NVIDIA stock has surged by more than 880 percent over the past three years. In 2024 and 2025 alone, AI and semiconductor stocks reached valuation levels that investors had last seen during the dot-com era. At the same time, global AI investments are expected to exceed 2.5 trillion US dollars in 2026, with roughly half flowing into data centers and infrastructure.
Anyone buying now is betting on the rally continuing. Anyone selling now is betting on a correction. And anyone waiting is paying a price as well the opportunity cost of missing gains in a market that has disproven skeptics for two straight years.
This article examines the arguments supporting the AI bubble thesis, the arguments against it, when a crash becomes more likely, and what all of this specifically means for investors. For additional context, we also recommend our analysis of the potential 2026 financial crisis in Europe.
Everything at a Glance
- AI spending is projected to surpass 2.5 trillion US dollars in 2026, representing a 44 percent increase compared to the previous year.
- NVIDIA, Broadcom, and other leading AI stocks are trading at price-to-sales ratios above 30 levels that have historically often preceded sharp market corrections.
- Economists such as Ruchir Sharma warn that the bubble could burst if US interest rates rise and cheap capital becomes scarce. Meanwhile, Goldman Sachs and J.P. Morgan argue that the growth is fundamentally justified.
- The key difference compared to the dot-com era: Microsoft, Alphabet, Meta, and Amazon are funding their data center investments through ongoing cash flow rather than debt or equity raises.
- For investors in 2026, the central question is no longer whether AI is transformative. The real question is whether current valuations have already fully priced in that transformation.
1. What Is an AI Bubble?
A bubble does not form simply because prices rise. A bubble forms when prices rise faster than the underlying economic fundamentals. The dot-com bubble around the turn of the millennium was not a misunderstanding about the importance of the internet. The internet truly was transformative. The bubble emerged because companies with no business model, no revenue, and no profits were being traded at multi-billion-dollar valuations.
That is why the key question in 2026 is not whether AI is real. The real question is whether current valuations reflect actual profits or whether they are pricing in profits that do not yet exist.
The Numbers Behind the AI Boom in 2026
Investment Volume and Growth Rates
- Global AI spending in 2026 is expected to exceed 2.5 trillion US dollars
- Growth compared to 2025: approximately 44 percent
- Forecast through 2029: 3.3 trillion US dollars, with an average annual growth rate of around 22 percent
- Infrastructure share: roughly 50 percent of investments are flowing into data centers, chips, and network infrastructure
2. What the Hyperscalers Are Spending
A helpful comparison from FAZ illustrates the scale of the current AI investment boom: the major cloud providers like Amazon, Alphabet, Meta, and Microsoft have announced roughly 660 billion US dollars in data center investments for 2026 alone. By comparison, Germany’s special infrastructure fund, approved by the Bundestag to modernize the country for the future, amounts to 500 billion euros spread across twelve years.
The key difference lies in the financing. The hyperscalers are funding these investments through ongoing cash flow. They have real revenue, real profits, and strong balance sheets. The German government is financing its infrastructure package through debt. This is not a criticism of German fiscal policy. It is simply a comparison that highlights how fundamentally different the current tech rally is from the dot-com era.
3. What Are the Arguments for an AI Bubble?
Valuations Beyond Historical Norms
The price-to-sales ratios of NVIDIA and Broadcom temporarily exceeded 30 in 2025. Historically, a price-to-sales ratio above 10 has been considered high, while anything above 20 has often been viewed as speculative. Levels above 30 have regularly signaled overvaluation in previous market cycles, often followed by sharp corrections.
This does not automatically imply a crash. But it does mean the market is pricing in uninterrupted growth for many years in order to justify current valuations. Any disappointment whether in quarterly earnings, sales forecasts, or interest rate expectations could destabilize this fragile structure.
The Profitability Problem
Billions are being invested into AI infrastructure. The question of when, and whether, these investments will generate proportional economic value remains unanswered. Critics, including prominent economists and investors such as Ruchir Sharma, argue that many AI applications currently promise productivity gains that are not yet visible in broader economic data.
Sharma goes even further. He believes the AI bubble could burst if US interest rates rise due to persistent inflation. Higher interest rates reduce the present value of future profits. And growth stocks with extremely high price-to-sales ratios are valued almost entirely on expectations of future earnings.
Interest Rate Risk
An analysis by n-tv highlights how closely the tech rally is tied to favorable financing conditions. Rising US interest rates would increase the cost of capital, reduce the present value of future growth expectations, and shift investment flows toward lower-risk asset classes.
The market already demonstrated this vulnerability in 2022. Within a single year, the NASDAQ lost roughly 33 percent as higher interest rates triggered a sharp correction in technology-heavy portfolios.
4. What Are the Arguments Against a Crash?
Fundamental Differences Compared to the Dot-Com Era
The dot-com bubble burst because companies with no revenue and no viable business models were traded at absurd valuations. Today’s situation is structurally different:
- Microsoft, Alphabet, Meta, and Amazon generate real profits in the hundreds of billions of dollars.
- The balance sheets of major tech companies contain cash reserves comparable to those of smaller national economies.
- AI chips such as NVIDIA’s H100 and B200 series are being purchased by paying enterprise customers, not speculative startups.
- Early AI applications in enterprise software, cybersecurity, and specialized infrastructure are already generating high profit margins.
Goldman Sachs and J.P. Morgan view the current growth as fundamentally justified. Their assessment: AI is not a speculative bet on a distant future monetization has already begun.
Competition Keeps Quality High
Finanzen.net highlights another important dynamic. The year 2026 is being shaped by new chip generations and High-Bandwidth Memory (HBM). Competitors such as AMD are attempting to close the gap with market leader NVIDIA.
This competition forces genuine innovation and prevents prices from being driven solely by hype. Where real competition exists, markets tend to regulate themselves more efficiently.
When Could the AI Bubble Burst?
There is no precise answer. But there are several conditions under which a serious correction becomes more likely:
- Rising US interest rates: If the Federal Reserve is forced to raise rates due to persistent inflation, growth stocks could lose their valuation support.
- Disappointing earnings expectations: If hyperscalers fail to meet the aggressive growth assumptions already priced into their quarterly results, corrections could happen quickly.
- Regulation: Government intervention in AI markets for example through data privacy laws or antitrust regulation could impact business models that have so far operated with limited regulatory pressure.
- Geopolitics: An escalating trade conflict between the United States and China could disrupt semiconductor supply chains and threaten production capacity. For more on the connection between trade wars and economic downturns, see our analysis of the potential financial crisis in Germany in 2026.
IN JUST 5 MINUTES:
In just 5 minutes: Become a sponsor – Your entry into attractive litigation financing opportunities
1
Register as a sponsor
2
Select a case
3
Set the bid amount and quota
4
Provide PayPal or credit card details
5
Participate in the litigation proceeds
5. What Does This Mean for Investors?
Anyone invested in a tech-heavy portfolio in 2026 should honestly answer two questions:
What percentage of my capital could lose 30 to 40 percent in the medium term without changing my financial situation?
Am I invested because I truly understand these companies, or simply because the market has been right so far?
For investors seeking diversification outside of economically and market-correlated assets, litigation finance offers a structurally different approach. The performance of a financed legal case does not depend on the NASDAQ, interest rate decisions, or valuation multiples. Courts issue rulings independently of whether markets are rising or falling.
To learn more about how this works as an investment model, see our article on litigation finance as an impact investment in 2026.
Participate in Market-Independent Investments as a Sponsor
Investors looking to reduce their portfolio concentration in market- and economy-correlated assets can access vetted legal cases through AEQUIFIN, whose performance is structurally independent of stock market movements.
For a deeper look into the opportunities and risks of litigation finance, read our comparison article.
Register now as a sponsor and review vetted legal cases directly.
