Nvidia’s Q3 showed the company’s GPU momentum return, delivering a substantial data center beat with 25% QoQ growth, surpassing an important $50 billion quarterly revenue milestone for the segment. More importantly, Nvidia’s guide pointed to this momentum continuing into the fourth quarter, implying that data center revenue could be on track to rise another $8 billion QoQ for 15% growth.
Yesterday I published an article entitled Why Nvidia Stock Could Reach $20 Trillion Market Cap by 2030 – a prediction that requires a 36% CAGR over a five-year period or about 8% growth QoQ. These two quarters alone meet the criteria for next year’s CAGR plus some.
Margins expanded sequentially, free cash flow sustained nicely, networking was up an impressive 162% and compute was up 56%. Management repeated they “currently have visibility to $0.5 trillion in Blackwell and Rubin revenue from the start of this year through the end of calendar year 2026.”
Below, we look at one other metric that hints that Blackwell Ultra can continue to deliver knockout earnings report while our sights are now set on Vera Rubin for H2 2026.
Nvidia Surpasses $50 Billion Quarterly Data Center Revenue
Nvidia surpassed the $50 billion quarterly data center revenue milestone in Q3, as it reported $51.2 billion in revenue for the segment, up 25% QoQ and 66% YoY. This is the highest QoQ growth rate for data center since fiscal Q4 2024. An impressive feat to deliver such strong growth at this scale considering the segment was just $18.4 billion at the time. On a dollar basis, data center revenue rose by $10.1 billion sequentially.
This sequential growth was driven by a strong inflection in Compute revenue, which surged 27% QoQ to $43 billion, its highest sequential growth rate since fiscal Q1 2025; however, this does come after a (1%) QoQ decline in fiscal Q2. Nvidia noted that Blackwell Ultra was ramping across all customer categories and became its leading architecture.
Q4’s guidance suggests that this $50 billion data center segment will quickly be in the rear view mirror, with the $65 billion guidance implying data center revenue of around $59 billion assuming similar mix shift as Q3. This represents another 15% QoQ growth on top of Q3’s 25%, or essentially the data center segment rising nearly 44% in just two quarters.
This would also correspond to a nearly $8 billion QoQ increase, meaning that if Nvidia maintains this growth cadence through mid-CY26, then it would reach our prediction for a $75 billion data center segment two quarters early. If this materializes, this would represent data center growth of 66% YoY, up from 56% last quarter.
It also could suggest Nvidia potentially reaching a $90 billion quarterly data center segment if this trajectory is maintained through the end of fiscal 2027. However, it is important to note that given the sheer scale of data center revenue, there is the potential for this inflection to be lumpy. Therefore, we are directionally bullish but not with too specific of a timeline; rather, these are milestones that tell us how quickly we could see $8 trillion market cap for example and then onward.
Blackwell Revenue Tops $100 Billion
As stated in our $20 Trillion pre-earnings writeup, after taking into account Blackwell revenue that will ship in FY2026, this should lead to a $320 billion data center segment next year. Here is what was stated in the $20 Trillion analysis: “Reading between the lines on Huang’s comments suggests strong upside to Nvidia’s data center revenue through 2026. Over the prior three quarters heading into fiscal Q3’s report, Blackwell revenue has totaled approximately $63 billion. Including Networking over that time frame, total revenue would rise to $78 billion, still a fraction of the total overall opportunity management is projecting. Thus, if we assume that Blackwell and Rubin ramp over the next five quarters, fiscal 2027 data center revenue could be nearly $320 billion, versus estimates for around $270 billion.”
Since we wrote that earlier this week, analyst estimates have been rising and now stand at $292 billion for next year. With information from this report, we look to be on track for a $317 billion data center segment next year (close to our original estimate this morning).
We calculated this from the prior three quarters heading into fiscal Q3’s report, Blackwell revenue has totaled approximately $63 billion. Now, Q3’s Compute revenue of $43 billion implies Blackwell has delivered around $104 billion in revenue in the past four quarters, assuming the only non-Blackwell revenue was the $2 billion disclosed from Hopper.
Including Networking and Q4’s guidance, Nvidia looks to be on track to generate $186 billion of its $500 billion opportunity in fiscal 2026. This would leave approximately $314 billion for fiscal 2027’s data center revenue to meet the $500 billion visibility, but if Nvidia can exceed that by 2-4%, it could be on track for $330 billion next year. Management sounded confident to achieve the $500 billion target and they hinted that they could exceed it as the CFO stated, “So there's definitely an opportunity for us to have more on top of the $500 billion that we announced.”
One Figure Says This Growth Inflection Will Continue
While we continue to hammer on the importance of Big Tech’s capex as the number one indicator for Nvidia’s data center growth continuing, there was potentially a more important, well overlooked figure in Nvidia’s report that signals this data center inflection will continue.
Nvidia’s total supply-related commitments, such as for CoWoS wafers, HBM memory, or other components, surged nearly 52% QoQ to $50.3 billion in Q3, with management noting that they are “ordering to secure long lead-time components, meet the demand for Blackwell, and support future architecture ramps.”
This is a notable increase from the prior five-quarter average of ~$30 billion, which is likely supporting the current ramp in data center revenue. This uptick in supply commitments, which is likely to translate into inventories and revenue over the coming four to six quarters, hints that Nvidia will continue ramping Blackwell output while preparing for Rubin’s production in the second half of 2026.
This also bolsters confidence in Nvidia’s order visibility to fill out and even exceed this cumulative $500 billion in Blackwell and Rubin revenue, as the company would not need to boost supply commitments by this degree if the demand signals were not there.
There is Global Demand for Nvidia’s GPUs
When hearing about an AI bubble, it’s important to remember there is global demand for Nvidia’s GPUs. The diversification across geographic regions, enterprises, startups – and of course, Big Tech, helps to insulate Nvidia should one customer or region slow their spending. Here is what was stated on the call:
“And then lastly, remember, we were just talking about the American CSPs. Each country will fund their own infrastructure. And you have multiple countries, you have multiple industries. Most of the world's industries haven't really engaged agenetic AI yet, and they're about to. All the names of companies that you know we're working with, whether it's autonomous vehicle companies or digital twins for physical AI for factories and the number of factories and warehouses being built around the world, just a number of digital biology start-ups that are being funded so that we could accelerate drug discovery. All of those different industries are now getting engaged, and they're going to do their own fundraising. And so don't just look at the hyperscalers as a way to build out for the future. You got to look at the world, you got to look at all the different industries and enterprise computing is going to fund their own industry.”
Commentary on Vera Rubin
If only a Nvidia investor could kick back and call it a day! Instead, given Blackwell Ultra now comprises 2/3 of revenue confirming a successful launch, our sights are now set on Vera Rubin commentary. According to management, Rubin is set for a fast ramp:
“The Rubin platform is on track to ramp in the second half of 2026. Powered by 7 chips, the Vera Rubin platform will once again deliver an X-factor improvement in performance relative to Blackwell. We have received silicon back from our supply chain partners and are happy to report that NVIDIA teams across the world are executing to bring up beautifully.
Rubin is our third-generation rack-scale system substantially redefined the manufacturability while remaining compatible with Grace Blackwell. Our supply chain data center ecosystem and cloud partners have now mastered the build to installation process of NVIDIA's rack architecture. Our ecosystem will be ready for a fast Rubin ramp.”
Financial Overview:
Strong Revenue Growth of 63%
Nvidia’s Q3 revenue grew by a solid 62.5% YoY and 22% QoQ to $57.01 billion. Revenue growth accelerated by 6.9 percentage points from 55.6% YoY growth reported in Q2. Revenue beat estimates by 3.5% and is the strongest beat in the last four quarters. The company’s strong revenue growth dispelled fears of an AI Bubble. Nvidia’s CEO Jensen Huang said, “Blackwell sales are off the charts, and cloud GPUs are sold out.”
The Blackwell revenue gained further momentum in the recent quarter. The GB300 sales were higher than the GB200 sales, notably accounting for 2/3 Blackwell’s revenue, driven by strong demand from cloud companies and hyperscalers. The Hopper platform contributed approximately $2.0 billion in revenue. While H20 sales were negligible at $50 million, management is working with the US and Chinese governments to ship products to China. Looking forward, Rubin is on track to ramp in the second half of 2026.
Management also provided a strong Q4 revenue guide of $65 billion at midpoint, representing a YoY growth of 65.3% and up 14% QoQ. It beat the estimates by 5.1%. Looking forward, analysts expect revenue to grow 40.5% YoY to $292.1 billion for FY2027 and 24.8% YoY to $364.6 billion for FY2028.
Networking Revenue Growth of 162%
The company’s networking revenue was an outlier, growing 162% YoY and 13% QoQ to $8.19 billion. Revenue growth accelerated by 84 percentage points from 78% YoY growth in Q2. Management stated in the earnings call that the company’s networking business is specifically built for AI and is now the largest in the world. The strong growth was primarily due to NVLink scale-up and robust double-digit growth across Spectrum-X Ethernet and Quantum-X InfiniBand.
Management stated in the earnings call that Meta, Microsoft, Oracle, and xAI are building gigawatt AI factories with Spectrum-X Ethernet switches, further highlighting the flexibility and openness of the company’s platform.
The company introduced Spectrum-XGS Ethernet in August, which will enable to connect distributed data centers into Giga-Scale AI Super-Factories. Nvidia is the only company with AI scale-up, scale-out and scale across platforms, reinforcing the unique position in the market as the AI infrastructure provider.
- Q3 gaming revenue grew by 30% YoY and was down (1%) sequentially to $4.27 billion. Management mentioned that channel inventories have reached more normalized levels heading into the holiday season.
- Pro visualization revenue grew by 56% YoY and 26% sequentially to $760 million. Colette Kress, CFO, said in the earnings call, “Growth was driven by DGX Spark, the world's smallest AI supercomputer, built on a small configuration of Grace Blackwell.”
- Automotive revenue grew by 32% YoY and up 1% QoQ to $592 million. The CFO highlighted, “We are partnering with Uber to scale the world's largest Level 4 ready autonomous fleet built on the new NVIDIA Hyperion L4 robotaxi reference architecture.”
- OEM and other revenue grew by 79% YoY and 1% QoQ to $174 million.
Margins
The company’s margins beat management guidance and are expected to expand in Q4.
- Q3 gross profits grew by 60% YoY to $41.85 billion. Q3 gross margin was 73.4%, beating management guidance of 73.3% by 10 basis points. Gross margin was up 100 basis points sequentially and down 120 basis points YoY. Q4 gross margin guide is 74.8%, up 140 basis points sequentially and up 180 basis points YoY. Adjusted gross margin was 73.6%, beating the management guidance by 10 basis points. Management expects an adjusted gross margin of 75% in Q4, up 140 basis points sequentially and up 150 basis points YoY.
- Looking forward, management mentioned in the earnings call that the input costs are increasing and are looking to hold gross margins in the mid-70s range for FY2027.
- Q3 operating income grew by 65% YoY and 27% sequentially to $36.01 billion. The operating margin was 63.2%, beating the management guidance by 80 basis points. Adjusted operating margin was 66.2%, beating the management guidance by 50 basis points. Management has provided a strong operating margin guide of 64.5% and an adjusted operating margin guide of 67.3% for Q4.
- Q3 net profits grew by 65% YoY and 21% QoQ to $31.9 billion with a net profit margin of 56% compared to 55% in the same period last year and 56.6% in Q2. Adjusted net profits grew by 59% YoY and 23% QoQ to $31.77 billion with an adjusted net profit margin of 55.7%, compared to 57% in the same period last year and 55.2% in Q2.
Adjusted EPS grew by 60.5%
Q3 adjusted EPS grew by 60.5% YoY and 23.8% QoQ to $1.30, beating estimates by 3.5%. GAAP EPS grew by 66.7% YoY to $1.30, beating estimates by 8.5%. GAAP EPS included $0.06 in gains in non-marketable and publicly held equity securities.
- Analysts expect adjusted EPS to grow 61.2% YoY to $1.43 in Q4 and accelerate to 89.5% YoY growth to $1.53 in Q1.
- Looking forward, analysts expect FY2027 adjusted EPS to grow 49.5% YoY to $6.83 and 26.7% YoY to $8.65 in FY2028.
Cash and Balance Sheet
The company has a strong balance sheet with solid cash flows primarily driven by strong revenue and profits.
- Q3 operating cash flow grew by 34.7% YoY to $23.75 billion with an operating cash flow margin of 41.7%, compared to 50.3% in the same period last year and 32.8% in Q2.
- Q3 free cash flows grew by 31.6% YoY to $22.09 billion with a free cash flow margin of 38.7%, compared to 47.9% in the same period last year and 28.8% in Q2.
- The company’s cash and marketable securities have been steadily increasing and were $60.6 billion at the end of Q3, up from $56.8 billion in Q2 and $38.5 billion in the same period last year. Debt remained constant at $8.47 billion for Q3 and Q2.
- The company returned $12.7 billion to shareholders in the third quarter through $12.5 billion of share repurchases and $243 million of cash dividends. The company expects to continue to use its strong future cash flows to buy back shares and invest in AI growth opportunities.
- Inventories grew by 32% sequentially to $19.78 billion to support strong revenue growth.
Conclusion:
The $20 trillion prediction came from looking at the original $10 trillion prediction ahead of earnings and realizing I’d have to bump this up given the commentary around the $500 billion from the Blackwell-Rubin cycle. Although we had already slated next year for a $300 billion run rate and a $75 billion quarterly data center segment, it helped to hear last month that management agrees this is possible. Where the disconnect happens with analyst estimates is what will happen after next year as this is where analyst estimates show minimal growth through 2030 revenue with $437 billion whereas I am calling for double that by 2030. While Blackwell Ultra gets us to a new milestone of $50 billion to $75 billion quarterly revenue, quite a bit of my thesis depends on Vera Rubin, Rubin Ultra and the Feynman generations.
Although the next five years will be a marathon, the Q3 report is a step in the right direction. I don’t expect consistent QoQ growth every quarter, yet as stated, we are already exceeding my CAGR for next year in two quarters’ time. Crunching these numbers matters quite a bit as we are talking about the world’s most valuable company and Nvidia will have to put up consistent growth for the stock to inch upward. The days of a sudden spike in the stock price are likely behind us, yet if Nvidia remains consistent, the incessant market narratives will eventually tire.
Overall, this was an excellent report — and after two years of dissecting every angle of Blackwell, I’m excited to finally shift coverage to the Rubin generation of GPUs. Woohoo! This now marks the fourth GPU generation I’ve retired for I/O Fund Members — and we’re officially moving on to the fifth.
I/O Fund Equity Analysts Damien Robbins and Royston Roche contributed to this analysis.
Please note: The I/O Fund conducts research and draws conclusions for the company’s portfolio. We then share that information with our readers and offer real-time trade notifications. This is not a guarantee of a stock’s performance and it is not financial advice. Please consult your personal financial advisor before buying any stock in the companies mentioned in this analysis. Beth Kindig and the I/O Fund own shares in NVDA at the time of writing and may own stocks pictured in the charts.
Recommended Reading: