Summary of Key Highlights
NVIDIA delivered an impressive financial report for the second quarter of its 2027 fiscal year: revenue of $96.2 billion (doubling year-over-year) and net profit of $59.6 billion (a 126% increase), setting new record highs. The growth was driven primarily by its data center business, which accounted for over 90% of total revenue, especially due to the surge in demand from AI cloud services and corporate clients. NVIDIA has shifted from selling individual chips to offering a comprehensive “chip + software + system” solution. The company has also begun to play a key role in AI infrastructure development, participating in large-scale infrastructure investments. However, market attention has shifted: investors are now concerned about the sustainability of the AI infrastructure boom and whether NVIDIA’s next-generation products will be able to maintain its lead. The stock price reacted with initial declines followed by gains after the report was released, reflecting these divergent views.
Detailed Analysis
1. Financial Performance: How impressive is the growth? How strong is the profitability?
- Revenue Doubling: Revenue soared to $96.2 billion, compared to $46.7 billion in the same period last year, almost doubling in just one year, generating approximately $650 million per day (equivalent to the annual profit of a mid-sized company).
- Even More impressive Profits: Net profit increased by 126% to $59.6 billion, with earnings per share rising by 128% to $2.46. The gross margin was 75%, meaning that for every $100 in sales, NVIDIA retained $75 in profit, which is higher than the previous year (72.4%) and the previous quarter (74.9%), indicating extremely high profitability.
- Shareholder Benefits: NVIDIA returned $26 billion to shareholders through share repurchases and dividends, with an additional $99 billion remaining for repurchases, essentially handing out a substantial bonus to shareholders.
- Stock Price Volatility: The stock price fell by 1.3% before rising by 4% after the report. Investors are no longer solely focused on whether the results exceeded expectations; they are more concerned about the longevity of the AI infrastructure trend and the competitiveness of NVIDIA’s next-generation products, leading to this mixed reaction.
2. The Driving Force Behind Growth: Why the Data Center Business Accounts for 90% of Revenue?
NVIDIA’s growth is largely due to the soaring demand for AI computing power:
- Data Center Revenue: Revenue from data centers reached $89 billion, accounting for 92% of total sales, with a year-over-year increase of 117%.
- Diverse Customer Base:
- Large-scale Customers: Revenue from major cloud providers like AWS and Google Cloud increased by 102%.
- ACIE Customers (including AI cloud services, enterprises, governments, and sovereign AI initiatives): Revenue increased by 138%, indicating that AI demand is spreading from a few large cloud providers to a broader range of organizations, including governments and countries that are building their own AI infrastructure.
- China Market Impact: Revenue from Hopper chips shipped to mainland China accounted for less than 1% and was not included in the third-quarter forecasts. This segment may impact future growth, but its current contribution is still small.
- Jeff Huang’s Perspective: “Computing is revenue.” AI model training and inference require extensive GPU usage, and whoever possesses the necessary computing power can profit. NVIDIA’s chips are essentially the “electricity” of the AI era.
3. Product Evolution: Moving from Chip Sales to Comprehensive AI Solutions
NVIDIA is no longer just selling GPU chips; it has transformed into an “AI platform provider”:
- Current Focus: The Blackwell Ultra chip is the main driver of growth, being widely adopted by large cloud providers and AI companies like OpenAI.
- Next Generation: The Vera Rubin platform has been fully deployed, offering not only GPUs but also a dedicated AI CPU for AI robots and assistants, as well as networking equipment and software systems, which are already in use with CoreWeave and Google Cloud.
- Inference Market Strategy: The Groq 3 LPX chip was introduced for AI inference tasks (such as answering questions in ChatGPT and image generation). As AI applications become more prevalent, the demand for inference will exceed that for training, making this a key area of competition.
- Software Ecosystem: The DSX platform helps customers build “AI factories” to manage large-scale computing clusters, and the Agent Toolkit simplifies the development of AI assistants, making it difficult for competitors to switch to other solutions.
4. NVIDIA’s New Role as an AI Infrastructure Leader
AI data centers require substantial resources such as land, power, and facilities, and NVIDIA is taking on the role of project manager:
- Future Investments: NVIDIA plans to invest $360 billion in various areas, including chip supply, cloud service agreements, capital expenditures, and equity investments. For example, it provided a guarantee of $105 billion for the SB Energy project in Ohio to build a 4.25GW computing facility that will house NVIDIA’s facilities for OpenAI.
- Collaborating with Investors: NVIDIA is working with institutions like Blackstone and Goldman Sachs to attract $500 billion in third-party capital for AI infrastructure development, positioning itself as a leader in this emerging industry.
- Financial Stability: Despite issuing $25 billion in bonds, NVIDIA still has $566 billion in cash. Jeff Huang emphasizes financial stability, noting that AI infrastructure is a long-term investment aimed at securing a monopoly in the future.
5. Future Challenges: Can Growth Continue? Competition is Looming
NVIDIA expects revenue to reach $108 billion in the third quarter, setting another record high, but there are several concerns:
- Declining Interest in AI Infrastructure: There are concerns that corporate and government investments in AI may slow down due to the high costs associated with building data centers.
- Intensifying Competition: AMD’s MI300 chips are competing in the data center market, and Google’s TPU (self-developed AI chips) are also challenging NVIDIA. Even major customers (such as Microsoft) are developing their own chips, posing a threat to NVIDIA’s market position.
- Inference Market Competition: The demand for inference chips is growing, but so are competitors. Whether NVIDIA’s Groq 3 LPX can outperform its competitors remains to be seen.
- China Market Impact: The lack of Chinese revenue in the third-quarter report suggests that if Chinese demand does not recover, it could affect NVIDIA’s growth.
In summary, NVIDIA is the dominant player in the AI computing power market, but to maintain its leadership, it will need to offer comprehensive platforms and invest in infrastructure. The AI industry is transitioning from a focus on chip production to the establishment of comprehensive ecosystems.