OpenAI ended 2025 with a loss of roughly $39 billion, while spending climbed to $34 billion. Yet investors keep pouring money into the company behind ChatGPT. According to documents seen by the Financial Times and reported by Reuters, OpenAI raised another $122 billion this year and is now valued at more than $850 billion.
That raises a fundamental question: why are investors tolerating losses that are extreme even by Silicon Valley standards?
The answer reveals a lot about today’s AI market, expectations around artificial general intelligence (AGI), and the global race for AI infrastructure.
Billions in losses, breakneck growth
OpenAI lost nearly $39 billion in 2025. That number sounds dire, but a large share reflects accounting charges tied to the company’s reorganization. Excluding those, the operating loss was closer to $8 billion. At the same time, revenue jumped to over $13 billion—triple the year before.
The burn remains massive in 2026. The Information reports OpenAI burned about $3.7 billion in the first quarter alone, despite $5.7 billion in quarterly revenue.
The biggest cost drivers are:
- AI model training
- Data centers
- Nvidia GPUs
- Research and development
- Global expansion of ChatGPT and API services
OpenAI spent more than $19 billion on research and development last year alone.
Why investors still back the bet
Investors aren’t optimizing for today’s profit. They’re betting on who will control tomorrow’s AI market.
The logic mirrors earlier tech waves. Amazon ran losses for years before becoming one of the world’s most profitable companies. Uber, Tesla, and Netflix were valued for future dominance long before they posted steady profits.
The same principle applies to OpenAI.
Backers believe the first companies to field the most powerful AI models can ultimately dominate:
- Business software
- Search
- Digital assistants
- Developer tools
- Education
- Healthcare
- Scientific research
That’s why players like Microsoft, SoftBank, Nvidia, and Amazon keep writing billion-dollar checks. Many investors see OpenAI as the leading candidate to become the core infrastructure layer of the AI economy.
Bubble territory—or not?
Comparisons to past tech bubbles are getting louder.
Critics note AI costs are rising faster than revenue. State-of-the-art models demand ever more compute, energy, and specialized chips. Converting users into paying customers is getting harder. Several analysts warn the sector is too dependent on constant capital injections.
There are also signs OpenAI has missed some internal growth targets. Reports in The Wall Street Journal and elsewhere say both user growth and revenue at times fell short of internal expectations.
Yet today’s landscape differs from the dot-com era in one crucial way.
Real demand for AI exists.
ChatGPT has hundreds of millions of users, companies are wiring AI into their software, and large enterprises are raising AI budgets year after year. The technology already has commercial traction—something many past hype cycles lacked.
How long can OpenAI keep this up?
The short answer: longer than many critics expect.
After its record $122 billion round, OpenAI has an exceptionally large war chest. The company is also moving toward an IPO, possibly as early as 2026. Reuters previously reported an IPO could push OpenAI toward a $1 trillion valuation.
The risks, however, are significant.
OpenAI has accumulated hundreds of billions of dollars in future infrastructure commitments. The coming years will demand enormous spending on data centers, chips, and energy. If revenue growth slows while costs keep rising, pressure on the model could intensify quickly.
What it means for European AI
OpenAI’s finances have direct implications for Europe.
On one hand, the torrent of capital underscores how far ahead U.S. AI companies have pulled. No European AI player has comparable resources or infrastructure.
On the other, this also opens doors.
If investors become more selective about loss-making AI firms, European companies can differentiate with leaner models, lower costs, and sharper focus on industry-specific use cases. Companies like Mistral AI are aiming for exactly that niche.
Meanwhile, digital sovereignty is rising on the EU agenda. That is fueling more political support for European cloud, chip, and AI efforts that reduce reliance on U.S. tech giants.
Bottom line
OpenAI is losing billions, but investors are fixated on the grand prize: controlling the world’s future AI infrastructure.
For now, the losses aren’t scaring them off. As long as revenue climbs, ChatGPT holds its lead, and investors believe in more powerful AI systems ahead, the capital will keep flowing.
The real test comes when the question shifts from how many users OpenAI has to how much profit it can actually generate. Until then, OpenAI remains the defining experiment of the global AI economy.
Latest comments
Loading