NVIDIA Posts Record $96.2 Billion Q2 FY2027 Revenue as AI Chip Demand Keeps Accelerating
NVIDIA just delivered another quarter that reset expectations for how large a single company’s AI business can get. The chipmaker reported second quarter fiscal 2027 revenue of $96.2 billion, up 18 percent from the previous quarter and more than double what it posted a year earlier, comfortably clearing Wall Street’s forecast of roughly $92.3 billion. For a company that has already redefined what rapid growth looks like in the semiconductor industry, this report showed that the pace still hasn’t meaningfully slowed.
The headline number tells only part of the story. NVIDIA’s data center segment, which captures the bulk of its AI chip business, brought in $89 billion for the quarter, up 117 percent from a year ago and ahead of analyst expectations near $85.7 billion. That segment alone now dwarfs the entire company’s revenue from just a few years back, a reflection of how completely the AI infrastructure buildout has reshaped NVIDIA’s business. Within that figure, hyperscale customers contributed $48.7 billion, while the newer AI clouds, industrial, and enterprise category, which NVIDIA now breaks out separately, added another $40.3 billion.
Profitability kept pace with the top line. Net income came in at $59.69 billion, or $2.46 per share on a GAAP basis, compared to $26.42 billion a year earlier. On an adjusted basis, earnings landed at $2.22 per share, beating the consensus estimate of roughly $2.09. Gross margin held at 75 percent for the quarter, though NVIDIA’s guidance for the current quarter trimmed that figure slightly to around 74 percent, a detail that briefly rattled investors more than the record revenue itself did.
That margin softness traces back to rising memory costs, a pressure point affecting the entire chip industry this year rather than something unique to NVIDIA. Operating expenses jumped 55 percent to $8.41 billion during the quarter, and the company has already signaled it will raise prices on AI servers by more than 15 percent to offset climbing component costs. Even with that headwind, the market’s reaction told its own story. NVIDIA shares initially dipped after the report landed, then reversed and climbed roughly 4 to 5 percent in after hours trading as executives walked through the details on the earnings call, eventually trading above $220.
CEO Jensen Huang used the call to make a broader argument about where the AI industry currently stands. He described AI as having reached what he called its inflection point, telling investors that the technology is now doing useful, productive work rather than sitting in an experimental phase, and that compute itself has effectively become revenue. Huang also pointed to a shift in the competitive landscape, noting that where a single lab drove most of the buildout a year ago, NVIDIA is now seeing demand from multiple frontier labs, a growing wave of AI native startups, and an expanding open model ecosystem running in parallel.
Much of that demand is now flowing through NVIDIA’s newest architecture. The company confirmed that its Vera Rubin platform has moved into full production, with racks already running at major cloud providers including CoreWeave, Google Cloud, Microsoft Azure, Oracle Cloud, and Nebius, the AI infrastructure company in which NVIDIA holds a stake. Management expects Vera Rubin to represent roughly 20 percent of data center revenue in the current quarter, an early signal of how quickly the platform is scaling. Separately, AWS confirmed plans to deploy an additional 2 million NVIDIA GPUs, underscoring just how deep the hyperscaler commitment to NVIDIA’s roadmap continues to run.
Looking ahead, NVIDIA guided third quarter revenue to $108 billion, plus or minus 2 percent, which would mark the first time the company has projected quarterly revenue above the $100 billion threshold. That figure alone came in well above Wall Street’s prior expectations. In a move that stood out even by NVIDIA’s own standards, the company also offered a preliminary outlook for fiscal 2028, projecting approximately 70 percent annual revenue growth. Chief Financial Officer Colette Kress described that figure as supply constrained rather than demand constrained, meaning actual customer appetite for NVIDIA’s chips likely exceeds what the company can currently manufacture and ship.
There is a geopolitical wrinkle woven through the results as well. NVIDIA disclosed that its forward guidance assumes no data center compute revenue from China, a market that contributed billions of dollars in prior quarters before export restrictions and shifting trade dynamics complicated sales there. Hopper chip shipments to China accounted for less than 1 percent of data center revenue during the quarter, a sharp decline from prior periods and a reminder that even a company posting record numbers everywhere else is still navigating real constraints in one of the world’s largest technology markets.
NVIDIA also returned a substantial amount of capital to shareholders during the quarter, combining 20 billion dollars in stock buybacks with 6 billion dollars in dividends for a total of roughly 26 billion dollars. That kind of capital return, paired with record revenue and an aggressive growth outlook, reinforces how central NVIDIA has become to the broader AI infrastructure economy, sitting at the intersection of cloud computing, enterprise software, and now increasingly sovereign AI initiatives from governments building out their own compute capacity.
The quarter also arrived against a backdrop of growing investor anxiety about whether the broader AI spending boom can sustain itself. Companies like Microsoft, Amazon, and Google have reported strong cloud growth numbers that helped ease some of those concerns, even as rising capital expenditure from Google and Meta has drawn scrutiny over how long this level of spending can continue without clear returns. NVIDIA’s results this quarter effectively answered that question for now, showing that demand for its chips remains strong enough to support both record shipments and higher prices simultaneously.
For a company that already sits among the most valuable in the world, this report reinforces just how unusual NVIDIA’s growth trajectory has become. Four consecutive quarters of accelerating growth, a supply constrained rather than demand constrained outlook, and a product roadmap that hyperscalers are racing to deploy all point toward a company still expanding into demand it cannot fully satisfy. Whether that pace holds through fiscal 2028 will depend heavily on how supply chains, memory costs, and geopolitical trade policy evolve over the coming year, but for now, NVIDIA’s numbers suggest the AI infrastructure buildout still has considerable room left to run.
