AI funding is concentrated

Venture capital is no longer spreading its bets across the AI field. It is piling into a few companies that look less like startups and more like private versions of the old platform monopolies. Crunchbase says global venture investors put about $300 billion into startups in the first quarter of 2026, the highest quarterly total on record. More than four-fifths of that money went to AI-related companies. In North America alone, AI soaked up more than 87% of all venture dollars. (news.crunchbase.com) Inside that surge, the money got even narrower. Crunchbase’s sector data shows funding for foundational AI startups in Q1 was already double all of 2025. But the real story is concentration. OpenAI, Anthropic, xAI, and Waymo absorbed an enormous share of the quarter’s capital. OpenAI and Anthropic alone accounted for roughly 57% of all venture investment in the quarter, according to the reporting cited in today’s coverage. That is not a hot market. That is a market being reorganized around a few giant balance sheets. (news.crunchbase.com) OpenAI is the clearest example of what that new structure looks like. On March 31, the company said it closed a funding round with $122 billion in committed capital at an $852 billion post-money valuation. A year earlier, OpenAI had announced a $40 billion raise at a $300 billion valuation. In two rounds, it moved from being a richly funded lab to something closer to private infrastructure. Its own description makes the shift plain: the money is for research, compute infrastructure, and becoming “core infrastructure for AI.” (openai.com) That language matters because these companies are no longer being valued mainly as software products. They are being valued as systems that need chips, power, datacenters, cloud contracts, and financing partners large enough to carry all of that. OpenAI’s Stargate project, announced in January 2025, set out to invest $500 billion over four years in U.S. AI infrastructure, with $100 billion to be deployed immediately. Later partnerships with SoftBank, Oracle, SB Energy, and Nvidia extended that buildout into energy supply and multi-gigawatt datacenter capacity. The frontier model business is starting to look like a utility with a chatbot attached. (openai.com) Anthropic fits the same pattern from the other side. In February, it announced a $30 billion Series G round at a $380 billion post-money valuation. Amazon had already invested billions and tied Anthropic closely to AWS. So even when capital appears to be flowing into a rival, it is still flowing through the same narrow set of cloud and compute channels. The market is choosing a few model makers, then binding them to a few infrastructure suppliers, then calling the result competition. (anthropic.com) That is why the governance questions now matter as much as the models. When a company needs outside investors, hyperscale cloud partners, energy developers, and chip suppliers all at once, board oversight stops being a formality. It becomes the only thing standing between growth and overcommitment. OpenAI’s internal debate over spending and IPO timing shows how quickly those pressures can collide. Reports published on April 6 say CFO Sarah Friar has raised concerns about whether the company is ready for an IPO as early as the fourth quarter of 2026, given rising infrastructure costs, internal readiness work, and plans that could involve up to $600 billion in spending over five years. (moneycontrol.com) That tension is the real story behind the giant rounds. Late-stage AI funding is not just buying product growth anymore. It is financing a capital structure. OpenAI’s valuation now sits alongside questions about loss tolerance, compliance, supplier dependence, and who gets to say no when the spending curve outruns the operating one. Reports about executive role changes tied to IPO preparation only sharpen that picture. The frontier AI race is still about intelligence, but the companies winning money are the ones convincing investors they can survive the physics of the buildout.