Nvidia on Wednesday reported fiscal second-quarter revenue of $96.2 billion, up 106% from a year earlier and ahead of analyst estimates. Non-GAAP earnings per share were $2.22, while Data Center revenue reached $89.0 billion, up 117% year over year.

The company said it expects third-quarter revenue of $105.8 billion to $110.1 billion — a range that implies continued quarter-over-quarter growth — and projected fiscal 2028 annual sales growth of 70% from the prior year, a forecast covering the fiscal year that runs through late January 2028.

“Demand is accelerating,” Jensen Huang, founder and CEO of Nvidia, said in a statement.

The results renewed debate over whether AI spending can remain sustainable, while also giving Nvidia and its supporters fresh evidence of continued expansion.

“This was a masterpiece quarter with stunning guidance that speaks to the massive demand Nvidia is seeing in the AI Revolution,” Dan Ives, partner and senior managing director at Yorkville Ives, told me. “The guidance for next quarter was well ahead of whisper numbers, and this will be a spark that is a boost for tech stocks and the broader sector.”

During the earnings call, CFO Colette Kress delivered a data-driven response to the AI-bubble narrative. Rather than relying on broad assertions about durable demand, Kress said Nvidia’s growth is becoming more diversified beyond hyperscalers such as Microsoft, Google, Amazon and Meta.

“Hyperscalers will remain a major growth driver, but non-hyperscaler growth, our AICE segment, spanning sovereign, regional NeoClouds, enterprise edge and air-gap data centers, will represent roughly half of our data center business,” Kress said.

Of the $89 billion in Data Center revenue, $49 billion came from hyperscalers, up 13% sequentially, while the rest came from what Nvidia calls ACI&E — enterprise, industrial and NeoCloud customers, which Kress grouped under “AICE.”

The categories she cited span a broad range of buyers. Sovereign AI programs are government-backed efforts to build domestic computing capacity; regional NeoClouds are smaller cloud providers that rent out GPU access; and air-gapped data centers run on networks physically isolated from the internet, an approach used where security is paramount, such as defense and intelligence work.

That means Nvidia’s business is no longer dependent only on the capital spending plans of a handful of Big Tech companies. It is increasingly spread across sovereign AI programs, regional cloud providers, corporate deployments and air-gapped or edge systems.

“Our AI-native start-up ecosystem developed and running primarily on the Nvidia compute platform is scaling at a rapid pace,” Kress said.

She pointed to venture funding as another sign of demand. “Global VC funding in AI, roughly 70% of which is spent on compute, exceeded $400 billion in the first half of 2026, surpassing the $265 billion raised in all of 2025,” she said.

Kress’s argument was that a significant share of those venture dollars ultimately goes toward GPU purchases or cloud rental, which benefits Nvidia.

She also highlighted customer growth. “Nearly 20 companies, including Cursor, owned by SpaceX, Figma and Together AI, now exceed $1 billion in annualized run-rate revenue, up from 13 companies in Q4 of last year, with vertical enterprise software logging the fastest growth,” she said.

The rapid growth in vertical enterprise software suggests that AI adoption is moving beyond experimentation and into embedded, revenue-generating business software.

“Jensen and Nvidia help put to rest some concerns about financing and balance sheet issues that have been an overhang on the tech sector and capex cycle buildout for the hyperscalers,” Ives said. He added that the demand and metrics “were off the charts” and pointed to an acceleration in AI spending. “The yields/debt issue is not going away, but this shows monetization is happening quicker than expected,” he noted.

His reference to yields and debt points to the financing side of the debate: part of the data center construction underpinning AI compute is debt-financed, which ties the sector’s outlook to credit conditions as well as to chip demand.

Sheryl Estrada Sheryl.Estrada@fortune.com

This story was originally featured on Fortune.com