The news
Sandhya Venkatachalam, the Groq backer who led early institutional investments in the AI chipmaker, told Crunchbase News on September 28 that her firm, Axiom Partners, expects about half of its startups to fail. It has a $52 million fund and plans 35 investments.
Venkatachalam is the founder and managing partner of Axiom Partners. According to the Crunchbase News interview, she previously led product at a data center hardware company that Cisco acquired, worked as a product executive at Skype before its sale to Microsoft, served as a general partner at Social Capital, where she backed Groq, and later invested at Khosla Ventures.
Her explanation for the high expected loss rate is simple venture math. She told Crunchbase News that one outstanding investment that grows very large can return the whole fund, and that accepting losses is what lets the firm invest before an opportunity is obvious. She traced that mindset to Groq: she said co-founder Jonathan Ross convinced her that inference, the work of running trained AI models, would be a far bigger market than training, at a time when, in her words, she "barely understood inference."
Axiom's thesis is what she calls AI for the real world: software that delivers measurable outcomes in construction, industrials and insurance, rather than another enterprise SaaS tool. She said such contracts often reach hundreds of thousands of dollars and are paid from labor budgets, not software budgets. She argued the defensible part is handling the last mile of the job, which requires deep integration with a customer's systems, data and workflows.
The firm also runs an unusual bench. Venkatachalam said Axiom works with part-time AI practitioners who keep their day jobs and receive carried interest, a share of fund profits, and uses an internal system it calls the Axiom Brain for market analysis and faster due diligence. She said she looks for nonobvious founders in nonobvious industries, and criticized Silicon Valley's habit of favoring founders with Stanford computer science degrees or OpenAI experience.
The interview landed the same day TechCrunch reported that Peak XV, the firm formerly known as Sequoia Capital India and Southeast Asia, raised the ceiling on its Surge seed program to up to $5 million per company from $3 million. Peak XV managing director Rajan Anandan told TechCrunch that the bar to raise a Series A has gone up significantly.
The numbers
- Axiom Partners fund size
- $52 million
- Planned Axiom investments
- 35
- Share Axiom expects to fail
- About half
- New Peak XV Surge maximum check
- Up to $5 million (from $3 million)
- Surge 12 cohort
- 18 startups, $50 million+ invested by Peak XV
Why CEOs should care
For founders, the two data points point the same way. Bigger seed checks buy more runway, but TechCrunch's reporting and Anandan's comments suggest investors now expect more progress before a Series A. A founder raising seed should ask each investor what milestones they believe the next round will require, and plan the budget backward from that answer rather than from a fixed number of months.
For corporate leaders buying from startups, a fund that plans for half its companies to fail is a reminder to check vendor durability. If your team signs an AI contract worth hundreds of thousands of dollars with a young company, ask about its runway, who its investors are, what happens to your data and integrations if it shuts down, and whether source code escrow or data export terms are in the contract.
For boards and limited partners, Axiom's model shows how some small funds are trying to compete: sector focus, outside practitioners paid through carry, and internal analysis tools. Those are fair topics for diligence. Ask how part-time advisers handle conflicts with their day jobs, and how a fund that expects heavy losses reserves money for follow-on rounds in its winners.
The bigger picture
Venkatachalam's comments fit a broader pattern in early-stage AI investing, where seed programs such as Peak XV's Surge are adjusting check sizes upward. According to TechCrunch, the Surge program, launched in 2019, has backed more than 180 companies, and the top 10 generate more than $1 billion in combined annual revenue.
Her pitch against generic SaaS also reflects a debate among investors about where AI startups can build lasting businesses. Selling outcomes into industries such as construction and insurance, and getting paid from labor budgets, is one answer. It depends on startups doing slow integration work that larger model providers may not want to do.
What’s next
Watch whether other seed programs follow Peak XV in raising check sizes, and whether Series A investors publish clearer milestones. For Axiom, the test is time: with 35 planned investments and about half expected to fail, the fund's result will depend on whether one or two companies grow large enough to return it.
What “Fact-checked” means
Fact-checking means testing a story’s facts against the evidence before it is published. This story went through at least two separate checks before this version was published.
- What we checked
- Its names, figures, dates, job titles, quotes and who said what were checked against the story’s sources, including its main source where it could be opened. The headline was checked for accuracy and overstatement.
- How
- A first check reviewed the whole story. If it passed, a second, skeptical check went back to the sources to look for mistakes in the most important facts. If a check flagged the story, it was edited to fix the problems found, and a separate re-check then reviewed the whole story again.
- Who
- The checks are made by our newsroom, as steps kept separate from the writing, under rules set by our editor, Hussein Mukhtar. A story the checks still flag is held for the editor, who decides whether it is fixed, published or dropped.
- If something is wrong
- “Fact-checked” does not mean error-free. If a material error is found after publication, we correct the story and add a note saying what changed. Report an error
