OpenAI has finalized a $110 billion funding round at a $730 billion valuation. That number — $730 billion — puts the company, which has never turned a profit, in the same neighbourhood as ExxonMobil and Berkshire Hathaway. According to Bloomberg, OpenAI is now also in early talks with TPG and Bain Capital to launch a $10 billion private equity venture targeting AI infrastructure investments. Thrive Capital, Josh Kushner's fund, raised $10 billion in new capital this month on the strength of its OpenAI position. In February, three companies — OpenAI, Anthropic, and Waymo — accounted for 83% of all venture capital deployed globally. These are not normal numbers, and they deserve more scrutiny than they typically receive in coverage that treats each round as a milestone to be celebrated.
The concentration dynamic is the most striking feature of the current moment. In a typical technology cycle, large venture rounds seed dozens of competing companies, and the market sorts out winners over time. What's happening in AI is structurally different: a handful of frontier labs have captured the majority of available capital, and the gap between them and the rest of the market is widening rather than narrowing. Seventeen US-based AI companies raised rounds of $100 million or more so far in 2026, but those seventeen together received a fraction of what the top three captured. The industry is behaving less like a broad-based technology wave and more like a utility race, where the infrastructure winners take a disproportionate share and others compete for the margin.
Humans&, a startup founded by alumni of Anthropic, xAI, and Google, raised a $480 million seed round in January — a figure that would have been a substantial Series B a decade ago. The company describes itself as building "human-centric AI," a framing that at seed stage tells investors more about the founders' résumés than about the product. The round reflects a real dynamic: in the current environment, a credible team from the right labs can raise pre-product capital at a scale that simply did not exist previously. Whether that is efficient capital allocation or a form of credential arbitrage is a question the market will eventually answer.
The OpenAI-PE venture is the more novel development. Private equity firms raising dedicated AI infrastructure funds represent a new phase: not startup bets, but asset-heavy investments in data centres, power infrastructure, and compute capacity. That is a different risk profile from venture — slower returns, less upside, but more predictable yield — and it signals that the AI buildout is now attracting capital that does not expect to participate in the equity upside of any single company. PE firms are, in effect, treating AI infrastructure as they would treat toll roads or utilities: not exciting, but necessary, and therefore investable at scale.
The valuation question hanging over all of this is whether the revenue will materialise to justify the numbers. OpenAI's $730 billion valuation implies an expectation of tens of billions in annual revenue within the next several years. The company is on a strong growth trajectory — ChatGPT's user base and enterprise contracts are expanding — but the path from here to a valuation that makes sense on traditional multiples requires either sustained explosive growth or a fundamental shift in how investors are pricing AI companies. Both are possible. Neither is guaranteed. The rounds are large enough, and the concentration is tight enough, that a significant re-rating of the sector would not be a contained event.
None of this is an argument that the AI investment wave is wrong, or that the underlying technology does not merit the attention it is receiving. It is an observation that the financial structures being built around AI — the valuations, the concentration, the PE infrastructure plays — are becoming as consequential as the technology itself. How those structures resolve will shape which AI capabilities get built, who owns them, and who pays for them. The funding numbers are not just finance news. They are an early draft of the industry's ownership map.