Axiom Partners’ Sandhya Venkatachalam Plans to Let Half Her AI Bets Fail

Sandhya Venkatachalam, a veteran of both hardware and software startups, has launched Axiom Partners with a clear-cut thesis: back artificial-intelligence companies that deliver tangible outcomes in industries such as construction, industrial manufacturing and insurance. The $52 million fund aims to make about 35 investments and is built around the expectation that roughly half of those companies will not succeed. Venkatachalam’s approach, shaped by early bets on AI chipmaker Groq, emphasizes early-stage risk assessment, founder diversity and real-world impact over conventional enterprise software metrics.

Venkatachalam’s career began in product leadership roles at a data-center hardware firm that was later acquired by Cisco, followed by a stint as a product executive at Skype before its sale to Microsoft. Those experiences gave her exposure to data and machine learning long before AI became a dominant venture-capital theme. She later joined Social Capital as a general partner, where she led the firm’s first institutional investment in Groq, an AI inference chip company. The Groq investment, made in 2016 when inference was still a niche concept, taught her the value of being slightly ahead of the market and exercising patience while a technology category matures.

At Axiom, Venkatachalam has translated those lessons into a structured investment model. The firm’s team is composed of active AI practitioners who split their time between their primary roles and Axiom, receiving carry in the fund. This part-time partnership model ensures that the investors stay close to the market and can contribute current technical insight during diligence. The firm also relies on an internal tool called the “Axiom Brain,” an AI system that scans market trends, identifies promising founders and accelerates the due-diligence process. The goal is to move quickly on opportunities that align with the fund’s focus on outcomes rather than software features.

The “AI for the real world” mantra guides Axiom’s screening criteria. Venkatachalam looks for startups that use AI to produce a concrete result for customers, especially in sectors that have been underserved by technology. She prefers companies that can command contract values in the hundreds of thousands of dollars, even at early prototype stages, indicating that customers are willing to allocate labor budgets to AI-driven solutions. Durability, in her view, comes from deep integration with a client’s data, workflows and operational systems, what she calls the “last mile” of the job. Such embedded relationships are difficult for competitors or large model providers to replicate.

While the fund’s ambition is to back companies that can become “digital workers” replacing traditional software tools, Venkatachalam acknowledges the inherent risk. The plan to make 35 bets with an expectation that half will fail reflects a deliberate tolerance for loss, balanced by the pursuit of outlier successes that could reshape entire industries. As the AI landscape continues to evolve, Venkatachalam’s strategy hinges on spotting the infrastructure and application layers that will be built on top of foundational models before the market reaches consensus. The coming years will reveal whether Axiom’s early-stage, outcome-focused bets can translate into the next generation of durable AI enterprises.