As long as the application side continues to generate demand, the AI infrastructure chain is far from coming to an end.

Written by: Jim, McTong MSX

Google and Nvidia, the applications and underlying entry points of the AI sector, have submitted their reports this week.

If the Google I/O discussed the imagination space of AI applications, then Nvidia's earnings report verifies whether the computational power demand behind these imaginations has been realized.

After the market closed on May 20, Eastern Time, Nvidia announced its FY2027 Q1 earnings report, with revenue reaching $81.615 billion, a year-over-year increase of 85% and a quarter-over-quarter increase of 20%; data center revenue reached $75.2 billion, a year-over-year increase of 92% and a quarter-over-quarter increase of 21%; meanwhile, Nvidia announced an additional $80 billion stock repurchase authorization and raised the quarterly cash dividend from $0.01 per share to $0.25 per share.

This set of data is strong enough, but what the market is really concerned about is not whether "Nvidia is still growing," but whether, under high market expectations, it can continue to prove that the AI main line is still intact, the demand for computational power has not peaked, and Nvidia's pricing power remains strong?

1. An Overview of Revenue, Guidance, and Gross Margin: Is the AI Engine Still Accelerating?

First, it is important to clarify that Nvidia's core business is no longer the traditional "graphics card," but data centers, which are the computational infrastructure behind AI factories.

This quarter, Nvidia's data center revenue reached $75.2 billion, accounting for over 92% of total revenue. Breaking it down, according to the old business classification, data center computing revenue was $60.4 billion, a year-over-year increase of 77%; data center networking revenue reached $14.8 billion, a year-over-year increase of 199%, also hitting a historical high.

This indicates a key issue, that is, AI demand is not just limited to GPU single points, but is expanding to the entire AI infrastructure—where GPUs are responsible for computation, networks connect computational power, and whole rack systems, NVLink, InfiniBand, Ethernet, optical communications, power, and cooling will all be part of the AI factory.

Therefore, the significance of this data center revenue is not just "Nvidia sells a lot," but indicates that global cloud vendors, AI model companies, enterprise customers, and sovereign AI have not noticeably reduced their investment in computational power. From this perspective, if future data center revenue continues to exceed expectations, the risk appetite for the AI hardware chain is likely to continue to spread; but if this metric starts to fall below expectations, the market will genuinely worry about the peaking of AI capital expenditures.

Of course, besides revenue, for a high-expectation company like Nvidia, the stock price post-earnings announcement often does not just look at this quarter's numbers, but should focus more on the guidance for the next quarter.

Nvidia's guidance for FY2027 Q2 revenue is $91 billion (with a 2% fluctuation), clearly higher than the approximately $86 billion to $87 billion range that the market generally anticipated before the earnings report. The company also clearly stated that this guidance does not assume any revenue from data center computing in China. This is significant because if the guidance still reaches $91 billion without including data center computing revenue from China, it shows that overseas cloud vendors, AI factories, enterprise AI, and demand from other regions are sufficient to continue supporting high growth.

In other words, what the market originally worried about was whether Nvidia's growth had been too fast, making it hard to exceed expectations in the future. But this guidance signals that at least for the next quarter, the demand for AI computational power has still not shown any obvious slowdown.

However, it is also important to note that as market expectations rise, Nvidia needs to deliver not just "good earnings," but "earnings that are significantly better than expected," so whether the stock price will surge in the short term still depends on whether investors believe this guidance is sufficient to cover its high valuation.

Meanwhile, Nvidia's high valuation comes not only from high revenue growth but also from its strong profitability.

This quarter, Nvidia's GAAP gross margin was 74.9%, and Non-GAAP gross margin was 75.0%. The company also guided the gross margin for the next quarter to be 74.9% for GAAP and 75.0% for Non-GAAP, with a fluctuation of 50 basis points.

This indicates that, although the Blackwell system, HBM, advanced packaging, and whole rack solutions will bring higher costs, Nvidia can still keep its gross margin stable around 75%. For the market, this undoubtedly represents two things:

• Nvidia still has strong pricing power. Customers are not just buying a chip; they are purchasing complete platform capabilities;
• Although competition in AI chips has intensified, it has not yet significantly compressed Nvidia's profit margins. Google TPU, Amazon Trainium, AMD GPU, and self-developed ASIC chips will bring competition, but at least from this earnings report, Nvidia's profitability has not been significantly shaken;

Of course, if in the future the gross margin falls significantly below 74%, the market will start to worry about product switching costs, client bargaining power, and the pressure from alternative solutions, which needs to be monitored over the long term.

2. Is Nvidia Starting to Transition to an "AI Cash Flow Platform"?

A very noteworthy change in this earnings report is the shareholder returns.

In Q1, Nvidia returned approximately $20 billion to shareholders, including stock repurchases and cash dividends. By the end of Q1, the company still had $38.5 billion remaining from its original repurchase authorization, and then the board approved an additional $80 billion stock repurchase authorization, and increased the quarterly dividend from $0.01 per share to $0.25 per share.

The significance behind this is not just that the company has cash on its books, but more importantly, Nvidia is sending a positive signal to the market, indicating that the AI dividends will not only be invested in ecological partners, AI startups, and supply chains, but will also start returning to shareholders.

After all, in the past, the market was concerned that Nvidia's substantial investment in AI ecosystem partners like OpenAI and Anthropic could be "circular financing," but if the company simultaneously increases buybacks and dividends, it can partially alleviate the long-term capital concern regarding capital allocation efficiency.

This also makes Nvidia gradually take on characteristics of an "AI cash flow platform" rather than just a purely high-growth AI stock.

3. What is the Market Looking For After Blackwell?

Another point of interest for Nvidia is whether the product cycle can continue.

This quarter, Nvidia emphasized the Vera Rubin platform, which includes Vera CPUs, BlueField-4 STX, and other products, and mentioned its collaboration with Google Cloud, including Vera Rubin-driven A5X instances and previews of the Google Gemini model on NVIDIA Blackwell and Blackwell Ultra GPUs.

This shows that Nvidia has not stopped the narrative with Blackwell, but is paving the way for the next-generation platform in advance.

For investors, this is important because if Blackwell is just a strong cycle, the market will worry about the growth downturn after the peak; but if Vera Rubin can smoothly connect, Nvidia will not just rely on a single generation of products for explosive growth, but will have ongoing platform iteration capabilities.

As for whether Google TPU and CPU pose a threat to Nvidia, I believe this should be viewed in two layers.

In the short term, TPUs, ASICs, and CPUs will indeed take on more tasks in some scenarios, especially for large companies' in-house models and inference workloads, but in the medium term, this looks more like a coexistence of multiple routes due to excessive AI demand rather than an immediate replacement of Nvidia.

Nvidia's true advantage lies not just in the GPU itself, but in the "platform capabilities" formed by the combination of GPUs, CPUs, networks, software, rack systems, and ecological partners. As long as customers need to quickly deploy large-scale AI factories, Nvidia still occupies a core position in the supply chain.

In Conclusion

This earnings report at least proves one thing: the AI main line is still intact.

Data center revenue continues to break records, next quarter's guidance exceeds expectations, gross margins are held around 75%, repurchases and dividends are significantly increased, and the product cycle extends from Blackwell to Vera Rubin; all of these indicate that Nvidia continues to stand at the core position of AI infrastructure expansion.

However, for the stock price, the question is not "how good the earnings report is," but "whether it is good enough to exceed the already high market expectations." If the market perceives this earnings report as merely validating expectations, short-term volatility may occur; if investors further raise AI capital expenditures and Nvidia's long-term revenue potential, then the AI chain still has the possibility of continuing to expand.

Moreover, from the perspective of the supply chain, Nvidia's strong earnings not only impact NVDA.M alone, but also prompt the market to reassess the entire AI infrastructure chain:

• ASIC / Manufacturing / HBM: AVGO.M, TSM.M, MU.M
• Networking: ANET.M, MRVL.M, CRDO.M
• Optical Communications: COHR.M, LITE.M, AAOI.M
• Power and Cooling: VRT.M, ETN.M, MPWR.M

Of course, if we trace it back from the entry point, since these days' Google I/O has proven that AI applications are still expanding, it is easy to understand why the computational power demand in Nvidia's earnings report is still being realized—as long as the application side continues to generate demand, the AI infrastructure chain is far from coming to an end.