AI neolabs raised $24B in two quarters, nearly 5x pre-ChatGPT OpenAI and Anthropic

Newly formed AI labs have raised $24 billion over the past two quarters, nearly five times as much as OpenAI and Anthropic collected in the years before ChatGPT was launched. The figure comes from research by Radical Ventures, reported by the Financial Times, and it covers a wave of startups that in many cases have no products, customers or revenue.

Investors are betting that specialized newcomers can still take a place in a market dominated by a handful of large AI groups, even though compute and top research talent remain scarce and expensive.

Billions within months of founding

Emulate, founded by former Google DeepMind researchers, is a month old and already on track for a post-money valuation of nearly $4 billion. Safe Superintelligence, launched by OpenAI co-founder Ilya Sutskever in 2024 without a disclosed product, raised money last year at a $32 billion valuation.

Running one of these labs requires two inputs that are hard to secure: leading researchers, and access to chips and servers. Both stay costly, which pushes new entrants into very large rounds very early.

Faster rounds, stacked valuations

Venture firm Chapter One calculates that funding rounds now come at a median interval of 7.6 months, with a 3.5-times increase in price. That pace shows how quickly capital needs and valuations are rising.

Headline numbers do not always reflect one clean price. Investors can commit funds quietly at one valuation and then join a more public round shortly afterwards at a higher price. The practice lets founders promote the larger implied valuation while backers point to a lower average entry point.

Why investors keep betting

The AI model market is still taking shape, unlike mature sectors such as supermarkets or ridesharing, which leaves room for new players to force their way into the top tier. Several of the newcomers aim at narrower fields, including world models for complex simulations and physical AI for autonomous machines.

Customer preferences are unsettled too. Businesses that use AI tools are still willing to switch providers for better performance or lower cost, an opening for smaller developers competing with incumbents. Suppliers gain from the spending: Nvidia has backed more than half of the cohort named in the report. Compute deals are also getting larger: Anthropic recently agreed a seven-year cloud contract with Akamai worth $11.6 billion, with an option to expand to about $20 billion.