A Peter Thiel-Backed AI Startup, Cognition Labs, Seeks $2 Billion Valuation

Cognition only began working on its product last year and doesn’t generate any meaningful revenue. It was valued at $350 million earlier this year in a $21 million deal led by Founders Fund. Peter Thiel, the billionaire investor who started Founders Fund, helped lead its investment in Cognition. 

The new fundraising deal hasn’t been finalized, meaning that the terms could change. The company recently has turned down offers at valuations closer to $1 billion—which would still be nearly triple the prior level. 

Cognition is the latest young startup whose value has soared thanks to a boom in artificial intelligence. In December, the French AI model developer Mistral hit a $2 billion valuation, a roughly sevenfold increase from a funding round the prior summer. Perplexity, a two-year-old AI search startup, was valued at $1 billion in a funding round in recent weeks. 

Cognition, co-founded by Chief Executive Scott Wu, was first conceived as a crypto company but pivoted as AI took over Silicon Valley, launched by a wave of generative-AI applications following the release of ChatGPT. 

Cognition introduced its AI coding tool, Devin, earlier this month and said it is able to autonomously complete complex coding tasks such as creating custom websites. Devin was trained using AI models from OpenAI, in which Founders Fund is also an investor.

Some investors say that Devin represents a major leap in AI intelligence and could presage the widespread automation of software development. 

Other companies building similar products are seeing an uptick in momentum. Last quarter, ’s coding tool GitHub Copilot grew its subscriber base by 30% to 1.3 million users. Magic AI, a startup competitor to Cognition, received $117 million from venture capitalists in February. Other coding assistant startups are also in the process of raising competitive funding rounds. 

Despite promising signs of growth, the ballooning valuations of new AI startups has stoked fears that the industry is headed into another bubble. So far, few startups have been able to show how they might recoup the steep costs associated with developing generative AI products. 

In a presentation earlier this month, the venture-capital firm Sequoia estimated that the AI industry spent $50 billion on the chips used to train advanced AI models last year, but brought in only $3 billion in revenue. 

Write to Berber Jin at berber.jin@wsj.com