Sarvam AI and the Big Question: Can India Pay to Build Frontier AI?
Sarvam AI is one of India’s brightest young tech companies. It is a startup (a new, fast-growing company) based in Bengaluru, India. The government picked it under the IndiaAI Mission. That is the government’s national plan to build India’s own AI tools. Sarvam’s job is to build a “foundational AI model” for India. A foundational model is a huge AI system. It is trained on tons of text and data. Other apps can then be built on top of it. Sarvam AI is growing fast. But its next big test is not about technology. It is about money.
The truth is simple. Building frontier AI costs a lot. Frontier AI means the most advanced, cutting-edge AI models in the world. It needs a lot of money. It also needs a lot of computer power. India now has the smart people. The big question is this: can India pay the bill?
What Sarvam AI is and why it matters
Sarvam AI started in August 2023. Two people founded it: Vivek Raghavan and Pratyush Kumar. In a short time, it became a big name in Indian AI. It builds AI models that understand Indian languages. It also builds AI agents. An AI agent is a software helper. It can talk to people and do tasks on its own.
The company is already used a lot every day. It says people have more than 2 million chats with its AI each day. That is about 10 million (1 crore) API calls a day. An API call is just one request an app makes to use the AI. Almost 80% of the money it earns today comes from these chat AI agents.
The IndiaAI Mission chose Sarvam to build a “sovereign” AI model. Sovereign AI means AI that a country owns and controls by itself. The country does not have to fully depend on foreign companies. For India, this is a big deal. It is about keeping data safe. It is about national pride. And it is about Indian languages. All three at once.
The HCLTech bet: is “patient capital” finally here?
The big recent news is about funding. Funding means the money investors give a company. Sarvam AI raised $234 million in one round. This pushed its valuation up to $1.5 billion. Valuation is the total worth of the company, as people guess it. That makes Sarvam a “unicorn.” A unicorn is a startup worth a billion dollars or more. Sarvam is India’s second AI unicorn. The first was Ola’s Krutrim.
The main backer is HCLTech. It is a large Indian IT services company. HCLTech promised $150 million. People are calling this a turning point. Why? Because of an idea called “patient capital.” Patient capital is money that is happy to wait many years to get returns. It does not ask for quick profits.
Frontier AI needs exactly this kind of money. Training a big model can take years. It can burn lots of cash before it earns much back. Many investors get scared by such long, risky bets. So a big Indian company giving long-term money is seen as good news. People say patient capital is finally coming to Indian deeptech. Deeptech means advanced, science-heavy technology that takes a long time to build.
Some people compare this to Microsoft backing OpenAI early on. But the size here is much smaller. It is a sign of belief, not a match in dollars.
Key facts at a glance
| Item | Detail |
|---|---|
| Company | Sarvam AI |
| Base | Bengaluru, India |
| Founded | August 2023 |
| Founders | Vivek Raghavan, Pratyush Kumar |
| Latest funding round | $234 million |
| Lead backer (HCLTech) | $150 million committed |
| Valuation | $1.5 billion (unicorn) |
| Annual revenue (ARR) | About $12 million |
| Daily AI interactions | 2 million+ |
| Daily API calls | ~10 million (1 crore) |
| National role | Chosen under IndiaAI Mission for sovereign model |
The real challenge: scaling India AI funding
Here is the tricky part. The $234 million round is big for India. But it is small in the global AI race. The world’s top AI labs raise billions of dollars at a time. Training the largest models can cost huge sums just for compute. Compute means the raw computer power and chips used to train AI. These are mostly powerful GPUs, a kind of strong computer chip.
Right now, Sarvam AI earns about $12 million in ARR. ARR stands for annual recurring revenue. That is the steady yearly money a company expects from its customers. This income is healthy and growing. But it is tiny next to the cost of training frontier models again and again.
So the gap is clear. India has the smart people. It now has its first wave of serious money. But it must prove one thing. It must keep feeding these models money and compute, round after round, for many years.
Where the new money will go
The new money will push three things. First, frontier model research. This includes agentic AI, coding, and cybersecurity. Agentic AI means AI that can act and finish tasks by itself. Second, large-scale compute. That means the chips and servers needed to train models. Third, wider use across businesses and government offices.
Why it matters (especially for India and founders)
For India, Sarvam AI is more than one startup. It is a test of the AI sovereignty dream. India wants its own models. It wants models that understand its languages and protect its data. But someone has to pay for them. Sarvam AI is the live test of whether that can happen.
For founders, the lesson is clear. Talent and clever products are not enough in frontier AI. You also need investors who can stay calm through years of heavy spending. The HCLTech deal shows such patient backers may finally be coming to India.
For the wider tech world, this round sends a signal. A homegrown IT giant just funded a homegrown AI lab. Others may now follow. That could open up a chain of long-term money for Indian deeptech.
Frequently asked questions
What is Sarvam AI?
Sarvam AI is a Bengaluru AI startup. It started in 2023. It builds AI models and AI agents made for Indian languages and needs. The IndiaAI Mission chose it to build a sovereign Indian foundational model.
How much did Sarvam AI raise, and who led it?
Sarvam AI raised a $234 million round. This was at a $1.5 billion valuation. HCLTech promised $150 million. That made it the main backer in the deal.
What does “patient capital” mean here?
Patient capital is money that is willing to wait many years for returns. Frontier AI needs this. Models cost a lot and take a long time to pay off. The HCLTech bet is seen as patient capital coming to Indian deeptech.
Why is funding the next big challenge for Sarvam AI?
Building frontier AI needs huge money and compute, again and again. Global labs raise billions. India’s biggest AI rounds are much smaller. So scaling India AI funding is the key hurdle ahead.
The takeaway
Sarvam AI shows that India can build serious AI. The HCLTech backing shows that long-term money is starting to appear. But the funding gap with global labs is still wide. The next chapter of India AI funding will decide one thing. It will show if the sovereign AI dream can truly grow big.
For more on India’s startup and tech funding moves, see our coverage of the India IPO watch on Zypp and Turtlemint and the latest global AI deals involving Getty, DeepMind and Samsung.
Sources and further reading: Financial Express and Inc42.
