Last updated: March 2026
February 27 was a lot, even by AI standards.
OpenAI announced the largest private financing in history: $110 billion, with $50 billion from Amazon, $30 billion from Nvidia, and $30 billion from SoftBank. Pre-money valuation: $730 billion. Post-money: roughly $840 billion. A private startup, ahead of most publicly traded companies on Earth.
That would’ve been enough for a full news cycle. Instead, it was one of four things that happened to OpenAI in the same 24 hours. The other three might matter more.
The Money
A year ago, OpenAI’s $40 billion raise was called the largest private fundraise ever. This one nearly triples it.
The $730 billion pre-money valuation represents a staggering trajectory. In March 2025, OpenAI was valued at $300 billion. By October, secondary market trades pushed it to $500 billion. Now $730 billion, in under a year.
For context: that valuation exceeds all but a handful of public companies globally. And the round isn’t even closed. OpenAI says venture firms, sovereign wealth funds, and other financial investors are still lining up.
Sam Altman’s framing was predictably grand: “AI is transforming the whole economy, and the world needs a lot of collective computing power to meet the demand.” Translation: we need more money, and we’ll keep asking for it.
Amazon’s $50 Billion Play
Amazon’s investment is the most interesting piece. It’s not just a check. It’s a strategic land grab.
The $50 billion comes in two tranches: $15 billion upfront, with $35 billion contingent on OpenAI hitting certain milestones. The Information previously reported those conditions likely include completing an IPO or reaching an AGI benchmark.
Beyond the cash, the two companies expanded their existing $38 billion AWS agreement by another $100 billion over eight years. OpenAI’s commitment was to using at least 2 gigawatts of Amazon’s Trainium chips (covering Trainium3 now and Trainium4 next year). In return, AWS gets exclusive third-party distribution rights for OpenAI’s enterprise platform Frontier.
Here’s the dynamic that matters: Microsoft still holds exclusive cloud hosting for OpenAI’s API and ChatGPT. But Amazon now owns the distribution channel for OpenAI’s enterprise agent platform. As one Chinese tech commentator put it: “Microsoft kept the old world. Amazon bought a ticket to the new one.” (This split matters for the broader AI agent landscape — enterprise agents are where the next wave of revenue comes from.)
Amazon also has a major investment in Anthropic. Regardless of which AI company wins, AWS sells the compute. Classic picks-and-shovels play.
900 Million Weekly Active Users
Buried in the funding announcement: ChatGPT now has 900 million weekly active users. Up from 800 million in October. Over 50 million paying subscribers. Over 9 million enterprise customers.
To put 900 million WAU in perspective: Instagram has about 2 billion monthly actives. WeChat has about 1.3 billion. ChatGPT’s 900 million is a weekly number, which means its monthly figure is higher still. For a product that launched in November 2022, that growth curve is absurd.
Codex, OpenAI’s coding tool, has also seen its weekly active users triple since the start of the year, hitting 1.6 million. OpenAI says January and February set records for new paid subscriptions.
The Insider Trading Scandal
Same day, different energy: OpenAI fired an employee for using confidential information to trade on prediction markets like Polymarket.
Fidji Simo, OpenAI’s head of applications, disclosed the termination in an internal memo. But the story goes deeper than one bad actor.
Unusual Whales, a financial data platform, analyzed prediction market activity tied to OpenAI events going back to March 2023. They found 77 suspicious positions across 60 wallet addresses. These bets targeted Sora’s release date, GPT-5’s launch timing, the ChatGPT browser rollout, and even Sam Altman’s employment status.
The most brazen trade: two days after Altman was fired by OpenAI’s board in November 2023, a brand-new wallet placed a heavy bet that he’d return. It paid out over $16,000. The wallet never traded again.
Before ChatGPT Atlas (the browser feature) launched, 13 fresh wallets appeared within a 40-hour window and collectively bet over $300,000 on the correct outcome.
This is the first time a major tech company has fired someone specifically for prediction market insider trading. It won’t be the last. As Wired reported, prediction markets are becoming big enough that the temptation for insiders is real, and the enforcement infrastructure hasn’t caught up. (Same week, Anthropic was dealing with its own crisis — the Pentagon standoff that ended with a federal ban on Claude.)
The Math That Doesn’t Add Yet
Here’s the number that should give everyone pause: OpenAI’s annualized revenue just crossed $20 billion. Its quarterly losses may exceed $12 billion.
At a $730 billion pre-money valuation, investors are paying roughly 36x annual revenue for a company that’s burning cash at an extraordinary rate. OpenAI is projecting $280 billion in revenue by 2030, split roughly evenly between consumer and enterprise. If they hit that target, the current valuation looks reasonable. If they don’t, this is the most expensive bet in tech history.
The company has also quietly revised its compute spending target down from $1.4 trillion to roughly $600 billion through 2030, according to CNBC. Even OpenAI is starting to acknowledge that the infrastructure buildout can’t be infinite.
Four stories in 24 hours. Record funding, record users, an insider trading scandal, and a Pentagon deal. The money is pouring in faster than ever, and the cracks are showing at the same time. That’s not a contradiction. That’s what a company looks like when it’s racing to build enough revenue to justify a valuation that assumes AI will restructure the global economy.