India's Biggest "Domestic AI Infrastructure" Company Is Now Majority-Owned by a US Private Equity Firm
Neysa went from $50 million raised to a $1.2 billion deal in three years. The fine print: Blackstone now owns the majority stake, not the founders.

In February, Blackstone announced it would lead a capital raise enabling Neysa, a Mumbai-headquartered AI infrastructure startup, to secure up to $1.2 billion, split between roughly $600 million in equity and a further $600 million in planned debt financing. The number alone would be a notable Indian AI story in any year. The detail that actually matters sits one paragraph into every announcement of the deal, easy to read past if you're only scanning for the total: Blackstone and its co-investors are acquiring a majority stake. Neysa, founded in 2023 by Sharad Sanghi and Anindya Das on the strength of roughly $50 million in prior funding, is no longer majority-controlled by the people who started it.
That's not a criticism of the founders' decision. It's worth being precise about what actually happened, because the framing every press release used, "India's leading AI infrastructure platform," "helping to enable the country's AI revolution," is doing quiet rhetorical work that the ownership structure doesn't support. This is a genuinely good outcome for Sanghi and Das personally, and probably the right call given what building at this scale actually requires. It is not, in any meaningful sense, India retaining control of its own AI compute layer. It's a US private equity firm acquiring control of it, with an Indian founding team staying on to operate it.
What Neysa actually does, and why the money is this size
Neysa builds and operates GPU-based AI infrastructure, cloud compute purpose-built for enterprises and government entities to train, fine-tune, and deploy AI models inside India rather than renting capacity from a foreign hyperscaler. Its customers span financial services, technology, healthcare, and public services. The new funding is earmarked specifically for deploying more than 20,000 GPUs, a scale of hardware commitment that simply wasn't fundable from Neysa's earlier $50 million base. Founder Sharad Sanghi framed the round around exactly that gap, arguing India's AI ambitions need infrastructure capable of handling genuinely heavy computing loads, the same underlying compute-scarcity problem this publication has covered before: India has the model-building talent and the policy ambition, and has consistently lacked the raw hardware scale to match either.
Sovereign AI without domestic compute ownership is a sovereignty argument that ends the moment the largest infrastructure check gets written from outside the country.
Neysa's deal is the clearest concrete test case of that exact tension to date, and it resolved in the direction the tension predicts. When a domestic AI-infrastructure company needs to jump from $50 million to over a billion in a single round, precisely the scale required to compete on GPU deployment, the capital available at that size overwhelmingly comes from global private equity and sovereign-scale institutional investors, not Indian venture funds, because Indian venture capital simply doesn't operate at billion-dollar single-check sizes yet. Blackstone had the check size. Indian investors, structurally, could not have written it.
This wasn't a one-off bet on India
Blackstone's Neysa investment doesn't read as an isolated India wager once you look at where else the firm has already deployed the same playbook. The company's existing AI-infrastructure portfolio includes QTS, the world's largest data center platform, AirTrunk, the leading data center operator across Asia-Pacific, CoreWeave, a specialized AI cloud infrastructure company that itself became one of the most closely watched AI infrastructure plays globally, and Firmus, an Australian AI infrastructure platform. Neysa is the same bet, in the same category, made in a new geography, by an investor that has already run this exact strategy at least four times elsewhere. That context matters for reading the deal correctly: this isn't Blackstone discovering something uniquely Indian about Neysa's technology. It's Blackstone applying a proven global thesis, AI infrastructure ownership as an asset class, to whichever market currently has the most attractive entry price and the least developed domestic capital competing for the same deal.
Neysa itself belongs to a category the industry has started calling "neo-clouds," a new generation of AI-focused infrastructure providers built specifically to run GPU workloads at scale, distinct from both the traditional cloud giants and from a company that merely resells someone else's compute. TechCrunch's own framing of the deal placed it explicitly inside global supply constraints for specialized chips and data center capacity, the same bottleneck driving every major economy's AI infrastructure investment right now. Neysa isn't solving a uniquely Indian problem. It's solving the same GPU-scarcity problem every neo-cloud globally is racing to solve, with India as the specific market, and Blackstone treating that race the same way it treats the identical race in the US, Asia-Pacific, and Australia: as an asset class to acquire control of, not a sovereignty project to fund from the sidelines.
The deal's own paperwork tells you how it was actually treated
It's worth noticing who got hired to structure this, because deal complexity is itself a signal. DC Advisory served as Neysa's lead financial advisor, running what its own description called a multi-stream due diligence process alongside business-plan refinement and investor outreach strategy, the standard playbook for a company being professionally prepared for a major institutional buyer, not a founder-led fundraise run informally through existing relationships. KPMG advised Blackstone on the financial side. Talwar Thakore & Associates handled Neysa's legal work, while Trilegal and Gibson & Dunn, an international firm with serious cross-border private equity experience, represented Blackstone. That's the exact advisory stack you'd expect around a control transaction at this size, not a growth round where a founder keeps the keys. The professional infrastructure surrounding the deal was already treating it as an acquisition with growth financing attached, months before the announcement made that framing explicit in its own fine print.
What actually changes, and what doesn't
Nothing about Neysa's day-to-day operations changes immediately. Sanghi and Das remain in place, running the company, deploying the GPUs, serving the same customers. The technology stays Indian-built and Indian-deployed. What's changed is who ultimately answers to whom when Neysa's strategic priorities and Blackstone's return expectations diverge, a question that doesn't surface in year one of a deal like this and reliably surfaces by year three or four, once the fund's own return timeline starts applying pressure the founding team didn't originally sign up for.
The honest framing for this deal, the one none of the announcements used, is that India's AI compute buildout is being financed the same way most capital-intensive infrastructure gets financed globally: by whoever can write the largest check at the moment the largest check is needed, regardless of that capital's national origin. That's a completely normal, unremarkable way to finance infrastructure. It's also the opposite of what "sovereign AI" rhetoric implies when a government minister or a company's own press release uses the phrase. Neysa's GPUs will run inside India's borders. The decisions about how aggressively to expand them, and on whose return timeline, now sit with a fund headquartered in New York.



