Nvidia projects 77% growth as Vera Rubin enters market

Think Nvidia’s (NVDA) growth story is getting old?

Think again.

Nvidia Vera Rubin AI supercomputer is already reshaping expectations.

Nvidia just told investors revenue is set to surge 77% to about $78 billion this quarter.

Wall Street was looking for roughly $72 billion to $73 billion. That is a massive beat for a company already valued at about $5 trillion. The milestone reinforces how Nvidia became the first $5 trillion company as AI demand soared, a valuation few thought possible just a couple of years ago.

It would be Nvidia’s fastest quarterly growth since early 2025 and its 11th straight quarter with growth above 55%. At this point, the data center business makes up more than 91% of total sales. Strategic capital moves like Nvidia investing in CoreWeave show how the company is strengthening its grip on AI cloud infrastructure beyond just hardware.

And there’s more fuel coming.

CFO Colette Kress confirmed the first samples of the new Vera Rubin platform shipped this week. The full rack-scale system combines 72 next generation GPUs and is designed to deliver about 10x the performance per watt of the prior Grace Blackwell platform. Nvidia expects major AI model builders and cloud providers to roll it out over time.

Management also said growth should outpace earlier projections tied to a $500 billion revenue opportunity across the Blackwell and Rubin product lines. Inventory and supply commitments now extend into 2027, giving Nvidia rare visibility for a company in tech.

Nvidia is becoming the infrastructure layer for AI across the global economy. If you believe AI spending keeps climbing, this stock sits at the center of that spend. AI is one of the 7 disruptive technologies that will change the world.

And yet, the stock barely budged after hours.

That tells you something important. The market has gotten used to upside surprises from Nvidia. For a $5 trillion company to keep posting numbers like this, expectations are sky high. The question now isn’t whether Nvidia is growing, but how long it can keep growing at this speed.

The disruption behind the news: Vera Rubin collapses the cost curve of intelligence.

Ten times more performance per watt changes the math for every AI workload.

Energy, not chips, has been the binding constraint.

Efficiency gains don’t just help adoption.

They expand Nvidia’s pricing umbrella.

At about $0.07 per kWh, a 100MW AI cluster spends roughly $61 million per year on electricity. That’s 100 × 8.76 million kWh × $0.07. If rack level output per watt rises about 10x, the same compute could run on about 10MW, cutting that to roughly $6 million per year. That’s a $55 million annual gap.

That $55 million per year is surplus value hyperscalers can rationally hand back to Nvidia through higher rack prices and still come out ahead. Efficiency doesn’t just lower costs for customers. It can increase Nvidia’s take rate at the system level. And Nvidia is directly marketing inference economics as an order of magnitude step change.

This kind of infrastructure expansion mirrors recent stock moves for example when CoreWeave stock jumped after Nvidia committed $2 billion to AI expansion, underscoring how Nvidia is shaping the broader AI ecosystem, not just selling chips.

If performance per watt scales anywhere near 10x at the rack level, hyperscalers can push far more tokens per dollar of power and cooling. AI inference becomes cheaper to deploy at scale. Training runs that once required massive capital outlays start to look routine. That pulls forward adoption across enterprise software, robotics, biotech, defense, and media. In fact, many investors are already asking whether robotics is the next big thing for Nvidia as cheaper inference unlocks physical-world AI deployment.

Nvidia now owns the rack, not just the chip. That’s the strategic kill shot. By selling rack scale systems, it locks customers into its architecture at a higher level of integration. Switching costs rise. Procurement cycles lengthen. Competitors have to match not only silicon but full stack systems.

Yes, Advanced Micro Devices is bringing Helios later this year, and Meta Platforms plans to deploy up to 6 gigawatts of AMD GPUs starting in 2026. That sounds big. But Nvidia already has supply lined up through 2027 and expects every major cloud to adopt Rubin. The difference between shipping samples now and promising deployments in 2026 is an eternity in AI cycles. The company’s deepening ecosystem ties are evident as Meta expands its Nvidia partnership with millions of AI chips, reinforcing Nvidia’s central role in hyperscale AI infrastructure.

The bigger threat is vertical integration. Amazon and Google are building in house chips. Nvidia is bluntly excluding China data center revenue from guidance due to export controls. That’s geopolitical risk already baked in. And it’s still guiding to 77% growth. That’s what dominance looks like.

What to watch next

Watch power availability.

Watch rack level adoption timelines.

Watch gross margins under system level integration.

If Rubin truly delivers 10x performance per watt, data center buildouts will accelerate wherever power can be secured. Utilities and grid operators become key gatekeepers in the AI economy. Regions with surplus energy win. Others fall behind.

Second, track how fast hyperscalers standardize on Rubin racks. If deployment cycles compress from years to quarters, Nvidia tightens its grip. If customers diversify to AMD or in house silicon at scale, pricing power weakens. The next 12 to 24 months will show whether this is a monopoly platform or a peak cycle.

Finally, watch margins. Selling full racks can dramatically increase average selling prices, but integration complexity can hurt profitability. If Nvidia maintains or expands margins while scaling system sales, it cements itself as the operating system of AI infrastructure.

This isn’t just a temporary upswing, but industrial consolidation happening right in front of us.

I’m backing the company that keeps shipping faster, denser, more efficient compute. When Nvidia Vera Rubin shifts the economics this dramatically, it signals the cost curve is still moving in Nvidia’s favor.

Nvidia (NVDA) has a Disruption Score of 4. Click here to learn how we calculate the Disruption Score.  

Nvidia is also part of the Disruption Aristocrats, our quarterly list of the world’s top disruptive stocks.