Startup seed rounds double as VCs turn more selective

India’s seed-and early-stage startups are raising almost twice as much per round as they did a year ago, with venture firms frontloading capital for more mature products and business models that require upfront spending, particularly in artificial intelligence, deeptech, infrastructure and healthtech.

According to data collected from Tracxn’s India Tech H1 2026 report, seed and early-stage startups raised a combined $3.34 billion across 608 rounds in the first half of 2026, compared with $2.96 billion across 1,055 rounds a year earlier. That pushed the average cheque size across these two buckets to about $5.5 million, from $2.8 million.

Early-stage led the shift, with funding rising to $2.8 billion from $2.2 billion even as the round count fell to 188 from 258. At seed-stage, total capital fell to $541 million from $758 million, but number of rounds dropped much faster, to 420 from 797, lifting the average cheque to about $1.3 million from less than $1 million.

Also Read: ETtech Deals Digest: Cred’s mega round lifts startup funding to $1.09 billion this week, up 290% on-year

Shift in VC market

“Earlier, people would come up with an idea, raise funding, hire a team and then build the idea out,” said Siddarth Pai, founding partner, 3one4 Capital. “The minimum viable products—the first usable versions of a product—coming to market now are a lot more mature, courtesy of AI. Even non-technical people can use AI to create a prototype, get some traction and then raise funds.”

The shift reflects a venture market where funding has become harder to access, but startups that clear diligence are being capitalised more heavily at the outset. Larger VC funds, tighter portfolio construction, mature products reaching investors and the rise of capital-intensive sectors such as AI, deeptech, infrastructure and healthtech are changing what seed and early-stage rounds look like, investors said.

Aggregate funding

Investors said larger early-stage cheques should not be read as a return to the easy-money years. Overall funding in Indian tech rose 12% year-on-year to $7.2 billion in H1 2026 from $6.4 billion a year earlier, but the number of funding rounds fell 43% to 652 from 1,149, according to Tracxn.

The number of companies receiving their first recorded funding round also fell 31% to 218, from 317 a year earlier.

Late-stage funding rose marginally to $3.8 billion from about $3.5 billion a year earlier, but the number of late-stage rounds dropped to 44 from 94. The data shows that growth capital has not broadly returned and remains concentrated in fewer, larger transactions.

Anand Lunia, founding partner at India Quotient, said the fall in deal activity is being felt more sharply among smaller early-stage investors than larger venture firms.

“Funds of our size or bigger ones are doing more or less the same amount of activity,” he said. “But angel syndicates are fewer and smaller firms are tighter or doing more shared deals."

That is one reason the average cheque size is rising. As smaller rounds reduce and larger institutional seed and early-stage cheques form a bigger share of the market, the average moves up even though the number of funded companies falls.

Lunia said seed rounds have also become larger because founders are preparing for a tougher path to the next round. “Median seed deal size has also gone up in response to the toughness of Series A,” he said.

Nishit Garg, Partner at venture capital fund, RTP Global, said the funding market has changed sharply from the 2021-22 cycle.

“You cannot raise a $20-25 million round on the same metrics that you could raise it in 2021-22,” Garg said. “The market cycle has changed. It has to be on performance today; in 2021-22, it could still be on hope.”

AI is also changing how young companies are built before they approach investors. Smaller teams can now build prototypes, test workflows and show early traction with fewer people, even as the cost structure of startups changes.

“Today, even when evaluating consumer companies, one of the first things we look at is whether the founder is AI-enabled or AI-first,” Garg said. “The cost structure of these companies is changing. Earlier it was about people; today it is about people and tokens.”

For founders, this has meant longer preparation before fundraising. A founder of a biotech company that recently raised seed funding said investors were interested in the company’s technology and vision, but the round took time because the company had to show the commercial path as well.

“When we first presented the technology and the larger vision, investors were interested. But the fundraising process still took time because we had to go beyond the science and show the commercial side of the business—how the product would be deployed, who would pay for it, and what the path to market would look like,” the founder said.

“Because the technology is novel, there is also an education process involved. It takes time for VCs to understand the science, the use case and the milestones we are working towards. So the scrutiny is naturally higher, especially when the company is still at an early stage,” the founder added.

The deeptech driver

The other driver is the changing sector mix. AI infrastructure, spacetech, semiconductors, battery technologies, advanced materials and healthtech do not follow the same funding curve as consumer internet or software-as-a-service companies. They require money for specialised talent, research, compute, hardware, pilots and technical validation before conventional revenue metrics become visible.

This has made early-stage capital more important for companies building in frontier areas, where a small seed round may not be enough to reach the next meaningful milestone. Tracxn’s H1 data shows large Series B rounds such as Neysa’s $600 million, Sarvam’s $234 million and Skyroot’s $60 million, underscoring how the early-stage bucket is now being shaped by companies underwriting platform, compute or technical risk rather than just product-market fit.

Vinod Sankar, partner at Java Capital, which specialises in deep-tech technology funding, said the number of deeptech companies being formed has risen significantly across areas such as semiconductors, battery management, space and nanoparticles.

“We are seeing a massive number of deeptech companies now. The numbers were much lower earlier,” Sankar said. “There is quantity. The question is whether these companies can build something scalable and show real outcomes.”

For investors, this requires a different underwriting lens. In consumer businesses, early demand, retention and transaction behaviour can show up quickly. In SaaS, investors have established benchmarks around revenue growth, churn and sales efficiency. In deeptech, early proof is often technical progress, not revenue.

“In deeptech, revenue comes late and scale also comes late. That is the nature of the industry,” Sankar said. “At the early stage, I look for technical milestones first, not revenue milestones. Once the technical milestones are met, revenue follows.”

That is why early-stage cheque sizes are rising even when the broader venture market remains cautious. Startups that fit the new investor template—mature products, technical validation, specialised teams and capital-heavy sectors—are being funded more aggressively upfront. But the fall in deal count shows the benefit is not spreading evenly across the ecosystem.