Not Another Dot Com Burst: How Are VCs Investing in AI? The discussion unpacked the current state of AI investing, the risks of overvaluation, and what it will take for AI startups to succeed in a rapidly evolving market.

AI has become the dominant theme in venture capital, attracting significant investment and raising questions about whether it's the next dot-com bubble or a lasting shift in technology. Rahul Agarwalla, managing partner at SenseAI, wasted no time addressing the elephant in the room. "Last year alone, about $500 billion was invested in AI. Of that, $100 billion went into venture funding, but the real story is the $400 billion going into building AI data centers," he noted. This level of capital deployment, Agarwalla argued, signals that AI is not just a passing trend.

Som Pal Choudhury, co-founder & partner at Bharat Innovation Fund agreed, highlighting how AI's rapid technological advancement distinguishes it from previous tech booms. "When ChatGPT came out in 2022, it was orders of magnitude faster and more cohesive than anything we had seen before," he said. "The last two years, frankly speaking, none of us have been able to even keep up with the pace."

Sheetal Bahl, partner at Merak Ventures drew parallels to the dot-com era but underscored why AI is different. "I think the parallel to the dot com boom is the fact that for a really long time, people didn't realize the value of something, and then they were undervaluing it, and suddenly they woke up. Something happens to bring that technology into the zeitgeist."

Investment strategy

The panelists were unanimous in stressing that AI startups need to offer more than just a thin layer of value. Agarwalla was blunt: "It can't be that you just consume OpenAI's APIs and slap a user interface on top. That's not going to cut it." SenseAI follows a structured framework called VDAT (Value, Data, Algorithms, Technology) to evaluate startups. "We want to fund the builders, not just the users of AI," he said.

Choudhury explained how Bharat Innovation Fund assesses opportunities across the AI stack. "We see three key layers: foundational models and hardware, middleware, and applications," he said. "India's strength is at the application layer—applying GenAI and agentic AI to specific industry problems like cybersecurity and sales automation. But we're also seeing promising middleware companies, especially in red-teaming AI models and fine-tuning pipelines."

Bahl took a more nuanced approach, preferring to invest in what he called "iceberg startups." "The part of the business you see—the application—is just the tip. The real value lies in proprietary data, algorithmic improvements, and defensibility across the stack," he explained.

Patience capital

While the potential for AI-driven returns is enormous, the panelists acknowledged the need for patient capital and realistic expectations. "Deep tech investments take time," Bahl said. "We've invested in space tech, EVs, and semiconductors. Those businesses can take five to seven years to generate meaningful revenue."

Choudhury agreed but highlighted that AI startups tend to scale faster than other deep tech sectors. "Getting to half a million or a million in revenue is possible within 18 to 24 months," he said. "The real challenge is scaling beyond that—especially from India, where the enterprise market is still immature. The (Silicon) Valley has mastered that playbook, but we're still figuring it out."

Agarwalla noted that the revenue ramp-up for AI-first companies can be surprisingly fast if the product-market fit is strong. "We're seeing startups go from zero to half a million in four months," he said. "But they need to work with real enterprise customers early. AI doesn't get built in an ivory tower—you need customer data and real-world feedback to fine-tune the product."

Surviving the future

All three panelists agreed that AI's market shakeout is inevitable. "Nine out of ten AI startups will fail," Bahl predicted. "Big tech can destroy 100 startups with a single product update. Surviving that requires more than just a great model—it requires proprietary data, defensible IP, and a clear value proposition."

Choudhury highlighted the importance of adaptability. "You need to constantly evaluate whether your AI layer is defensible. If it's too easy to replicate, the big players will roll you over," he said.

Agarwalla pointed to the importance of business fundamentals. "Unless you can monetize quickly and sustainably, you're not going to survive," he said. "AI's pace of change is brutal. The winners will be those who can translate technological advantages into sustainable business models."

The AI gold rush

As the discussion wrapped up, the moderator, Punita Sabharwal, managing editor of Entrepreneur India, noted that while AI investing is clearly hot, the underlying message was one of discipline and focus. Capital may be flowing into AI at unprecedented rates, but the investors leading the charge are proceeding with caution.

The panellists were speaking at the Entrepreneur India Tech & Innovation Summit. The discussion unpacked the current state of AI investing, the risks of overvaluation, and what it will take for AI startups to succeed in a rapidly evolving market.