Sharon AI signs $1.32B contract as AI labs lock in compute

A five-year, $1.32B compute agreement could turn Sharon AI from an emerging neocloud into a major regional supplier of frontier-model infrastructure. The size of the contract also shows how aggressively AI laboratories are reserving capacity years before it comes online.

What happened

Sharon AI signed a five-year cloud-computing agreement with an unnamed global AI laboratory. The company expects the contracted service to begin generating revenue in the first half of 2027 through data-centre infrastructure in New Zealand.

Sharon AI says its planned AI-factory footprint totals 132MW, with 116MW already contracted, and that it expects to deploy more than 62,000 Nvidia GPUs by mid-2027. Those figures imply a substantial buildout across power, cooling, networking and data-centre capacity—not simply a software resale arrangement.

The customer’s identity, minimum-usage obligations, cancellation protections and payment guarantees were not disclosed. That makes the headline contract value important but not identical to recognised revenue.

Why it matters

Frontier AI developers increasingly need to secure compute through long-term agreements because chip supply, power access and data-centre construction remain constrained. Specialist neoclouds can offer dedicated capacity without forcing customers to rely entirely on the largest hyperscalers.

For Sharon AI, the agreement could provide a credible demand anchor for financing and building new infrastructure. It also creates concentration risk if a large share of future revenue depends on one customer.

The bigger picture

AI competition is becoming a race to control physical capacity as much as model quality. The winners will need dependable power, capital and operational execution—not merely access to GPUs. Sharon AI still has to deliver the infrastructure on schedule and prove that its customer commitments translate into durable cash flow.