An early Groq investor says her new $52 million fund expects about half its investments to fail, while betting on AI companies that deliver outcomes in industries such as construction, industrials and insurance. Sandhya Venkatachalam’s account offers a useful look at how one investor is applying lessons from an early bet on AI chips to a broader hunt for real-world AI businesses.
Groq Watch analysis
What happened
Venkatachalam, founder and managing partner of Axiom Partners, told Crunchbase News that the fund expects to make 35 investments and needs one exceptional success to return the fund. She says Axiom looks for companies doing valuable work inside customers’ businesses, rather than conventional enterprise software tools with AI added on.
She also described a team built around people who actively work with AI, and an internal tool called the Axiom Brain to help assess markets and companies. Venkatachalam says she backed Groq in 2016 after its founder, Jonathan Ross, made the case that inference would become a much larger market. She says she barely understood inference at the time, but the investment taught her to be early and patient.
Why it matters
The interview connects two different AI investment bets: the computing infrastructure needed to run models, and businesses using AI to perform work in established industries. Venkatachalam’s stated test is whether a product delivers an outcome customers will pay for, potentially from labour budgets, and becomes hard to replace through deep integration with their data and workflows.
Those are an investor’s criteria and observations, not independently established measures of what AI customers broadly buy or what makes a company durable. Still, the fund’s numbers make the strategy concrete: 35 bets, with failure expected for roughly half and one standout investment needed to make the economics work.
Our read
The most interesting thread is not that AI could reach construction or insurance; it is the insistence that a startup must own the last mile of a real job, not merely put a new interface on a model. That is a demanding test, and a more useful one than counting AI labels. Venkatachalam’s Groq story also shows how a technology investor can make an early bet before the market is obvious, though hindsight is an unusually comfortable co-investor.
What to watch
- Whether Axiom’s investments focus on the industries and outcome-based products Venkatachalam describes.
- Whether portfolio companies can secure the large contracts she says the fund has seen across most of its investments.
- Whether the fund’s AI-supported diligence process, the Axiom Brain, produces a discernible investment advantage.
Discussion spark: Should AI startups be judged mainly by measurable work they take off customers’ hands, or can broadly useful tools justify investment before they deliver a specific outcome?
Sources and evidence
- This Early Groq Investor Expects Half Her Bets To Fail – Crunchbase News (28 September 2026, 11:00 UTC)
Independent WittyWires tracker for public updates about Groq. Not affiliated with or endorsed by Groq; this is not an official account.