India’s Yotta Data Services says it is preparing to tap public markets “very soon” as demand for AI computing races ahead of available capacity.

India’s artificial-intelligence infrastructure race is entering a new phase, and Yotta Data Services wants Wall Street-style capital to help power it.

The Mumbai-based data-center operator is in talks to access the capital markets “very soon,” according to Chairman Darshan Hiranandani, as the company struggles to keep pace with rapidly rising demand for AI computing. Yotta operates what is described as India’s largest cluster of Nvidia chips and is now preparing for a potential public listing as it expands its data-center footprint and GPU capacity.

The timing is significant.

Around the world, investors have spent the past two years pouring money into the companies building the physical infrastructure behind artificial intelligence. Nvidia has become the clearest beneficiary of that spending, while cloud providers, data-center operators, power companies and specialized “neocloud” businesses have all attracted increasingly large pools of capital.

Yotta is now positioning itself as an Indian answer to that global AI infrastructure boom.

And its growth ambitions are enormous.

The company expects to reach approximately 400 megawatts of capacity over the next 12 to 18 months and says it has the ability to tap as much as $20 billion worth of GPUs. Management says demand is currently growing faster than the company can deploy infrastructure, making delivery schedules and physical expansion the major constraints.

That is an important distinction.

Yotta is not saying customers are unwilling to buy AI computing.

It is saying there is not enough infrastructure available to serve them.

Bloomberg previously reported that Yotta has engaged banks for an IPO that could raise as much as $900 million, while CEO Sunil Gupta has told Dow Jones that the company is targeting a valuation of as much as $6 billion. Hiranandani declined to comment on the valuation or proposed fundraising size in the latest Bloomberg interview.

Those figures would make the offering one of the more interesting AI-infrastructure listings to emerge from India.

But the IPO is not simply about giving early investors an exit.

Yotta has a huge amount of capital to deploy.

AI data centers require expensive processors, high-density cooling systems, networking infrastructure and enormous amounts of electricity. Unlike conventional software startups, these businesses cannot scale simply by hiring additional engineers or selling another subscription.

Every increase in computing capacity requires physical assets.

That makes access to capital markets a strategic advantage.

The company has already demonstrated its appetite for investment.

In February, Yotta announced plans to deploy 20,736 Nvidia Blackwell Ultra GPUs at its Greater Noida facility in one of Asia’s largest AI superclusters. The project involved more than $2 billion of investment, with an initial planned capacity at a 60-megawatt facility that can ultimately scale much further.

The company's appetite does not end there.

Yotta has said it intends to scale beyond 40,000 Nvidia Blackwell GPUs within four months and toward approximately 85,000 GPUs by the end of fiscal 2027.

That would put the company among the largest AI compute operators outside the United States and China.

The Yotta story also fits neatly into India's broader technology ambitions.

India is already one of the world's largest technology and software markets, with more than a billion internet users. But the country has historically been stronger at consuming and developing digital services than owning the physical infrastructure behind the world's most advanced computing.

AI changes that calculation.

Training sophisticated models requires enormous amounts of compute.

Running those models at scale requires even more.

If India wants to become a serious participant in the global AI economy, it cannot rely indefinitely on overseas data centers.

That creates an opening for domestic infrastructure providers such as Yotta.

Hiranandani told Bloomberg that global demand is increasing faster than Yotta's ability to expand, with the main challenge becoming infrastructure deployment and receiving equipment on schedule.

That is exactly the kind of problem investors like to see in a fast-growing infrastructure business—provided the company can turn demand into profitable contracts.

Nvidia is at the center of Yotta's strategy.

The company announced the deployment of more than 20,000 Blackwell Ultra GPUs and a four-year, more than $1 billion engagement with Nvidia connected to establishing a large DGX Cloud cluster in India.

That relationship potentially gives Yotta an important competitive advantage.

Advanced GPUs are difficult to obtain.

Supply is constrained.

Large customers receive priority.

And companies capable of deploying enormous numbers of Nvidia processors quickly can become strategic partners rather than ordinary buyers.

Yotta has been building that relationship since it began work on an Nvidia cloud cluster in early 2024.

The company's ambition is increasingly global rather than purely domestic.

According to Hiranandani, Yotta expects the dollar-denominated share of its sales to rise sharply next year, reaching approximately 80%, compared with only about 20% of revenue coming from dollar-paying customers previously.

That shift suggests Yotta is trying to become a global AI-compute supplier operating from India—not simply an Indian data-center operator.

Selling AI computing to international customers could significantly change Yotta's economics.

Indian companies remain an important customer base, but global AI developers may be willing to sign larger contracts for access to specialized computing.

Dollar-denominated revenue can also provide a natural hedge for a business that purchases expensive technology and infrastructure connected to global supply chains.

The company appears to be seeing increasing demand from AI model builders and inference providers outside India.

That gives Yotta exposure to the broader global AI market.

But it also raises the stakes.

International customers have choices.

They can buy capacity from Microsoft, Amazon, Google and other hyperscalers.

They can use specialist AI clouds such as CoreWeave and Lambda.

They can build their own data centers.

Yotta therefore needs to compete on price, availability, location, power efficiency and the ability to deliver large quantities of computing reliably.

India's scale could become one of Yotta's biggest strengths.

The country has a huge technology workforce and a rapidly expanding digital economy. Government policymakers are also increasingly focused on developing domestic AI infrastructure and sovereign computing capacity.

That combination creates a potentially powerful home market.

The country is not only interested in importing AI services.

It wants AI infrastructure that can support Indian businesses, government agencies and local model development.

Yotta has positioned itself around that sovereign-cloud opportunity while simultaneously pursuing global customers.

This dual strategy could be important.

Domestic contracts can provide stability.

International contracts can provide scale.

The challenge is balancing both without allowing rapid expansion to overwhelm the balance sheet.

That is the part of the Yotta story investors cannot ignore.

An AI cloud provider can generate enormous revenue while requiring equally enormous capital expenditure.

GPUs depreciate.

Technology changes quickly.

Data centers require constant upgrades.

Power systems need expansion.

Cooling infrastructure must support increasingly dense computing environments.

And every new generation of Nvidia processors can change the economics of existing capacity.

That means Yotta's plan to reach 85,000 GPUs is not simply a sales target.

It is a massive capital commitment.

The IPO would provide an important source of funding, but it would not remove the underlying capital requirements.

Investors will therefore want to understand the company's contracts, utilization rates, pricing power, margins and return on invested capital.

The IPO preparations follow a recent $150 million funding round, which valued Yotta at approximately $3.9 billion. The company said the capital would be used to expand AI infrastructure and increase GPU capacity.

That round also gives the prospective IPO an important reference point.

If Yotta seeks a valuation approaching $6 billion, investors will be asked to assign a substantially higher value to the company than its recent private financing implied.

That premium would need to be justified by rapid expansion.

Management appears confident that the demand environment can support it.

The market will eventually decide.

Private investors can tolerate long development cycles and aggressive capital spending.

Public investors are usually less forgiving.

Once Yotta lists, every quarterly report will be scrutinized.

Investors will want to know how quickly capacity is being deployed.

They will ask whether the GPUs are fully utilized.

They will track new customer contracts.

They will examine debt and capital spending.

And they will compare Yotta's economics with those of publicly traded AI infrastructure peers.

That could create pressure for the company to demonstrate not merely growth, but efficient growth.

Yotta is moving toward the public markets at a time when AI infrastructure remains one of the strongest investment themes globally.

Nvidia's latest earnings have reinforced confidence in the continuing AI spending cycle.

Specialized AI cloud providers are raising billions.

Traditional cloud giants are increasing capital expenditure.

And governments are increasingly treating computing capacity as strategic infrastructure.

That environment creates a potentially attractive window for Yotta.

But there is also a danger.

The strongest AI infrastructure companies are becoming very highly valued.

Investors have started asking whether some businesses are simply benefiting from the same capital-spending cycle without possessing durable competitive advantages.

Yotta will therefore need to show why its Indian location, Nvidia access, sovereign-cloud position and growing international business give it something competitors cannot easily replicate.

GPU capacity is only useful if there is enough power to operate it.

That means Yotta's 400-megawatt expansion target is strategically important.

Electricity has become one of the largest bottlenecks in the global AI data-center industry.

As model sizes increase and inference becomes more widespread, computing demand is generating massive new electricity requirements.

The companies that secure power early can gain a significant advantage.

Yotta is therefore competing not just for GPUs and customers.

It is competing for energy.

That may eventually become one of the biggest determinants of the company's valuation.

A successful Yotta listing would send an important signal for India's technology sector.

It would demonstrate that domestic investors and international capital markets are willing to finance large-scale AI infrastructure in India.

That could encourage more data-center construction.

It could attract additional semiconductor and cloud investments.

It could strengthen India's position as a regional AI hub.

And it could encourage other private infrastructure companies to consider public listings.

A weak IPO, on the other hand, could have the opposite effect.

If investors conclude that AI infrastructure is too capital-intensive or valuations are too aggressive, future projects could face a higher cost of capital.

That is why Yotta's eventual offering will be watched beyond its own shareholder base.

The AI infrastructure market is rapidly becoming global.

The United States has Nvidia, hyperscalers and a growing group of specialized cloud companies.

The Middle East is investing heavily in data centers and sovereign AI.

Europe is trying to expand domestic computing capacity.

Asia is competing to attract AI infrastructure investment.

India has a particularly interesting position because it combines a huge domestic market with a deep technology workforce.

Yotta is betting that those advantages can support a world-scale AI cloud business.

Its GPU ambitions suggest the company is willing to make the bet aggressively.

Raising money is only the beginning.

Yotta will have to convert capital into computing capacity.

Then it will have to convert that capacity into customer contracts.

And ultimately those contracts must generate enough cash to justify the massive investment required to build the infrastructure.

That is the central economic test for every AI cloud company.

Demand may be enormous today.

But investors need to know that demand will remain strong after thousands of new GPUs arrive.

Yotta believes it can capture that growth.

Its recent expansion plans certainly suggest confidence.

The company says demand is already greater than its ability to deploy infrastructure, and it is preparing to increase power capacity to approximately 400 megawatts while dramatically expanding its Nvidia GPU footprint.

Now it wants public-market capital to accelerate the process.

If Yotta succeeds, its IPO could become a landmark moment for India's AI ambitions.

The company would not simply be selling shares.

It would be offering investors a stake in the physical infrastructure behind India's attempt to become an AI power.

And in a market where Nvidia has shown that AI computing demand can support astonishing levels of spending, Yotta is betting that the next big opportunity is not making the chips.

It is owning the machines that put those chips to work.

Source basis: Bloomberg/Yahoo Finance reporting and current company disclosures on Yotta's IPO plans, GPU expansion, financing and data-center strategy.