Four AI firms drew $188 billion of Q1 venture funding
Creandum investors say OpenAI, Anthropic, xAI and Waymo captured 60% of venture capital deployed in the quarter, changing startup fundraising expectations.
By Sofia Marchetti · World Affairs Correspondent
3 min read
Four AI companies absorbed most of the venture capital deployed in the first quarter of 2026, according to Creandum investors Carl Fritjofsson and Cameron Sellers. The concentration matters for startups because investors are writing fewer, larger checks while demanding clearer defenses against AI platforms.
Fritjofsson, a partner at Creandum, and Sellers, a vice president at the firm, wrote in a Fortune commentary that venture investors put $300 billion of new capital to work in Q1. They said that was more than twice the prior quarter’s total and more than 70% of all venture spending in 2025.
Of that amount, $188 billion went to OpenAI, Anthropic, xAI and Waymo, according to the Creandum investors. They said those four companies accounted for 60% of all venture capital deployed during the quarter.
Funding rose, but deals fell
The headline numbers mask a sharper split in the market, Fritjofsson and Sellers said. They reported that U.S. investment rose 190% from a year earlier in Q1, while the number of deals declined 26%.
That pattern shows capital flowing into larger rounds for fewer companies, according to the Creandum investors. Still, they said the $112 billion that did not go to the four AI giants remains close to recent quarterly highs, leaving room for smaller rounds.
Fritjofsson and Sellers argued that frontier AI lab financings do not compete directly with pre-seed, seed and Series A rounds. They cited Stripe data showing the top 100 AI-native companies are growing from $1 million to $30 million in annual recurring revenue five times faster than earlier generations of software companies.
Startups face a higher bar
The Creandum investors said early-stage founders are under pressure to show stronger pricing power, faster growth and better defenses than they did a year ago. They warned that companies built around thin layers on top of existing AI models face pressure from platform owners above them and falling prices below them.
They pointed to Anthropic’s Claude Cowork as an example of how quickly model-company launches can hit software valuations. According to Fritjofsson and Sellers, the product’s launch erased hundreds of millions of dollars from the market values of some major software companies within hours.
They said the debate over whether software-as-a-service companies are threatened by AI remains unsettled as of June 2026. Their view is that investors are now asking which software companies gain strength as models improve and which ones become easier to replace.
Defensibility drives investor interest
Fritjofsson and Sellers said investors are paying close attention to robotics, defense, photonics, biotech and new computing approaches. They said those areas offer assets and expertise that foundation model companies cannot easily copy.
Robotics and defense companies often require physical integration, complex systems and long customer relationships, according to the Creandum investors. They said photonics and novel compute depend on years of scientific work and manufacturing skill, while biotech can combine proprietary data, regulation and specialized knowledge.
The authors said startup founders should expect a turbulent rest of 2026. They said capital is concentrating across stages, the revenue bar for raising money has roughly doubled over five years, and the profile of funded companies has changed sharply from three years ago.
This story draws on original reporting from Fortune.