” that leads to long delays in restructuring. This makes them slower to pivot when a new technology (like AI) suddenly shifts the market.³⁹

8.5.2 Corporate laws

Compared to other issues, this is a relatively small one, but it is symptomatic and it is related to the European Commission’s “EU Inc.” initiative (the EC has been trying to put that in place for some time, but without a breakthrough and anything that may come out of it will probably end up being of limited use, if any).

In Germany and some other countries, corporate laws are inflexible with respect to the number of shares (requiring a minimum € amount per share). By contrast, U.S. companies can just split up their stock into 10 million or even 100 million shares, allowing fine-grained allotments to employees (typically in the form of stock options) that still look attractive.

Another problem in Germany and some other countries is that every little thing, even relocating within the same street, requires notarization. Moreover, investment contracts tend to be long and come with voluminous exhibits, all of which must be read out aloud. As a result, investors and founders often spend 12-hour days at the offices of German notaries public. It is an antiquated, parasitic⁴⁰ approach that slows down transactions and distracts from more important tasks.

8.5.3 Other regulations

The EU’s digital overregulation adversely affects European companies to a greater extent than their larger U.S. rivals.

8.6 Wealth tax on paper money, capital gains tax on venture investments

8.6.1 Wealth tax on paper money

Startups have high paper valuations, but that does not mean any actual liquidity for their founders. Some European countries nevertheless impose wealth tax on them even prior to any liquidity event. Germany had such a tax in place and will presumably have one again soon (as the country is in rapid decline).

Whether graphics card maker SPEA could have become another NVIDIA in a different environment is another question, but its founder, Uli Seng, was probably the best entrepreneur Germany has had in the computer hardware field since Heinz Nixdorf. His company raised funds from the German subsidiary of U.S. venture fund Atlas. As result of venture funding, his shares had a nominal value on the order of ten or twenty million euros. He then had to pay a 1% wealth tax on that paper value. In order to do be able to do that, he had to sell 2% of his shares every year to the investors, paying a 50% income tax and using the remainder just to pay the wealth tax. I once met him (his company was based in Starnberg, where I lived for 20 years) and I talked to his investors years later. One day I read an interview in which he explained that he left Germany to build hotels in Croatia because he did not want to deal with the German startup environment anymore.

8.6.2 Capital gains tax on venture investments

In connection with risk aversion, this topic came up before. Institutional and private (business angel) venture investors are treated more favorably under U.S. tax laws.

8.7 Conclusion (and mathematical example): too hard to scale in Europe

All of those factors combined (and I do not even claim that the list is exhaustive) result in European startups being, on average, less ambitious and having less potential from the get-go, but even if there is the ambition and the potential, the problem is that they typically fail to scale.

Here’s a simplified example. Let’s assume that there are two companies in the same segment, and all other things are equal, but one is in the U.S. and one in Europe. As a result of faster, larger rounds of funding and a larger, more homogeneous market in which it is more efficient to acquire more customers, the U.S. company grows at a rate of 80% a year and the European one at 30%. After three years, will have grown by a factor of 5.8, and the European company by a factor of 2.2. In a winner-takes-all market, it is obvious who will have more money for R&D and marketing, and who will acquire (or simply displace) whom.

European companies can also try to address the U.S. market early on, benefiting from lower R&D costs at home while focusing on where there is the largest market opportunity. But for a long time, the perspective of European investors was that they should first focus on their domestic market and then go to the U.S. “from a position of strength” (just that there usually wasn’t any opportunity left for them in the U.S. market by then). In the 1990s and 2000s, many European startups were just copycats of U.S. startups, hoping to be acquired by the first movers, which worked for some while others simply got displaced.⁴¹

9. Focus on profits, not just sales: tech has been eating conventional industries’ lunch for decades, and there is no end in sight

Here are some ballpark figures to understand why tech is so key.

9.1 Sales growth

• In the early phase of computerization, the tech industry, which was initially hardware-centric, had huge growth rates. There were years in which it doubled in size. But in absolute numbers, it was almost a niche industry.
• Over the last couple of decades, the tech industry has grown about three times as fast every year (with only a few outlier years where it was not the case) as the overall economy:

Metric	Growth Rate (Approx.)
Global/OECD GDP	~2% – 2.5%
Tech/ICT Sector	~6% – 7.5%
Growth Multiple	~3x

• When you have that difference basically every single year, the exponential effects over a decade or two are substantial.
• That discrepancy in growth rates also means that the tech industry’s share of the overall economy has been increasing. But if a sector that accounts for 5%-10% of the overall economy grows at a rate of 7% while the other 90-95% of the overall economy grows at a rate of 2%, the growth of tech does not come at the expense of the overall⁴² non-digital economy’s sales.
• Presently the disrepancy is even wider due to massive AI investment. Some even say that the U.S. would presently be in a recession if not for AI-related capital expenditures.

Presently, the tech sector accounts for (depending on definition) 6-10% of the entire economy based on sales.

9.2 Profits: that’s where the difference is REALLY huge

What often (sometimes intentionally, sometimes not) causes confusion in debates over the economic importance of industries or in trade contexts is that sales are talked about more frequently than gross margins and profits. Legal professionals actually know the difference in their own business: they want to maximize their per-partner profit.

The tech industry is hugely more profitable because its marginal costs are often next to zero. These are the 2024 profits of the Magnificent Seven:

Company	2024 Net Income (Approx.)	Sector Classification
Apple	$93.7B	Information Technology
Microsoft	$88.1B	Information Technology
Alphabet (Google)	$87.3B	Communication Services
Meta (Facebook)	$52.3B	Communication Services
Amazon	$37.7B	Consumer Discretionary
NVIDIA	$60+B	Information Technology
Tesla	$15B	Consumer Discretionary
Total Magnificent 7	~$434 B	—

If one adds other publicly-traded U.S. tech companies outside the Magnificent 7, that’s a total annual profit volume of approximately $800B. That’s approximately 40% of the entire profits of the S&P 500, and well over 70% of the entire profits of the European STOXX 600.

Tech profitability is not the only reason, but a major part of the reason why U.S. corporate profits have grown more than twice as fast (more than quadrupled) since 2009 as STOXX 600 profits (which only doubled).

To some extent, tech platforms directly suck oxygen out of the conventional economy and increase their own profits that way. Delivery services are an example. Restaurants, a low-margin type of business, are forced to pay high commission rates. In some countries, taxi drivers offer their services via Uber, which forces them to work for less and split the proceeds with the platform. Amazon and other online retailers are a large part of the reason why so many traditional inner-city stores have closed down.

9.3 Market capitalization

The combination of vastly higher profitability, fast growth and network and lock-in effects (“moat around the castle”) explains why, as shown further above, the European share of global market capitalization went down from 30% in 2005 to 16% in 2024 (and why Europe has no more company in the global top 20, and only 3 of the world’s top 50 assets are European).

Among the global top 100 companies, Europe’s share (based on value) has gone down from 40% to 10% in less than 25 years:

What some people (also in IP) don’t understand is that market caps are a highly important indicator as well as an asset. The fact that there is some volatility does not render market cap ratios irrelevant. Volatility just means that one can’t infer anything from short-term (such as intra-day) fluctuations.

Market cap is a currency for mergers and acquisitions (M&A). Large tech companies can buy large non-tech companies,⁴³ but not the other way round. Market cap is also a currency for human resources: stock awards and stock options (for management, engineers, other staff) and puts companies at the top of the M&A food chain. For example, for all those NVIDIA employees, the (roughly) $5T market cap is very real because it means most of them now own tens of millions of dollars. Market cap is what the stock market expects future profits to be. Market cap is what venture investors have in mind when deciding on investments with a view to a future initial public offering (IPO). And so forth.

9.4 Comparison to other industries

There is nothing like tech in terms of the combination of margins, volume, growth rates, and moats around the castle. Pharma has high margins compared to many other industries, but low ones compared to tech, and unless you’re BioNTech or Moderna in the midst of a pandemic, the growth potential is moderate. Luxury goods (LVMH etc.) have high margins (still nowhere near as high as tech) and the growth potential is limited (Veblen goods lose appeal when too many customers own them).

Tech has been absorbing other industries and industry segments. Think of all the functions a smartphone performs. Look at how the fastest-growing players in the automotive industry are effectively tech startups, with software and other new technologies playing an ever more important role.

10. From big-eat-small to fast-eat-slow to scale-or-fail: what Europe would have had to change 20+ years ago and why it’s too late now

10.1 The fast-eat-slow window of opportunity was open for well over a decade; now it’s closed

For a long time, Europe didn’t view it as a fundamental threat that the U.S. was the primary cradle of new hardware and software companies. Back in the 1970s or 1980s, there was more of a concern over the rise of Japanese tech companies (Sony Panasonic, Toshiba, NEC, and others).

Tech was understood to be the future, but politicians and the general public were not interested in closing the gap (just like they ignored the changing demographics as the pensions gap still seemed to be far away from actually impacting public finances).

The market capitalizations of U.S. tech companies were not frightening at that time. For example, Microsoft went public in 1986, and at the closing of its first day of trading was worth approximately $800M. Four years later, its value was about ten times as high ($6-8B), but at the time, European giants like Royal Dutch Shell ($45B-$50B), BP ($25B-$30B) and Glaxo ($20B-$25B) were worth a lot more. Even Daimler (Mercedes) was well over $10B, and BMW’s market cap was not far below Microsoft’s.

None of those companies could have acquired Microsoft (a) due to a different industry focus and (b) because its key shareholders like Bill Gates wouldn’t have sold. But generally, the world of business in the 1970s and 1980s still worked like the Old Economy: the big eat the small. That was the conventional food chain.

The 1990s were transformative, mostly because of the web, but industry dynamics already changed in the first half of that decade (for example, with computers becoming multimedia devices). I remember a conversation with an Intel executive who said that the industry was now so dynamic that the big were no longer eating the small, but the fast were eating the slow, which was consistent with what I heard from other industry players during that period. What the metaphor meant was that fast execution (of winning strategies, of course) enabled new entrants to win. In Intel’s core business, that was actually not the case: “Wintel” (Windows + Intel) was too strong. But in many other segments of the tech industry, it applied, and above all, during the dotcom era, which gave rise to companies like Amazon. Google was founded toward the end of the first major dotcom wave, and became really strong only in the early 2000s. Facebook (now Meta) was founded in 2004.

During the fast-eat-slow era, Europe lacked the key success factors. But with a more ambitious investment mentality, more funding and a market environment enabling companies to scale faster, that was the time when Europe could have created a few companies like Amazon. “Coulda woulda shoulda didn’t.” It would have been the time to create a tech ecosystem that would have positioned Europe for the next waves. But Europe missed them all: after PC hardware, PC software, video games, e-commerce, social networks, crypto, smartphones etc., Europe has now missed out on AI and has practically already been marginalized in quantum computing (as also evidenced by its weak patent position).

The current era, where AI and quantum computing are the two key topics, is a scale-or-fail world where a company like OpenAI can raise a first round in the tens of billions of dollars and even seed rounds sometimes involve billion-dollar amounts. Now it’s too late. The amounts are beyond what Europe can accomplish. It’s now standing on the sidelines. It has a few decacorns, but other than SAP (founded in 1972) and music streaming platform Spotify, it has nothing that is worth hundreds of billions of dollars while America has multiple companies worth trillions of dollars each.

To play the game, Europe would have had to build a vibrant ecosystem in the 1990s and (at the latest) in the 2000s.

10.2 Subsidy-like loans as leverage for venture capital and business angel investments

The lack of available venture capital was and still is a major bottleneck.

From 1989 to 2003, Germany had a program named tbg⁴⁴ in place that made a temporary impact. It funded (in the form of loans for which no security had to be provided and which matched private-sector investments, combined with what was an insurance covering 50% of potential losses) various companies in SAP’s ecosystem, Mobilcom (which became a major mobile virtual network operator), Intershop (which was like an early Shopify), some laser tech companies, and biotech companies such as Qiagen and Morphosys. Indirectly, that program even benefited BioNTech’s founders before they created BioNTech itself. But instead of doubling down on it, a social democratic and Green coalition government with more of a focus on traditional industries shut it down.⁴⁵

Nowadays, European governments act as high-volume limited partners (LPs) of venture funds that invest in Europe. While this increases the amount of available capital, it is also criticized by some private-sector investors as a counterproductive market distortion.

10.3 Investor mentality: go big or go home

The most important change could not have been achieved by governments. It would have required European society and particularly the investment community to adopt a more American attitude (“go big or go home”) where moonshots (enormously ambitious projects) would not be categorically ridiculed.

10.4 Market size

A strategy would have had to be developed to have European companies compete internationally at the earliest opportunity. Governmental support of certaint ypes could have run afoul of state aid rules in the relevant target markets. The EU, however, would have had to create a large domestic market addressable by startups as a single market.

That involves a problem as hard to solve as the investor mentality problem discussed in the previous section: multilingualism. While it is a feature of Europe that many people like, it has been one major reason for which European tech startups have been unable to scale fast enough to become global leaders.

10.5 Attracting global talent

Here, again, language is a barrier. The return on investment (ROI) for learning a language like German or French is much less favorable (greater input, lower yield) than for English. And especially tech people tend to have at least some basic knowledge of English before they emigrate.

Since at least the 2010s, it has become more common for European tech startups to use English in their internal communications so they can hire developers from anywhere. But in everyday life, the language barrier still exists.

By adopting English as a Europe-wide official language down to the smallest village (in addition to existing national languages), Europe could have become a more attractive destination for foreign entrepreneurs and engineers.

It was never realistic to believe that one could attract many of the best migrants with all of those disadvantages, plus low salaries and high tax rates.

10.6 Tax, corporate and labor laws

Europe would also have had to cut red tape, lower taxes, and enable some degree of “hire and fire” in tech, particularly for employees who can easily find a new job.

11. What did the U.S. and China do better? What are the UAE and Saudi Arabia doing now?

11.1 U.S.-China-Europe comparison

When economists, politicians and others talk about Europe’s decline, most of them focus on conventional issues, such as industrial and labor policies or excessive bureaucracy, which are also real and massive (and deserve to be addressed), but which are still less important than the tech gap. Another frequently cited factor which is not specific to tech is that the U.S. better recovered from the 2008 financial crisis because of the greenback’s role as the world reserve currency.

But when future economists pass judgment on Europe’s descent into the Second World, the crucial role of tech will probably be better understood with the benefit of 20/20 hindsight. They will find plenty of evidence that there was a limited window of opportunity, a period during which each major economy had to be in the best position it possibly could be in to seize the AI opportunity. The U.S. and China made it work. Europe did not.

Counterintuitively, the U.S. and China made it work within systems that one would consider to be diametrically opposed. And Europe? There’s a quote often attributed to Margaret Thatcher: “If you sit in the middle of the road, you get run over from both sides.” Europe is definitely getting squeezed between the two superpowers that got the digital economy, which now increasingly means AI, right. So why did Europe’s “middle way” not work out as well as the U.S. and Chinese alternatives?

Here’s my comparison of the three major economies:

United States	European Union (incl. UK)46	China
Western democracy	Western democracy	Single-party people’s republic
Long tradition of innovation and courageous tech entrepreneurship	Old school: innovation in long cycles and highly conservative entrepreneurship	Went from manufacturing to engineering in a few decades; enabled entrepreneurship
Hands-off economic and regulatory policies	Economic freedoms restricted, bureaucratic burden, strong labor unions	Heavy-handed governmental control over the economy, yet companies move fast
One language	20-30 languages47	One language
Open and homogeneous market enables rapid expansion	Open but hetereogeneous market with significant resource requirements for expansion48	Strategically regulated and homogeneous market enabled local companies to take the lead
Domestic talent pool plus magnet for world’s best entrepreneurs, execs, and engineers	Domestic talent largely in old economy, limited attractiveness to world’s best migrants49	Huge domestic pool; closed society
Flexible workforce	Inflexible labor market due to employment protection laws	Flexible workforce
Meritocratic	Egalitarian	Communism, yet meritocratic in tech
Strong defense of national interest; transactional attitude to international politics	Neglect of national interest; savior syndrome	Strong defense of national interest; transactional attitude to international politics
Proud of its history, but relentless focus on the future	Backwards-oriented, focus on its own history	Proud of its history, but relentless focus on the future
Cheap, abundant, reliable energy, whether renewable or non-renewable	Expensive, too limited and unreliable energy	Cheap, abundant, reliable energy, whether renewable or non-renewable50

All in all, it appears that the U.S. and Chinese roads to digital success were disparate in some ways, but shared some common elements that Europe was lacking, such as the existence of a large, homogeneous market.

Europe’s digital failure will go down as a tragedy. A continent that needed to succeed in tech fell far behind because it just didn’t have some of the key ingredients of success in the fast-moving digital era that rewards those who take high risks and make it work.

11.2 Kingdom of Saudi Arabia and United Arab Emirates leverage resources to maximize benefits from AI

Saudi Arabia and the UAE obviously have two major advantages over Europe: national wealth from natural resources, and access to vast amounts of inexpensive and reliable energy.⁵¹ They have strategically made long-term investments through sovereign wealth funds, built world-class infrastructure, managed to attract global talent (particularly Dubai), and they have recently made bold investments in AI. They also aim for intelligent regulation as opposed to EU-style excessive and misguided regulations.

12. Could Europe have a bright future even without a world-class tech sector (and without having large tech companies)?

12.1 In relative terms, no way: the gap between the U.S. and Europe will continue to widen

Whether Europe can still do well in absolute terms will be addressed further below. That depends on different scenarios, but there is no realistic hope for Europe’s public finances and colonization will continue.

In relative terms, there is no plausible scenario in which the economic gap between the U.S. would not widen, probably every single year, otherwise over the course of any given three-year or five-year period. The U.S. has vastly greater natural resources, a stronger agriculture, and the greatest financial rewards are in tech. None of the industries in which Europe might hope to succeed, and not even all of them combined, comes close to the profit and growth potential of tech.

Therefore, the U.S. and China own the most valuable IP. Europe will have to pay for using it, and some of that money will be spent on acquiring more and more of Europe.

12.2 Could AI and quantum already be the end game?

The late 19th century quote according to which “everything that can be invented has been invented” may have been misunderstood and/or misattributed.⁵² One has to offer such a prediction with caution.

It is, however, possible that AI and quantum computing are the last major software and hardware frontiers. That is different from, for example, smartphones, which were considered “the future” but not forever.

Should Artificial Superintelligence be created in the next decade or two, then AI systems will make most if not all future inventions.

12.3 Most optimistic scenario for Europe: AGI-driven abundance

Artificial General Intelligence, which would then lead to Artificial Superintelligence, has the potential “to make everyone rich” as Elon Musk describes it. This would require a market in which AI is for the taking. If those who own it get to charge substantial amounts, Europe has a problem. Also, even if this vision materializes, Europe could get into serious trouble during the transitional period.

AGI-driven abundance would make many goods cheaper, but finite resources, which also includes real estate, would potentially become even more expensive.

If, on balance, AGI has a huge deflationary effect, tax revenues would decrease while debt would have the same same nominal value and in many jurisdictions it would be difficult to lower pensions. Therefore, only some of a government’s expenses would be reduced (such as bhuilding new infrastructure). As a result, governments would struggle to repay their debt unless central banks step in to buy up and clear all past debt. Under normal circumstances, that would destroy the markets’ faith in a currency. In a once-in-history situation where AGI changes everything, it might work. Negative interest rates would be another vehicle.

Artificial Superintelligence would even enable humanity to harvest the Asteroid Belt, the Moon, and other celestial bodies. When that point is reached, even many natural resources will become abundant.

12.4 An AI bubble burst wouldn’t solve any of Europe’s problems

The most pessimistic scenario for AI is that there is a bubble which will burst because the markets will lose faith in the idea of building huge data centers and that the current massive spend will lead to AGI in the not too distant future.

This could happen, but it would ultimately just be a market correction. Even a major market correction would not solve a single one of Europe’s problems:

• All Big Tech companies have profitable core businesses. They would take some write-downs, their investors would accept some losses, and life would go on.
• None of that creates a market opportunity for European tech companies.
• In a hypothetical scenario where NVIDIA, Google and Microsoft lost half of their current value and the German DAX40 remained stable (which is not even likely), those three companies combined would no longer have an aggregate market cap of four times that of the DAX40, but still twice that of the DAX40. In other words, Europe would still be Europoor, even if the gap narrowed (temporarily anyway).

12.5 Could Europe’s startup ecosystem create significant wealth?

Even if Europe’s current startup ecosystem is dwarfed by that of the U.S., it is a legitimate question whether policy makers should still try to support it and at least grow it to the extent that they can.

It is not entirely impossible, but statistically not very likely after more than a half-century, that Europe can create another company the size of SAP. If it happened, it would be positive and could encourage other entrepreneurs and investors. It wouldn’t result in a European Silicon Valley equivalent, but it could be helpful. There would be a high risk of such a company being acquired before it reaches massive scale

Whether any particular measures to foster the European startup ecosystem are likely to yield a good ROI would have to be discussed based on particular proposals. Some fundamental questions would have to be asked. One of them is how to prevent that promising companies are acquired by larger foreign players.

An economist whose claims of a “Super Europe” are beyond ridiculous argues (among other things) that it is a good model for Europe to build startups, sell them to U.S. acquirers, and then buy U.S. stock with that money. But if, for example, SAP had sold out to a U.S. rival when it was worth $3B (as opposed to its current $300B), 99% of the value would have been created by a U.S. owner.

12.6 Digital public infrastructure (Eurostack)

The EuroStack movement is a strategic industrial and political initiative aimed at ending Europe’s status as a “digital colony” of the United States. Its primary goal is to build a complete, sovereign European technology stack—spanning from chips and cloud infrastructure to AI and software—that is owned, operated, and governed within the EU.

The movement is driven by a coalition of technologists, economists, and it has support from companies like Airbus, Ionos, and Orange. Their strategy centers on three main pillars: Buy European, Fund European, and Build European. It is unclear how this would (and, in fact, could) work.

12.7 Applying AI

Some economists (including some who are notorious propagandists supporting the EU and certain national governments) argue that Europe doesn’t need to own foundational AI technologies as long as it can derive substantial value from their application. At first sight, this seems logical. But restaurants and taxi drivers who see their margins reduced by online platforms are also applying new technology, but the price is high.

As discussed further below in Section 13, European governments will lack tax revenues from tech that they would need to handle tech-driven (even if only temporary) mass unemployment.

12.8 Medium-sized companies (“Mittelstand”)

Some suggest that it is a strength of Europe, or at least of its largest member state, that it has a less concentrated economy. There are “hidden champions” (world market leaders in certain niches) among the larger one of its medium-sized companies (“Mittelstand”).

That may have been true in the old economy. In the AI era, scale is more important than ever. Also, more and more of those companies are facing fierce competition from China. And a number of them are dependent on Europe’s automotive industry.

Medium-sized companies are also limited in their ability to make acquisitions.

12.9 Is heavier tech regulation a solution?

European government agencies are ill-equipped to make decisions in technology markets. They can’t micromanage the tech sector.

It is important to have rules, and limiting gatekeeper power in mobile ecosystems would definitely help (not only, but also) European companies offering apps on those platforms. But the EU’s ability to impose its rules on Big Tech is not unfettered. It can do certain things, such as the USB-C charger requirement,⁵³ but if it goes too far, it risks blowback from the U.S. government.

In the U.S., it is already viewed very unfavorably that the EU has imposed fines on U.S. tech companies amounting to €3.8B in 2024 while its own publicly-traded tech and internet companies paid only €3.2B in income tax during that period:

12.10 Historical references and philosophical theories

In conversations with Europeans, particularly on social media, it is noticeable that arguments about Europe’s economic problems are often countered with historical references and philosophical theories. Sometimes such arguments are fielded for “cope” or denial. In a few situations, people initially acknowledge that Europe is strategically lost, but have some remaining hope based on highly unspecific ideas.

The most common reaction is to point to Europe’s social safety net, including healthcare, and (allegedly) less inequality. The problem with those arguments is three-fold:

• Even if all of that was correct, Europe still has to compete with other economies that, to put it bluntly, just don’t care: they simply compete. In other words, Europe may have been too good for its own good.
• Europe’s waning economic strength will make it harder and harder to afford the current system.
• Europe’s model was apparently the wrong one for an era during which the most important challenge was/is digital innovation.

The same three points apply to virtue-signaling arguments such as “at least Europe did the right thing, also on climate change, trying to convince the others to do the same.” If the whole world had been prepared to adopt the European way, maybe it would have been better. But there never was a chance that the U.S. or China would do so.

What no one can present is a credible theory according to which U.S. society (with its alleged inequality and other issues) would collapse to the extent that the U.S. would lose its technological edge (including its IP).

The vaguest arguments are the following ones:

• “This is a highly complex matter and one can’t predict the future.” Of course, many things could happen. Extraterrestrials could land on Earth. But when there is overhwelming evidence that the European model isn’t working in the Digital Era, it is perfectly reasonable to draw conclusions from that observation, particularly after researching numerous data points.
• “Reports of someone’s death have often been greatly exaggerated.” It is true that doom sometimes doesn’t occur when predicted, but that does not explain how Europe can compete.
• “Everything comes in cycles and Europe can come back.” This, however, is not a strategy for overcoming company- and ecosystem-level network effects.
• “Europe has previously risen from the ashes.” Or: “Germany was rebuilt after losing two world wars.” The current situation is fundamentally different. What Europe (or, specifically, German) rebuilt was old, slow-moving industries; there were differentiated (“Made in Germany”) product markets, not today’s winner-takes-all markets. The times were different, and it was not as easy as now for talented people to go elsewhere to found, or work for, companies. People tend to be rather rational about their individual choices. There isn’t a sufficient pool of people whose European patriotism would prevent them from doing what’s best for them.
• “Europe has 500 million people. There wil be enough smart people among them to make Europe succeed again.” That argument is wrong for the reasons just discussed (most people will just choose what’s best for them) as well as because other countries like China and India now have huge numbers of people educated at a level comparable to, if not above, the European one.
• “We can’t just give up. We must do something.” That is an honorable position to take, but doesn’t solve any of the problems. If there aren’t enough people with that attitude, it won’t help. It may be better to take care of oneself (Section 14).

Some of that amounts to “hope springs eternal.” But that’s not the answer. Also, some believe one shouldn’t talk about Europe’s problems because it’s not in the European interest. But the situation doesn’t get better if we remain silent, and everyone must prepare for a future for which Europe as a whole is ill-equipped.

13. Most likely scenario: U.S. virtuous cycle and EU vicious cycle continue or accelerate

Unfortunately for Europe, the range in which the by far most likely scenarios fall will create an uncomfortable situation in the years ahead and, even more so, in the 2030s. The most likely development is that, with or without a temporary market correction,

• the tech sector continues to outgrow the overall economy;
• Europe’s tech deficit continues to grow accordingly, which means increasing colonization;
• Europe’s high energy costs and limited quantities of reliably available energy will adversely affect its competitiveness (this is a Europe-wide issue with respect to using AI, and in some parts, especially Germany, there is deindustrialization);
• structural unemployment due to technological progress increases; and
• the public finances of (and some smaller) Western European countries deteriorate rapidly, due to a toxic mix of economic weakness, the pensions gap, and the costs of unselective mass immigration (plus the eurogroup moral hazard problem mentioned before, which lets countries like France get away without much-needed reforms). Increased military expenditures come on top.

The net effect is that there will be at least steady, and possibly even accelerating, decline. At a point where lenders lose faith in just one major eurogroup country, the eurogroup will have to let the € fall apart or take on greater debt, be it in the form of eurobonds or quantitative easing by the ECB, with inflationary effects. The debt of a country like France (€3 trillion; 10 times the amount that Greece owed during the last major € sovereign debt crisis) is too much for a rescue package, especially now that Germany is in bad shape.

Over the course of the next few years, that is the greatest risk, combined with potential mass layoffs particularly in the automotive industry, partly because of its reduce ability to compete (in emerging markets, Chinese car makers are already selling huge numbers of combustion-engine vehicles), partly because electric vehicles have fewer parts. By the 2030s, public finances of countries like Germany will be in terrible shape.

One problem facing European governments in the near to mid term is that they will have to deal with mass unemployment but (unlike the U.S.) without having substantial revenues from tech to redistribute.

Some argue that no matter how many jobs AI and robotics kill, new jobs will be created to replace them. But even tech leaders don’t all agree (for example, legendary entrepreneur and venture investor Vinod Khosla). The fewest would claim that enough new jobs will be there immediately. And unlike after past technological revolutions, it is fairly possible (in my view, almost certain) that this time around many new jobs will have to be created by governments. For many business purposes, human workers will become an economically inefficient choice. But there will be opportunities in fields where human interactions matter, and some of those will require government money.

Europe will probably need help from the U.S. in 5-10 years. Depending on the agenda of whoever will govern the U.S. at the relevant time, a Marshall Plan 2.0 may effectively fracture the EU. U.S. plans to pull select countries away from the EU have already leaked and may come to fruition in the wake of a € collapse.⁵⁴

14. How you can prepare and protect yourself

Europe is losing substance. As discussed further above, it spends half a trillion euros on Big Tech products, services and licenses every year because it lacks digital sovereignty. That amount dwarfs Europe’s trade surplus in physical goods. Furthermore, some tech platforms increasingly tax traditional businesses, and most of those platforms are not European. As a result, Europe is getting poorer, relative to the U.S. economy.

It should be obvious that Europe can’t pay for Big Tech products and services an amount comparable to the entire federal budget of its largest country and still prosper. It loses substance in the sense that more and more of Europe is foreign-owned. And it must soon tax economic substance.

European governments can’t (and will find it ever harder to) make ends meet based on the taxes they collect on sales, income, and business profits. But their tax rates are so high that further tax hikes risk driving out business and people.

Therefore, it is foreseeable that some European countries will impose wealth taxes and possibly massive “one-time” charges. Germany did so after World War II. That statute (an article of the country’s Basic Law) still exists, as does (in formal terms) the government agency that administered a 50% wealth charge, which included forced mortgages on real estate.⁵⁵ When a government consolidates its finances through a massive charge like this, it must do so retroactively (in that case, by four years). Otherwise, too many wealthy people will leave during the legislative process. Real estate can’t be moved, but prices immediately go down when the political debate begins. In post-war Germany, many large properties had to be subdivided so their owners could sell parts of them in order to afford the wealth charge.⁵⁶

The risk of this happening again is increasing with AI progress and a widening tech gap, as well as due to Europe’s other problems. On December 8, 2023, the national convention of a German government party already voted in favor of a renewed “one-time” wealth charge, which is far from a formal bill, much less the enactment of a law.⁵⁷ But this is how it begins.

Another problem facing countries like the UK, France and Germany is the exodus of wealthy people as well as HENRYs (“high earners, not rich yet”). For now, the numbers are significant, but not huge. There are different factors (professional and private) that make people stay in their countries even when options with major advantages are on the table. But polls show that the number of people contemplating emigration from those countries is at a record high.

Countries like Germany already complicate the emigration of businesspeople by effectively forcing them to sell all of their assets that come with hidden reserves. In the alternative, those assets are subject to an exit charge based on a hypothetical date on the day someone leaves a country.

Caught in a debt trap and headed for a dual-gap disaster, various Western European countries will make it even harder to leave. They have failed and will continue to fail at creating incentives to come or stay, so they need a new equivalent of the German Wall, beyond existing exit charges. There will foreseeably be taxation by citizenship (which the U.S. already has in place, but U.S. taxes are much lower than those of large Western European countries), and when that comes, countries will make it harder to renounce one’s citizenship. Germany has already taken an administrative step in that direction.⁵⁸

Many members of the IP community who are based in Europe cannot leave without giving up their profession. Others have other reasons to stay. But there are still a few things you can do:

• Diversify your investments, also regionally. For example, the UAE or some Latin American countries may be better equipped now for an Artificial General Intelligence and robotics world than Europe. If almost all tech comes from the U.S. or China, those countries will benefit from low energy costs and natural resources.
• Keep an eye on macroeconomics. What is ahead is far worse than the eurozone sovereign debt crisis of the early 2010s. Now it’s about the big countries. All major policies of Western European countries over the course of the past 10-15 years have failed, and the costs are huge. They will want you to pay for that, more than ever. The EU and national governments will make it sound as if they had the situation under control and had a plan. They don’t.
• Think twice before buying real estate in certain countries. If you think it’s a great investment, or if you want to live in a particular house or apartment, it’s still better to consider all the pro’s and con’s. Take into account that AI and robotics actually favor shrinking populations; ask yourself whether countries that have neither the most valuable IP nor natural resources (including access to cheap, abundant and reliable energy) have a great future in the AI and robotics era; and be aware of the possibility of forced mortgages.
• Citizenship is a highly personal choice, but if you can renounce the citizenship of a country in whose future you don’t believe, give it serious consideration. Ideally you would still be able to stay in the country in which you want to live for now, but reduce the risk of being taxed based on citizenship if you choose to leave. Don’t assume that you can always get out easily.

Just like Europe missed the window of opportunity to make it big in the Digital Era, make sure you won’t miss the opportunity to get out at the right time. Of course, some people love their country so much that you’d rather lose a double-digit percentage of their wealth than leave.

---

Why did I write about so much more than the lack of European AI patents per se?

Why this article:

• Good news is always more popular, but food for thought is sometimes necessary. I’m sure that my predictions will be validated. Those who acted upon them will be grateful. Those who didn’t believe me may or may not remember, but it will be well-documented that I saw it all coming and warned people. It would be nice to be proven wrong, but I will unfortunately be right. I care about the patent litigation community, and knowing that many of its members are bound to a single jurisdiction (even in the UPC era), I want to help them salvage as much of their hard-earned money as possible.
• As explained further below, I have insights into the mechanics of the digital economy that I gained outside the world of IP.
• Europe’s bleak future is also a long-term problem for the UPC. Other jurisdictions will become more and more important at Europe’s expense. The UPC will peak at some point not because its own potential is exhausted, but because of Europe’s economic decline relative to other jurisdictions.

Why me: I know in theory and in practice how the Digital Economy works:

• In 2017, I already discussed most of these issues, mostly on email and to some extent on the internet. It was the year in which I decided to leave Germany (which just didn’t work out as quickly as I had hoped). The problems I identified back then were all real, and all of my concerns have been validated. Nothing worked out better. My negative predictions included the challenges that the European automotive industry was going to face (just that in 2017 I focused on American projects such as the Apple Car, understanding the potential of Chinese car makers only a few years later). One thing even I did not anticipate was Germany’s energy insanity, an unprecedented level of criminal self-destructiveness.
• In 2003, I had a meeting in Menlo Park (then the venture capital center of the world) with two Benchmark partners. At the time, Benchmark was famous for its investment in eBay. Years later, the fund also invested in Uber. When we were finished with our main topic, one of them had to leave for a meeting: Bruce Dunlevie, who had invested in Rambus in 1990 (years before founding Benchmark; he remained on Rambus’s board until 2011). The other, Kevin Harvey, asked me to stay. Kevin wanted to know why a country like Germany never created a startup that they would want to invest in. I gave him various reasons. Not one of them has been addressed since then by any European government. And now it is too late.
• Early on, I focused on transatlantic technology partnerships. In 1988, between my last two high-school years, I did my first business trip to California (Berkeley). In 1996, I co-founded an online gaming startup, which we sold to Telefónica in 2000. I had raised money from an international group of investors and had various Silicon Valley meetings during the dotcom boom of the second half of the 1990s. Then I became an adviser to the CEO of MySQL and a small shareholder in this Swedish company that was acquired by Sun Microsystems for $1 billion in the European Venture Capital Deal of the Year (an EVCA award). Benchmark invested in MySQL, and the fact that I knew them before and helped the two parties come together was the reason for the 2003 Menlo Park meeting. It was because of MySQL that I took an interest in patent policy in 2004. I have also been involved with other (less successful) startups and helped a venture fund with due diligence.
• I learned a thing or two about the dynamics of competition in the digital economy when working on antitrust matters such as Oracle’s acquisition of Sun (where MySQL was the strategically most important part of the deal), mobile app store gatekeeper issues, the cloud computing market, and similar topics.
• Unless one understands the way digital startups succeed or fail as well as the power play between Big Tech corporations, it is impossible to fully understand the situation in which Europe finds itself now. Even economists I hold in high regard and with whom I agree on a number of important questions tend to underrate the importance of the digital sector and, in my opinion, do not fully appreciate what makes companies succeed or fail in that sector.
• This website has an audience that has the intellectual capacity to understand the issues, provided in each case that someone is prepared to face the facts without letting emotions get in the way. Whoever works in the patent community is by definition not a one-trick pony: patent law requires two disparate talents. I do, however, believe that I can help many in IP understand certain tech business dynamics.

---

Footnotes

1. Germany is the largest European economy and files more EPO patent applications, in absolute numbers, than any other European country (March 25, 2025 article by government agency Germany Trade & Invest). Some smaller countries are more innovative on a per-capita basis (page 5 of EPO Patent Index 2024 (PDF)), but the next-largest European economies are much weaker. Per million inhabitants, five smaller European countries (Switzerland, Sweden, Finland, Denmark and the Netherlands) rank higher than Germany, though Switzerland and the Netherlands do not achieve that result because of local R&D. Germany’s 300 EPO patent applications per million inhabitants tower far above other large European countries such as France (160), the UK (88), and Italy (82). ↩︎
2. Germany is currently the weakest performer among the world’s major advanced economies, having entered a period of structural stagnation that began in late 2019. As of late 2025, the German economy is essentially the same size it was six years ago. According to the European Commission and the ifo Institute, Germany’s price-adjusted GDP in 2024 was reported to be roughly at its 2019 level. Projections from the IMF and the German government for 2025 show growth of only 0.1% to 0.2%. This confirms that the economy has effectively stood still for over half a decade. While the total economy is flat, the per capita numbers (GDP divided by population) are arguably worse. Because Germany’s population has grown slightly due to migration, the “slice of the pie” for the average person has actually shrunk. Real GDP per capita in Germany reached an all-time high in 2019 ($44,207 in constant USD) and has not surpassed that level since. In 2024/2025, it remains slightly below the 2019 peak. ↩︎
3. The EU’s own Joint Research Centre has sounded the alarm: Quantum technology: 32% of the world’s companies, but only 6% of patents are from the EU (October 12, 2025 JRC news announcement). ip fray published the list of the top 20 quantum patent holders as the last image of a December 17, 2025 article. The only European entity is ranked 17th: CNRS (Centre National de la Recherche Scientifique (CNRS website). ↩︎
4. E.g., December 22, 2025 NIUS article. While NIUS is unpopular with the German government and often portrayed by mainstream media as being “far-right” or “right-wing populist”, leading German politicians have accepted invitations to be interviewed by NIUS, among them Jens Spahn (chair of the governing party’s parliamentary group) and a couple of prime ministers (governors) of German federal states. Of non-mainstream German media that discuss the most pressing issues facing that country, NIUS is the most popular and most centrist one. ↩︎
5. In 2024, China produced as much AI research as the U.S., UK, and 27 EU member states combined (July 10, 2025 article by Digital Science). There are, however, debates over whether the U.S. was nevertheless the leader in terms of impact. The U.S. and China combined have 60% of the world’s AI researchers (July 4, 2025 article by Xinhua). ↩︎
6. Singular Computing v. Google (January 25, 2024 ai fray article); ParTec v. Microsoft (June 10, 2024 ip fray article); ParTec v. NVIDIA (October 28, 2024 ip fray article) ↩︎
7. KEEEX v. OpenAI, Adobe et al. (August 2, 2025 ip fray article) ↩︎
8. November 5, 2025 article by Discovery Alert ↩︎
9. E.g., JP Morgan 2026 Outlook (PDF) ↩︎
10. A Canadian musician was wrongly accused of offenses by Google, causing real-world harm including a canceled concert (December 23, 2025 CBC article). ↩︎
11. E.g., February 24, 2025 sanctions order in Wadsworth v. Walmart (Findlaw) ↩︎
12. October 1, 2024 article on UCLA website ↩︎
13. July 24, 2025 article by CleanTechnica ↩︎
14. How the Merkel Government Makes up Its Statistics (October 3, 2018 article by Dr. Daniel Stelter) ↩︎
15. If children weren’t born during a certain era to replace those who retire, there is no going back in time to change it. Other than through taxes and debt, governments can just raise the retirement age. Migration could help, but as explained herein as well, Europe’s migration policies have been a failure since the post-war period and even more so since 2015. Instead of solving the problem, they have created additional hidden public debt. Medical progress, while desirable in its own right, exacerbates the problem (retirees live longer), though it can also be invoked to justify a longer work life. ↩︎
16. At minimum, there was the equivalent of willful blindness. Restructuring a pension system from a generational contract into a retirement savings or long-term public investment-based system cannot be done when the car is about to hit the concrete wall. It must be put in place at least a couple of decades before. Most European countries are way beyond the point of no return. ↩︎
17. Democracy: The God That Failed (Wikipedia page) ↩︎
18. The U.S. is also fiscally irresponsible, but has multiple advantages over Europe to deal wih the problem: a stronger economy that is fit for the AI/robotics future, vast natural resouces (including a strong agriculture) making it self-sufficient, and current U.S. tax rates are far lower than those of large Western European countries, making it possible to raise taxes without choking off the economy. ↩︎
19. In 2024, Professor Bernd Raffelhüschen, an expert in public finance with a focus on social security systems, published a study (PDF (in German)) according to which the long-term costs of Germany’s migration policies amount to €5.8 trillion. He explains that high-skilled immigration could make a positive contribution, but bases his estimates on the bottom-line impact of current policies. ↩︎
20. April 12, 2025 article by R&D World ↩︎
21. While a net payer is very likely positive for a country, even net recipients are desirable if they perform jobs that society needs done but which do not pay well enough. ↩︎
22. For example, there never was a need to open the floodgates over the war in Syria. The culturally more compatible countries that are closer to Syria than to Central Europe have about 20 times the area of Germany, and it would have been much cheaper to provide them with food and shelter there. Migrants from other continents pay human traffickers many thousands of dollars. The “bottom billion” live off about one dollar a day, so they cannot afford trafficking. ↩︎
23. The electricity needs of AI were harder to anticipate than the impact of further automation on the labor market. ↩︎
24. For lack of any other mechanism, the EC shoehorned Ireland’s corporate tax practices into a state aid case against Apple, and the ECJ ultimately supported this legally questionable theory. ↩︎
25. “Moral” in this context means “non-physical” and “hazard” does not mean an immediate accident, but a risk factor. A physical hazard would be a tree on a road. A moral hazard is something that creates risks because of what it leads decision makers to do. ↩︎
26. Trends in International Mathematics and Science Study ↩︎
27. E.g., German math professor Bernhard Kroetz operates a popular YouTube channel and sometimes talks about his conversations with colleagues from other European countries, such as France, where the situation is even worse. ↩︎
28. Schulphysik ohne Physik (YouTube video by Professor Bernhard Kroetz) ↩︎
29. European banks are also fairly small now compared to major U.S. rivals. ↩︎
30. North America accounted for 74% of global exit volume (IPOs and M&A) in early 2025, while European exits remained stagnant, seeing a 65% decline in IPO activity compared to the previous year. ↩︎
31. “A new study finds Israeli tech companies have expanded sharply across Europe, employing more than 30,000 workers; Activity is concentrated in EU states, with strong growth in R&D, AI, cybersecurity and life sciences, and rising hubs in Central and Eastern Europe .” (December 24, 2025 YNet article) ↩︎
32. Companies like Apple and Google pay much lower salaries in Europe than in the U.S. ↩︎
33. E.g., “You old daughter of a b….” (April 28, 2025 Focus article); a February 28, 2023 article by German state-owned TV station ZDF does not mention migration explicitly, but contains a chart that shows a massive increase between 2018 and 2022; November 4, 2024 News4Teachers article ↩︎
34. Even professional soccer players, whose intelligence is often underestimated, compare competing offers on that basis. In that field, some contracts even state after-tax amounts. ↩︎
35. SAP connects over 5 million companies globally. When a buyer joins the network, they immediately gain access to a massive pool of pre-vetted suppliers, and suppliers can transact with more buyers. Third-party software developers (like Salesforce or specialized industry apps) are more incentivized to build integrations and add-ons for SAP. Hundreds of thousands of certified SAP consultants globally (e.g., from firms like Accenture or Deloitte) are available, as are employees with SAP-specific skills. ↩︎
36. For example, the so-called “Paypal Mafia” is is a group of former employees and founders of Paypal who have gone on to shape the modern tech landscape through the creation of numerous massive enterprises, including Elon Musk (SpaceX, Tesla, OpenAI, Neuralink, The Boring Company, xAI), Peter Thiel (Palantir Technologies, Founders Fund, Clarium Capital, Valar Ventures, Mithril Capital), Reid Hoffman (LinkedIn, Inflection AI), Max Levchin (Affirm, Slide, Glow), David Sacks (Yammer, Geni.com, Craft Ventures), Steve Chen (YouTube), Chad Hurley (YouTube), Jawed Karim (YouTube, YVentures), Jeremy Stoppelman (Yelp), Russel Simmons (Yelp), Ken Howery (Founders Fund), Luke Nosek (Founders Fund, Gigafund), Keith Rabois (OpenStore), Premal Shah (Kiva), and Yishan Wong (Terraformation). ↩︎
37. The first high-profile startup to go against this trend is the new AI startup founded by Yann LeCun, a Frenchman who was Meta’s AI chief, in Paris. If he actually managed to create Artificial General Intelligence there, it could be a game changer. But this is an absolute outlier. ↩︎
38. November 20, 2025 article by Science|Business ↩︎
39. August 21, 2024 article by EQT ↩︎
40. Legal services, including those provided by notaries public, are necessary for a functioning economy. For certain transactions, notarization is reasonably required. However, there are purposes for which the requirement to hold a shareholder assembly in front of a notary public is unreasonable. ↩︎
41. German auction platform Alando was acquired within only about six months of its creation by eBay). German MySpace imitation Lokalisten, German Facebook imitation StudiVZ and German LinkedIn imitation XING were acquired by local media conglomerates. Lokalisten and StudiVZ shut down a while ago, and XING is of marginal relevance and probably also destined to be shut down. The Holtzbrinck Publishing Group could have traded its German StudiVZ (“student directory”) platform for 5% of Facebook’s shares, but foolishly declined. ↩︎
42. Substitution still happens in certain segments. Here are some examples: emails instead of physical letters or postcards; some of the students who use educational software would otherwise pay for tutoring; some who use a search engine and/or social media stream or talk to an AI chatbot would otherwise buy a book or magazine. ↩︎
43. Amazon’s acquisition of organic grocery store chain Whole Foods is an example. For the most part, however, tech companies abstain from such deals because too large a non-tech component in their portfolio could bring down their multiples. ↩︎
44. Full name: tbg Technologie-Beteiligungsgesellschaft der Deutschen Ausgleichsbank ↩︎
45. The tbg was integrated into the Kreditanstalt für Wiederaufbau (KfW). ↩︎
46. During the most critical period (which was when the fast-eat-slow window of opportunity was open), the UK was part of the EU. Also, the structural and societal issues that prevented the UK from making it big in AI are rather similar to those facing countries like Germany and France. ↩︎
47. The exact count depends on whether one includes non-EU countries that are part of the European market. Also, some regional languages, particularly Catalan, are commercially relevant, while Irish is not. ↩︎
48. Language and cultural barriers as discussed. ↩︎
49. It results in adverse selection (on average, not in every individual case) when a country’s social welfare benefits are more attractive than its tax rates, its tech salary levels, and the availability of venture capital. The language barrier (which does not apply to the UK, of course) was addressed before. ↩︎
50. The climate change movement often points to China’s huge investments in renewable energy. In absolute terms, it is true that China does a lot in that regard. But it doesn’t make itself, and particularly not its data centers, dependent on renewable energy. ↩︎
51. Europe would have had to build more nuclear reactors well ahead of time. ↩︎
52. June 23, 2023 article by Quote Investigator ↩︎
53. It works relatively well at the moment, but if the EU had done the same in the era of USB-A, it would have impeded some very important progress. ↩︎
54. December 12, 2025 article by Investment Migration Insider (IMI) Daily ↩︎
55. The 50% charge was collected in 120 quarterly installments (thus 1.67% per year over the course of 30 years). On the remaining forced debt, interest was collected. ↩︎
56. For 20 years, I lived in Starnberg, a town by a lake in the Greater Munich Area. I learned about those post-war forced mortgages when a local told me that Starnberg’s two most expensive streets, Prinzenweg and Almeidaweg (which offer a largely unobstructed lake view from a hill), resulted from the subdivision and partial sale of a huge property that belonged to the Almeida family, which could not possibly have afforded the interest on their forced mortgage. ↩︎
57. YouTube video by attorney-at-law Dr. Tim Greenawalt (Expats Global): SPD beschließt Vermögensabgabe ↩︎
58. Dual citizenship was previously an opt-in regime in Germany: anyone becoming citizen of another country automatically lost their German citizenship short of requesting permission for dual citizenship. Since 2024, an opt-out from dual citizenship is now required. Germany does not allow anyone to become stateless, but presenting proof of a new citizenship is not enough. In practical terms, renunciation will be allowed except in some outlier circumstances (such as related to military service). But this is a first procedural change that could lead to significant complications further down the road. ↩︎