Nvidia’s $108 Billion Forecast Resets the AI Spending Debate

Nvidia has once again delivered numbers that would look extraordinary for almost any other company.

The bigger question is whether they are extraordinary enough for Nvidia.

The AI chip leader reported fiscal second-quarter 2027 revenue of $96.2 billion, up 106% from a year earlier and 18% sequentially. Data Center revenue reached $89.0 billion, rising 117% year over year. Non-GAAP earnings came in at $2.22 per diluted share.

But the number that matters most for the broader AI trade may be Nvidia's next-quarter forecast.

Management expects fiscal Q3 revenue of approximately $108 billion, plus or minus 2%, even without assuming any Data Center compute revenue from China.

That forecast strengthens one of the most important arguments behind the AI bull market: despite historic spending on data centers, demand for compute still appears to be accelerating.

Nvidia's Growth Problem Is Now Expectations

Nvidia's results illustrate the unusual position the company occupies.

Revenue has more than doubled in a year, yet investors are no longer asking whether Nvidia can grow quickly. They are asking how long growth at this scale can continue.

The initial reaction after earnings reflected that tension. Nvidia shares first moved lower before reversing higher during and after management's earnings commentary.

CEO Jensen Huang indicated revenue could grow around 70% in the following fiscal year, reinforcing the view that the current AI infrastructure buildout may have substantially further to run.

That is important because Nvidia has become more than a semiconductor earnings story.

It has become one of the market's primary indicators for global AI capital expenditure.

If Nvidia's forward demand remains strong, the implications extend to memory suppliers, networking companies, data-center operators, power infrastructure and other semiconductor companies.

Rubin Is Beginning to Matter

One of the most important changes this quarter is the arrival of Nvidia's next major architecture.

The company said the Vera Rubin platform is ramping into full production, with systems already running at partners including Google Cloud, Microsoft Azure, Oracle Cloud Infrastructure, CoreWeave and Nebius.

Rubin could represent roughly 20% of Nvidia's Data Center revenue during the current quarter.

That reduces one risk investors have worried about repeatedly during Nvidia's rapid rise: product transitions.

If customers delayed purchases while waiting for the next architecture, Nvidia could suffer an air pocket between product cycles.

So far, the transition from Blackwell toward Rubin does not appear to be producing that kind of collapse in demand.

Instead, the market is seeing another generation of infrastructure arrive while spending on existing AI systems remains enormous.

The AI Spending Boom Is Still the Central Story

Nvidia's numbers need to be viewed alongside spending by its largest customers.

Major technology companies including Microsoft and Meta are collectively directing hundreds of billions of dollars toward AI infrastructure. Combined AI-related infrastructure spending among major technology companies could exceed $730 billion this year.

That helps explain why Nvidia can continue growing even from such a large revenue base.

There are now multiple frontier-model developers, cloud providers, sovereign AI projects and specialized AI companies competing for infrastructure.

The market is therefore moving beyond the first phase of the AI boom, when a relatively small group of hyperscalers dominated spending.

The critical question is whether those customers can eventually earn adequate returns on all that capital.

There Are Still Risks

The earnings report was not entirely clean.

Nvidia expects gross margin to decline to approximately 74% in Q3, from 75% in Q2, as component and memory costs rise.

China also remains uncertain. Nvidia's Q3 guidance assumes no Data Center compute revenue from the country, reflecting continuing export-control uncertainty.

Then there is the broader financing question.

As the AI ecosystem grows, investors have become increasingly sensitive to whether some infrastructure demand is being supported through complicated financing arrangements, strategic investments or vendor-backed commitments.

Those concerns do not negate Nvidia's reported revenue.

They do mean investors increasingly need to distinguish between end-user AI demand and the financing structures supporting infrastructure expansion.

What Nvidia's Quarter Really Tells Investors

The most important conclusion from Nvidia's August earnings is not that the company beat expectations again.

It is that the AI infrastructure cycle has still not shown a clear demand ceiling.

Nvidia generated $96.2 billion in quarterly revenue and is already guiding for $108 billion in the next quarter. Rubin is ramping, Data Center revenue more than doubled, and management continues to describe accelerating demand.

That does not guarantee Nvidia's stock will continue rising. Expectations, valuation, margins and capital intensity still matter.

But the latest results make one bearish argument harder to defend:

There is still little evidence that the AI infrastructure boom is ending.