The number that broke the internet this week: $30 billion. That’s how much Nvidia — the world’s most valuable company — is about to pour into OpenAI’s latest mega funding round. But the headline number isn’t the story. What happened before this deal, and what it reveals about AI’s increasingly tangled power structure — that’s the story.
The $100 Billion Deal That Died
Rewind to September 2025. Nvidia and OpenAI announced a blockbuster: Nvidia would invest up to $100 billion to support OpenAI’s chip deployment across massive data centers. Stock soared. AI’s ultimate power couple was born.
Except the deal never closed.
By January 2026, the Wall Street Journal reported the arrangement was “on ice.” Nvidia’s own quarterly report admitted there was “no assurance” of definitive agreements. The $100 billion commitment was a handshake that got complicated fast.
The new deal is fundamentally different. And that difference is revealing.
Why This Deal Is Cleaner — and Smarter
The original $100 billion arrangement was essentially circular: Nvidia invests cash, OpenAI uses it to buy Nvidia chips, money goes round and round. Critics called it a valuation-inflating merry-go-round.
The new $30 billion is a straight equity play. Nvidia buys OpenAI stock. No commensurate chip-purchase commitment. No circularity.
Three things this signals:
- Both companies heard the skepticism about circular AI deals
- Nvidia sees value in OpenAI beyond hardware sales — it wants equity upside
- OpenAI gets flexibility to diversify chip suppliers (which it’s already doing with AMD and Broadcom)
By taking equity instead of a supply agreement, Nvidia ensures it profits from OpenAI’s rise regardless of whose chips end up in the racks. It’s a hedge disguised as a bet.
The Valuation Trajectory Is Staggering
The round values OpenAI at roughly $730–830 billion pre-money. The growth curve:
- 2023: $29 billion
- Early 2024: $80 billion
- Late 2024: $157 billion
- Early 2025: $300 billion
- February 2026: $730–830 billion
That’s roughly 25x in three years for a company still burning cash at an extraordinary rate. Nvidia isn’t alone at the table — SoftBank, Amazon, and Microsoft are all expected to participate in what could be a $100 billion total round, one of the largest private capital raises in history.
The Market Share Problem Nobody Wants to Talk About
While OpenAI’s valuation climbs, its dominance is eroding fast.
ChatGPT’s share of U.S. daily AI users dropped from 69.1% in January 2025 to 45.3% in January 2026. Web traffic metrics tell a slightly different story (64–68% depending on the source), but the trend is unmistakable.
Google’s Gemini is eating consumer share. Anthropic’s Claude is reportedly leading in enterprise AI adoption. xAI’s Grok is growing. And DeepSeek proved you don’t need OpenAI-sized budgets to build competitive models.
So why is Nvidia writing a $30 billion check for a company losing market share?
Follow the Chips, Not the Chat
Nvidia’s bet isn’t about ChatGPT’s chatbot numbers. It’s about infrastructure.
OpenAI will use much of the fresh capital to purchase AI chips — and a significant chunk will likely be Nvidia GPUs regardless of formal commitments. OpenAI’s models run on Nvidia hardware. Its upcoming projects require massive compute. The money flows back whether it’s contractually obligated or not.
But OpenAI is hedging too, with AMD and Broadcom partnerships and custom silicon exploration. Nvidia’s near-monopoly on OpenAI’s compute is probably ending — which might explain why equity made more sense than another supply deal.
If OpenAI hits a $1 trillion valuation, Nvidia’s $30 billion stake could be worth multiples of that. Win-win, even if OpenAI starts buying someone else’s chips.
Everyone Invests in Everyone. Everyone Competes with Everyone.
Zoom out and the web is dizzying:
- Nvidia makes the chips that train the models
- OpenAI builds the models that drive chip demand
- Microsoft provides cloud infra, invests in OpenAI, and competes with it
- Amazon competes with Microsoft on cloud but also invests in OpenAI
- SoftBank builds data centers and invests in the companies filling them
Some analysts have compared this to the pre-2008 financial sector — so interconnected that distinguishing counterparties from competitors becomes impossible. That comparison might be dramatic. But when Nvidia’s earlier $100 billion deal collapsed, markets trembled and software stocks sold off. The interdependence is real.
What to Watch
OpenAI’s path to profitability. The company recently started testing ads in ChatGPT. With market share declining and costs astronomical, it needs to prove it can make money, not just raise it. An IPO looms.
Nvidia’s portfolio play. This isn’t a one-off. Nvidia is quietly building equity stakes across the AI ecosystem — becoming not just the arms dealer, but a financial beneficiary of every company building on its platform.
Regulators. When the world’s most valuable company invests $30 billion in one of the most influential AI companies, antitrust scrutiny follows. Both the U.S. and EU are watching.
China. DeepSeek, ByteDance’s Seedance, and other Chinese AI companies keep proving this isn’t a two-horse race. The geopolitical stakes are only growing.
The Bottom Line
This deal only makes sense in a world where AI is the most important technology on the planet — and where the companies building it are so entangled that traditional notions of competition and investment have broken down entirely.
It’s a bet on OpenAI’s future. A hedge against Nvidia’s customer diversification. A signal that the biggest players are doubling down even as cracks in the narrative emerge.
The real question isn’t whether this deal closes. It’s whether the AI industry can generate enough real-world value to justify the trillions flowing into it.
We’re either witnessing the greatest technological transformation in human history — or the greatest investment bubble. Maybe both.