After the bell on Wednesday, NVIDIA dropped its fiscal Q2 2027 scorecard, and honestly, it was a knockout. Both revenue and net profit absolutely crushed Wall Street’s forecasts. On top of that, the CFO just teased a hefty 70% revenue growth projection for fiscal 2028, while the Street had been bracing for something closer to 45%.
When the numbers first hit the tape, shares did a little jitterbug, dipping briefly before catching a serious bid. The momentum came straight from management’s bullish 70% growth outlook, sending the stock soaring as much as 4.7% in extended trading.
That pop helped soothe the sting of seven straight down sessions, which had dragged the stock down 7.5% in total, its longest losing streak in four years dating back to September 2022. At that low point, shares had fallen more than 10% from the all-time closing high of $235.74 set back on May 14.
NVIDIA’s rosy guidance didn’t just lift its own stock—it lit a fire under the whole AI infrastructure complex. Memory, optical comms, and AI cloud names all rallied in after-hours trading. SK Hynix ADRs jumped over 4%, Applied Optoelectronics added 4%, CoreWeave popped more than 5%, and Nebius surged a solid 6%.
Let’s dig into the actual numbers. NVIDIA pulled in a massive $96.22 billion in total revenue for the quarter, up a jaw-dropping 106% from the same period last year and 18% sequentially from Q1. Wall Street analysts polled by LSEG had penciled in around $92.17 billion on average—so NVIDIA beat that by a whopping $4 billion plus. Any lingering concerns about momentum cooling off? Yeah, you can toss those out the window.
On the profit side, net income landed at $53.95 billion, doubling year over year. The earnings engine is firing on all cylinders, keeping perfect pace with that explosive revenue growth.
Profitability metrics? Also rock solid. Both GAAP and non-GAAP gross margins held steady at a muscular 75.0%, sitting right at the top of the industry charts. Adjusted EPS came in at $2.22, up 120% year over year and comfortably ahead of the $2.10 analysts had modeled. Bottom line: the earnings quality blew past what the capital markets were pricing in.
Breaking it down by segment, data center remains NVIDIA’s undisputed crown jewel—the absolute engine pulling the entire earnings train. That division generated $89 billion in revenue for the quarter, up 117% year over year and 18% quarter over quarter. It now accounts for roughly 92.5% of the company’s total revenue pie.
Peek inside the data center customer mix and the story gets even more interesting. Hyperscale cloud providers chipped in $48.71 billion, a 102% year-over-year jump. Meanwhile, AI cloud players, industrial firms, and enterprise customers brought in $40.3 billion, growing at a blistering 138% clip. Here’s the kicker: enterprise AI compute procurement is now growing faster than the big cloud guys. The downstream demand map is clearly broadening out.
The smaller segments—consumer gaming GPUs, professional visualization, automotive—held their own without any drama. They’re not huge contributors to the top line, but they’re running steady. And on the demand side, order flow from enterprises and AI startup labs keeps pouring in, shaping up to be the second growth curve after the cloud giants.
On the earnings call, NVIDIA’s CFO Colette Kress said the company’s growth is actually going to accelerate next year, guiding to 70% revenue growth for fiscal 2028. She added, “Even at our current scale, demand is still accelerating. Customer forecasts suggest our growth next year will double.”
Kress also threw in a crucial caveat: the real demand for compute procurement that customers are tracking could actually be heading toward 100% growth. But here’s the rub—supply chain constraints around chips and memory components mean NVIDIA can only reliably deliver enough product to support that 70% revenue growth figure.
That’s a telling signal. The ceiling on NVIDIA’s growth right now isn’t a lack of market demand. It’s purely a bottleneck on the upstream supply side.
CEO Jensen Huang piled on, saying that if it weren’t for supply limitations, the company’s fiscal 2028 outlook could be “much higher.”
Looking ahead to the current quarter—fiscal Q3 2027—NVIDIA guided revenue to roughly $108 billion, give or take 2%. On the margin front, Q3 GAAP gross margin is expected to hover around 74%, with a 50-basis-point wiggle room on either side. That’s a slight step back from Q2’s 75% level, which validates the narrative that rising memory costs are starting to bite. It also came in just below the 75% analysts were looking for, which triggered a brief after-hours dip of more than 4%.
On that same call, Jensen Huang doubled down on his long-term optimism for the AI industry. His words: “Artificial intelligence has reached an inflection point. AI is producing real utility. Tokens are productive and generating revenue—compute itself is revenue. Compute demand is accelerating right now, and we’re in a golden era for new AI labs and startups. Multiple frontier labs are scaling up their compute infrastructure simultaneously.”
This blowout report is a shot of adrenaline for the entire AI compute supply chain. Since the AI wave kicked off, NVIDIA has been cranking out triple-digit revenue growth quarter after quarter. But even with all this momentum, there are a few landmines worth watching.
For one,rivals like AMD and Google keep rolling out next-gen compute silicon,and the competitive heat is rising. Meanwhile, the global memory chip shortage shows zero signs of letting up, which is driving NVIDIA’s memory procurement costs significantly higher. Both of these could throw a wrench into future growth.
Morgan Stanley’s numbers put this in perspective: NVIDIA’s next-gen flagship AI rack system, Vera Rubin, carries a price tag of roughly $7.8 million. Memory costs alone have exploded from around $370,000 to $2 million—a gut-wrenching 435% increase. That means memory’s share of total system cost has jumped from under 10% in the previous generation to 26% now.
There’s also the growing web of equity investments NVIDIA has made in core customers like OpenAI and Anthropic. The line between vendor and customer is getting blurry, and some analysts are starting to poke around for potential risk factors hiding beneath those headline revenue numbers.
