Early Groq Backer Anticipates Half Of Her Investments Will Fail

Sandhya Venkatachalam spent the initial phase of her career developing technology enterprises. She managed product development at an early data center hardware firm acquired by Cisco, then served as a product executive at Skype prior to its acquisition by Microsoft. These positions exposed her to data and machine learning long before artificial intelligence dominated venture capital.

She subsequently worked as a general partner at Social Capital, where she spearheaded early institutional investments in AI chip manufacturer Groq, and later invested with Khosla Ventures. Today, she operates as the founder and managing partner of Axiom Partners, a $52 million fund that supports startups utilizing AI to handle labor in sectors like construction, industrials, and insurance.

During an interview with Crunchbase News, Venkatachalam elaborates on why she looks beyond traditional founder backgrounds, what creates lasting durability in an AI enterprise, and how her early investment in Groq influenced her strategy.

This interview has been edited for clarity and brevity.

You invested at Khosla Ventures before starting Axiom. What did you take from that experience, and what did you want to do differently?

Venkatachalam: One lesson I absorbed was Vinod Khosla’s expansive perspective regarding the origins of exceptional founders. Silicon Valley has typically favored a restricted definition of who can establish the next premier AI enterprise: someone possessing a Stanford computer science or machine learning pedigree, or tenure at OpenAI. We actively search for nonobvious founders, particularly within nonobvious industries.

Another takeaway involved our framework for risk assessment. Instead of subjecting a company to every conceivable due diligence inquiry, we concentrate strictly on the risks that dictate its next set of milestones. Can this team execute its stated objectives? And if so, could the ultimate impact be monumental?

This approach requires acknowledging that numerous bets will falter while keeping our sights set on outliers. I believe my investors back me to identify tomorrow’s categories rather than participate in the familiar ones everyone already recognizes.

At Axiom, the fundamental distinction is that we structured the firm around individuals who actively work with AI. Without regular involvement in building, productizing, pricing, or bringing AI to market, maintaining relevance is exceedingly difficult. Our roster includes practitioners who perform those exact duties in their separate professional capacities. They keep our investment perspective current, and founders appreciate working with them because they have encountered parallel challenges.

We also integrate AI across all internal operations. We created a system termed the Axiom Brain to interpret market trends, discover compelling individuals and companies, and accelerate due diligence alongside other workflows. For me, the primary benefit lies in our capacity for rapid action.

You mentioned that some of those AI practitioners have other jobs. How does their role at Axiom work?

Venkatachalam: They dedicate specific time to Axiom and collaborate with us on a part-time basis. They also receive carried interest in the fund. They function as active partners in the endeavor rather than names listed on an advisory roster.

Their external employment is vital to the model. Certain top-tier angel investors remain active operators and builders. I do not require their full-time attention. In truth, they would offer less value to Axiom if they abandoned the hands-on work that keeps them close to the marketplace.

Axiom says it invests in “AI for the real world.” What does that mean when you’re evaluating a startup?

Venkatachalam: Our perspective is that AI should serve a significantly wider demographic than the early adopters presently utilizing it. We examine industries historically underserved by technology, where AI can deliver tangible outcomes rather than merely supplying another software tool.

This pursuit may direct us toward construction, industrials, or insurance. Certain targets incorporate hardware, sensors, or robotics, while others rely entirely on software. The unifying thread is that they execute vital work for customers operating in real-world sectors.

We generally avoid products resembling conventional enterprise software tools. We prefer observing AI directly generate measurable results.

You’ve described a shift from software people use to digital workers that perform jobs. Are customers actually paying for AI from labor budgets?

Venkatachalam: Yes, and that represents a core investment criterion. Even when a portfolio company operates at an alpha or design-partner stage, we perform diligence to verify whether customers are willing to purchase the solution under those terms. We frequently observe contract values reaching hundreds of thousands of dollars, eclipsing the modest contracts typically associated with midmarket software tools.

This purchasing behavior has manifested across the majority of our portfolio companies.

AI products are becoming faster to build and easier to imitate. What makes one durable enough to become a large company?

Venkatachalam: When you perform essential labor within a customer’s organization, and that labor carries high financial value, you become difficult to replace. You manage what we designate as the last mile of the job.

Within industrial environments, for instance, delivering an outcome requires deep integration with customer systems. You must comprehend their data, train models upon it, master critical workflows, and stand behind the final output. Accomplishing this demands far more than applying a user interface on top of a foundational model.

Such relationships and capabilities prove difficult for competing startups to replicate. They also encompass tasks that major AI model providers may lack interest in pursuing independently.

Before Axiom, you backed Groq when AI inference was far from an obvious investment category. What led you to it?

Venkatachalam: My background spans both hardware and software. At one point, I grew curious about why Google manufactured its proprietary networking switches instead of purchasing them from established vendors. Investigation revealed that the company was also building its own silicon processors.

This inquiry connected me with Jonathan Ross, who had participated in that initiative before departing to establish Groq. I began exploring why major technology firms developed chips specifically for model training. Subsequently, Jonathan argued that the vastly larger future market would center on inference.

To be transparent: in 2016, I barely grasped inference. However, if one believed these models would proliferate, it followed logically that people would build atop them and require supporting infrastructure. That realization drove my investment decision.

How did that experience shape what you look for now?

Venkatachalam: It demonstrated the value of arriving slightly early and exercising patience. You do not need to be wildly contrarian, but you must identify an opportunity before it becomes apparent to everyone else.

In a sense, our overarching thesis remains unchanged. We continue to ask what will be constructed on top of AI infrastructure and models. We seek to invest while those answers are still forming, prior to the establishment of consensus.

What happens when one of those early bets doesn’t work out?

Venkatachalam: We anticipate that outcome. Managing a $52 million fund involves executing roughly 35 investments, and we fully expect approximately half of them to fail—whether through outright closure or simply falling short of our growth expectations.

This model relies entirely on capturing an exceptional outlier. We require one outstanding investment to return the entire fund. Investing early enough to catch a company that expands massively can offset numerous unsuccessful bets.

That willingness to absorb losses is fundamental to investing before an opportunity becomes obvious. It is thoroughly integrated into our fund strategy.

Related Crunchbase query:

• Global Venture Funding To AI Startups In 2026

Frequently Asked Questions

What is Axiom Partners?

Axiom Partners is a $52 million venture fund founded by Sandhya Venkatachalam that backs startups utilizing AI to handle labor in industries like construction, industrials, and insurance.

What type of founders does Axiom Partners look for?

The firm seeks nonobvious founders, particularly within nonobvious industries, moving beyond the traditional Silicon Valley profile of Stanford computer science graduates or OpenAI alumni.

What is Axiom’s investment thesis regarding AI?

Axiom focuses on “AI for the real world,” backing companies that deliver measurable operational outcomes and act as digital workers rather than standard software tools.

What was Sandhya Venkatachalam’s connection to Groq?

While serving as a general partner at Social Capital, Venkatachalam led early institutional investments in AI chipmaker Groq during 2016 when inference was not yet an obvious investment category.

What failure rate does Axiom expect across its portfolio?

The fund plans for about half of its 35 investments to fail or miss growth targets, relying on exceptional outlier outcomes to return the fund.