Nvidia Guides Q3 Revenue Above $100B, Sees 70% Growth Next Year

Original | Odaily News (@OdailyChina)

Author | Azuma (@azuma_eth)

In the early hours of Aug. 27 Beijing time, Nvidia reported results for the second quarter of fiscal 2027, ended July 26, 2026. The company beat expectations on both revenue and profit, data center growth accelerated, and guidance for the next quarter set another record.

Specifically, Nvidia's Q2 revenue reached $96.22 billion, up 106% year over year and 18% sequentially, beating analyst estimates by 4%. Non-GAAP net income was $53.95 billion, up 118% year over year, and non-GAAP EPS came in at $2.22, about 6% above analyst estimates. Nvidia also guided Q3 revenue to $108 billion, meaning the company is about to officially enter the era of $100 billion in quarterly revenue.

Those numbers are already impressive, but they weren't the most striking figure in the report. On the post-earnings investor call, CFO Colette Kress delivered a signal the market had not fully priced in — the company expects fiscal 2028 revenue to still grow about 70% (excluding any data center revenue from China), far above Wall Street's prior 45% estimate.

More importantly, Nvidia stressed that this 70% growth outlook was given even under supply constraints. In other words, the main question facing the company no longer seems to be "how long can AI demand last," but rather "how much of Nvidia's demand can the supply chain actually meet."

Following the earnings release and call, Nvidia shares rose more than 4% in after-hours trading to $219.53.

Revenue Mix: AI Demand Still Dominated by Hyperscalers

A closer look at Nvidia's $96.22 billion revenue mix shows that the data center business remains the absolute core revenue source.

In Q2, data center revenue reached $89 billion, up 117% year over year and 18% sequentially, accounting for more than 90% of total revenue. More importantly, data center revenue growth continued to outpace overall company growth — in Q1, total revenue grew 85% year over year while data center revenue grew 92%; by Q2, those growth rates accelerated to 106% and 117%, respectively. Even with quarterly revenue approaching $100 billion, Nvidia's core business is not slowing down — it is accelerating.

Within the data center segment, Q2 growth was still primarily driven by hyperscalers. Nvidia disclosed that hyperscaler-related revenue in Q2 was approximately $48.7 billion. Meanwhile, Amazon, Microsoft, Google, Meta, and other large tech companies continue to expand AI infrastructure investment. Just before the earnings release, AWS announced plans to deploy an additional 2 million Nvidia GPUs.

This means that, at least based on the current revenue mix, the primary driver of AI compute demand remains capital expenditure by large cloud providers. This is also the most critical variable for assessing Nvidia's future growth trajectory and the sustainability of the AI cycle. If capex from Microsoft, Amazon, Google, Meta, and others continues to grow, Nvidia's high growth will have direct order support; conversely, if these companies' AI investments enter a digestion phase, Nvidia will be the first to feel the pressure.

On the other hand, Nvidia's Edge Computing business recorded $7.2 billion in revenue, up 27% year over year and 13% sequentially. While still significantly smaller than the data center business, it represents the company's long-term push into "physical AI" and an effort to reduce reliance on a single segment.

Vera Rubin: A New Product Cycle Has Begun

The most important product-level signal from the earnings report and call was the ramp of Vera Rubin — Nvidia confirmed on the call that Vera Rubin has entered full mass production and began shipping earlier this month… Every 1 GW of compute deployed with Vera Rubin corresponds to roughly $40 billion in revenue opportunity, and it is expected to contribute about 20% of data center revenue in Q3.

This announcement marks Nvidia's gradual transition from the Blackwell cycle to the Rubin cycle.

For a company with such rapid product iteration, this transition between old and new platforms is critical. In the past, a major concern about the AI chip cycle was whether each generation could smoothly hand off to the next; if a significant gap emerged between two product generations, revenue growth would naturally decelerate quickly. This time, Rubin's commercialization pace is clearly fast enough.

As Nvidia's rack-scale supercomputing platform designed specifically for AI, Vera Rubin integrates seven custom chips including Rubin GPUs, Vera CPUs, NVLink6, and ConnectX-9, delivering a leap in performance. Product evolution ultimately feeds back into value capture — the anchor of "1 GW of compute → $40 billion in revenue" means that once Vera Rubin ramps, as data center scale continues to expand, the value Nvidia can extract from each new AI data center will keep rising.

Additionally, Jensen Huang himself emphasized on the call that Vera Rubin's ramp speed will be the fastest in company history… This may be the foundation for Nvidia's confidence in guiding fiscal 2028 revenue growth of about 70%.

AI Demand Remains, but Supply Has Become the Bottleneck

On the question of whether AI demand has peaked, Nvidia's management sent a stronger signal than ever on the post-earnings call: demand is not slowing — it is accelerating.

Jensen Huang stated bluntly on the call that AI has entered a new phase. Over the past year, AI compute demand was primarily driven by a handful of leading model companies; now, demand is spreading to more frontier models, open-source models, enterprise AI, agents, and robotics.

The most important sign of this shift is the accelerating growth in inference demand. As large models move from training to real-world applications, AI compute consumption is no longer a one-time cost during model training but grows with user activity and token generation. Huang even used the phrase "tokens are becoming productive and profitable" to describe the change.

Furthermore, Nvidia said on the call that based on current capex plans from major tech companies, there are no clear signs of demand slowing. The world's top five cloud service providers are expected to spend nearly $800 billion in capex this year, potentially reaching about $1.3 trillion next year.

These observations collectively underpin Nvidia's judgment on future AI demand. However, with demand so strong, supply has become the constraint limiting Nvidia's further growth. In other words, the main bottleneck the company faces is not a lack of orders, but whether components, production capacity, and delivery capability can keep pace with order volume.

Nvidia's management made clear on the call that supply will remain the primary bottleneck limiting growth at least through fiscal 2028. Huang even stressed that without supply constraints, the company's fiscal 2028 outlook would be "much higher."

"Supply" here is not limited to GPUs themselves. As AI chip scale expands rapidly, HBM, advanced packaging, wafers, networking equipment, and data center power could all become critical factors constraining compute deployment speed. HBM in particular — Nvidia directly addressed the sharp rise in memory prices on the call and guided Q3 non-GAAP gross margin to approximately 74%, down from 75% in Q2, as rising memory costs pressure future profitability.

To break through this bottleneck, Nvidia is extending its role to become a "capital orchestrator" for AI infrastructure. The earnings report disclosed that Nvidia has established strategic partnerships with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR, aiming to mobilize over $500 billion in third-party capital for AI infrastructure construction. At the same time, Nvidia is working with SB Energy at the PORTS-Pike Technology Campus in Ohio to secure land, power, and building shell capacity in advance to support its compute deployment.

The $100 Billion Threshold and the Next Starting Point

Nvidia's earnings report and management's comments on the call once again provided a phased answer to the market's most pressing question — has AI demand peaked, and is the cycle over?

Based on the fiscal 2028 revenue growth guidance of 70%, Nvidia's current answer is that "the cycle is far from over" — the ceiling is still distant, and the current growth rate does not even reflect true demand, only the limit of what the supply chain can support. When a company with quarterly revenue approaching $100 billion still dares to guide growth far above market expectations under supply constraints, it can only mean that the orders Nvidia sees are even thicker than the numbers it is willing to put in its guidance.

And Nvidia itself is adapting to market development needs by continuously expanding its commercial role. Today's Nvidia is no longer just a semiconductor company that designs GPUs; it is evolving into a general contractor for "AI factories," providing chips, networking, CPUs, software, deployment solutions, and even participating in infrastructure financing and land/power lockups.

Of course, high expectations come with high pressure. The modest decline in gross margin guidance and policy uncertainty in global markets are inevitable frictions in expansion at the $100 billion scale… But at least for now, the core logic behind the market's willingness to pay a premium for Nvidia remains intact — in the AI infrastructure race, Nvidia is still the player that sets the rules, with no visible challenger.

When Q3 revenue officially crosses the $100 billion threshold, Nvidia will enter a new order of magnitude. And the 70% growth guidance for next year suggests that this threshold may be just the starting point for the next leg.