HCLTech is likely to lead a major funding round in Bengaluru-based Sarvam AI, in what could become one of the most closely watched technology deals in India’s startup ecosystem this year.
The planned round is expected to be worth $300 million, with HCLTech likely to invest around $150 million. Bessemer Venture Partners is expected to put in about $50 million, while the remaining $100 million may come from Nvidia, Prosperity7, Activate, Glade Brook and other investors, including existing backers.
If completed, the round would value Sarvam AI at about $1.5 billion post-money. That would mark a sharp rise from its 2023 valuation and place the company among the most valuable homegrown artificial intelligence startups in India.
For HCLTech, the move is not just another startup investment. It comes at a time when India’s traditional IT services industry is under growing pressure to prove that it can adapt to the AI shift rather than be disrupted by it.
For Sarvam, the deal could give it the financial muscle to scale its work in foundational AI models, especially at a moment when India is trying to build more local alternatives to AI systems dominated by global technology companies.
The round has not been formally closed yet. Sarvam AI, HCLTech, Bessemer Venture Partners, Prosperity7, Glade Brook, Activate and Nvidia did not respond to queries mentioned in the original report.
India’s IT services companies have spent decades building large global businesses around software development, maintenance, outsourcing and back-office work. But the rise of generative AI has unsettled that model.
As AI tools become better at coding, automation and enterprise workflows, investors have started asking whether large IT companies can move fast enough. The worry is simple: if AI begins to reduce the need for some traditional outsourcing work, Indian IT firms will need deeper AI capabilities, not just AI branding.
That is what makes HCLTech’s likely investment in Sarvam AI important. It would be a rare case of a major Indian IT services company placing a large bet on a domestic AI startup working on foundational models.
For years, India’s biggest IT firms have often been criticised for being conservative with frontier technology investments, preferring buybacks, dividends and client-led digital transformation work over aggressive research-led bets. A $150 million investment in Sarvam would be a different kind of move.
It would also show that the AI race in India is no longer only about startups raising venture capital. The country’s older technology giants may now have to decide whether they want to build, partner, invest, or risk falling behind.
Sarvam’s rise in India’s AI race
Sarvam AI was founded in July 2023 by Vivek Raghavan and Pratyush Kumar. Both had previously worked at AI4Bharat, an initiative backed by Infosys co-founder Nandan Nilekani.
The company has positioned itself as a full-stack generative AI player. Its work ranges from research and training custom AI models to building enterprise-grade platforms for deployment and authoring.
Sarvam last raised $41 million in a Series A round in December 2023. That round was led by Lightspeed Ventures, with participation from Peak XV Partners and Khosla Ventures. Existing investors such as Khosla Ventures, Lightspeed and Peak XV Partners are expected to participate in the new round on a pro rata basis.
Investor interest in Sarvam appears to have grown after the company launched two large language models earlier this year at the India AI Impact Summit in New Delhi. The models include a 30-billion-parameter model and a 105-billion-parameter model, both trained from scratch in India.
Sarvam co-founder Pratyush Kumar had earlier said the models were cheaper than something like Gemini Flash while outperforming it on many benchmarks. He also said Sarvam’s models showed stronger Indian language performance even when compared with Gemini 2.5 Flash.
Those claims, if consistently reflected in real-world use, would be especially important in India, where language diversity remains one of the biggest barriers to mass AI adoption.
The sovereign AI angle
Sarvam’s story is closely tied to India’s sovereign AI ambitions.
The IndiaAI Mission has been backed by a Rs 10,000 crore fund to support India’s push toward sovereign foundational models. So far, Rs 111 crore has been disbursed in GPU subsidies.
Sarvam has emerged as the biggest beneficiary of that support to date. The Bengaluru-based startup received access to 4,096 NVIDIA H100 SXM GPUs through Yotta Data Services and received nearly Rs 99 crore in subsidies.
That matters because foundational AI models are expensive to build. They require high-quality data, specialised talent and enormous compute capacity. For Indian startups, access to GPUs can often decide whether they remain research-heavy companies or become serious model builders.
Sarvam is not alone in this race. Conversational AI firm Gnani.ai and IIT Bombay-led BharatGen are also working on foundational AI models in India. But Sarvam’s fundraising plans, GPU access and recent model launches have pushed it into the centre of the country’s AI debate.
The funding discussions have gone through a few changes. Earlier, Sarvam AI was in talks to raise $250 million from HCLTech, Nvidia, Accel and others. Later, Accel passed on the round, while Bessemer Venture Partners became actively engaged.
Bessemer’s US Partner Sameer Dholakia was leading the discussions, based on the reported details.
With Bessemer now expected to invest about $50 million, the round appears to have expanded from the earlier $250 million plan to $300 million.
If the round closes at the current size, it would be Sarvam’s first major cash infusion in about two-and-a-half years.
A bigger message for Indian IT
The biggest story here may not be the valuation alone. It is the timing.
Indian IT companies are facing a difficult question: can they remain service providers in an AI-led world, or do they need to become deeper technology owners?
HCLTech’s likely investment in Sarvam suggests one possible answer. Instead of only selling AI services to global clients, Indian IT firms may need stakes in companies building the models and platforms that could shape the next phase of enterprise technology.
This does not mean the deal is risk-free. Foundational AI is capital-intensive. The market is crowded. Global companies have far larger resources. Model performance claims need to survive enterprise adoption, cost pressures and real-world scaling.
But for India, the deal could still carry strategic weight. A large domestic IT services company backing a homegrown AI model startup would give India’s AI ecosystem more credibility, more capital and possibly more enterprise access.
For Sarvam, the money could help it move from promise to scale. For HCLTech, the investment could become a statement that it does not want to watch the AI transition from the sidelines.
And for India’s technology sector, it could mark a shift from asking whether the country can build serious AI companies to asking how quickly those companies can be funded, tested and taken to market.
