NVIDIA unveils a new "money-making weapon": Sovereign AI is in full swing, with strong demand for government GPUs

According to the financial news app Zhitong Finance, executives from NVIDIA stated at the performance conference on Wednesday local time that as governments around the world are accelerating the construction of various artificial intelligence large models in their own languages, they are rushing to purchase NVIDIA AI GPUs. NVIDIA's Chief Financial Officer (CFO) mentioned that with generative AI becoming the central stage for enterprises and government departments, this further drives the already strong demand for AI GPUs. Despite NVIDIA's Q3 revenue outlook not exceeding the market's highest expectations leading to a post-market decline in stock price, many analysts still support the future prospects of NVIDIA (NVDA.US) stock price and emphasize that government departments joining the GPU purchasing legion will continue to boost NVIDIA's performance significantly.

NVIDIA predicts that its AI GPU sales will continue to grow significantly in the third quarter, as NVIDIA's AI GPU, the core AI infrastructure, provides essential driving force for popular generative AI tools such as OpenAI's ChatGPT and Anthropic's Claud. Although NVIDIA's latest revenue forecast did not meet investors' "sky-high" expectations, the company stated that new customers are spreading from the U.S. domestically to around the world, including global government departments seeking to build AI large models based on their own languages. They are intensifying hardware purchases to enhance independent underlying infrastructure to support generative AI blueprints.

After the financial report was released, NVIDIA's CFO Colette Kress stated at the performance conference with analysts that governments building AI applications and AI large models in their own languages are expected to contribute "billions of dollars" in revenue to NVIDIA's overall performance by the end of the fiscal year ending in January 2025. It is reported that this latest expectation significantly exceeds the earlier prediction by NVIDIA executives that such revenue would contribute "tens of billions of dollars" to NVIDIA's performance.

Sovereign AI - Strong Driver for NVIDIA's Future Performance Growth

NVIDIA's CFO Kress stated at the performance conference, "Countries around the world are eager to have their own generative artificial intelligence, especially hoping to integrate their own language, culture, and national big data AI large models." She described expertise in the field of artificial intelligence and core infrastructure such as AI GPUs as "a top priority for the nation."

She cited the example of the National Institute of Advanced Industrial Science and Technology in Japan, stating that the institute is building an artificial intelligence supercomputer with thousands of NVIDIA H200 AI GPUs as the core artificial intelligence infrastructure.

Undoubtedly, governments worldwide are using artificial intelligence as a measure to strengthen national security.

Sean Liao, a semiconductor analyst at the renowned research firm IDC, stated, "Large AI models are trained on data, and for political entities - especially at the national level, some of their data is confidential, which requires their large models to be tailored to their unique political, economic, cultural, and scientific and technological needs." "Therefore, they need to have their own AI large models and customized independent underlying hardware and software configurations, rather than outsourcing hardware infrastructure and technology to cloud giants." In addition, some large technology companies have been working hard to leverage government support and incentive measures to build AI platforms in regional languages.

IBM stated in May that the Saudi Data and Artificial Intelligence Authority will use the company's AI development platform Watsonx to train its "ALLaM" Arabic language model.

Bob O'Donnell, Chief Analyst at TECHnalysis Research, mentioned that besides large cloud service providers like Microsoft making large-scale purchases of the company's AI hardware infrastructure, various governments hoping to create their own large AI models and independent underlying hardware systems are expected to bring new growth opportunities for NVIDIA's AI GPUs.

It is understood that NVIDIA CEO Jensen Huang has mentioned "sovereign AI" multiple times this year, implying a surge in demand for national-level AI hardware. Therefore, in addition to providing AI GPUs to super-scale cloud computing service providers such as Microsoft, Amazon, and Google, NVIDIA is now also striving for business diversification. "Sovereign AI" is undoubtedly one of Jensen Huang's most valued areas.

As early as February this year, NVIDIA CEO Jensen Huang mentioned in a media interview that countries around the world are planning to establish and operate their own AI infrastructure domestically, which will significantly increase the demand for NVIDIA hardware products.

During the interview, Jensen Huang stated that countries including India, Japan, France, and Canada are discussing the importance of investing in "sovereign AI capabilities." "From their natural resources to data, they are seen as something that should be refined and produced for their countries. Recognition of sovereign AI capabilities is a global concept."

With the unmatched strong moat formed by the "CUDA software-hardware collaborative platform + high-performance AI GPU" and the AI developer ecosystem, NVIDIA's highly demanded Hopper architecture AI GPU and the upcoming mass delivery of Blackwell architecture AI GPU have undoubtedly become the preferred AI underlying hardware for governments worldwide.

"Buy on Dips" Legion is Ready with Bullets

In terms of overall performance expectations, NVIDIA expects third-quarter revenue as of October to be around $32.5 billion, indicating a growth of about 80% compared to the same period last year, higher than the analyst average estimate of $31.7 billion. However, the market's highest expectation for NVIDIA's Q3 revenue reaching an astonishing $37.9 billion has raised concerns about its explosive growth slowing down to some extent, prompting a post-market price drop of up to 8%. But before the Thursday pre-market session, buying following the "buy on dips" strategy gradually entered the market, narrowing NVIDIA's stock price decline from the post-market drop of 8% to around 2% currently.

Concerns about how long the massive spending on artificial intelligence will continue and when AI monetization will be achieved, as well as concerns about the demand for and shipment schedule of Blackwell chips, weighed down NVIDIA's stock price during the trough in early August. However, the rebound since then indicates that investors are still willing to buy the stock on dips, as the frenzy of large technology companies' AI infrastructure deployment shows no signs of stopping Despite concerns that the scale of AI infrastructure being deployed may exceed actual demand, with major customers such as Google, Microsoft, and Meta continuing to increase their capital expenditure budgets in this financial quarter, there are worries that an AI bubble may burst.

However, Jensen Huang insisted during the earnings conference that the significant investments made by tech giants in purchasing AI GPUs mark just the beginning of a new era of technology and economics.

On the eve of NVIDIA's financial report, Wall Street giant Morgan Stanley advised investors who are fond of tech stocks not to give up on AI-related stocks even if NVIDIA's financial report or next quarter's performance guidance disappoint. Morgan Stanley believes that the stock prices of AI-related companies have already readjusted after the August tech stock crash and are far from reaching the extreme valuation levels seen during the bursting of the internet bubble.

NVIDIA's valuation also experienced a significant contraction following the tech stock crash in August, with the stock's forward price-to-earnings ratio at around 38x, lower than the 44x in June and over 60x last year. The forward price-to-earnings ratio of the S&P 500 index is around 21x.

Dan Ives, a senior analyst at Wedbush Securities, stated after NVIDIA's latest earnings announcement that any selling pressure on NVIDIA is temporary. "I think this stock is actually going higher, investors are very focused on NVIDIA. There may be some crazy reactions ahead. But overall, we believe NVIDIA's financial performance can be rated A+."

Ives also mentioned that it is difficult to say that NVIDIA's performance and outlook for the next quarter are bad, but investors have become accustomed to the company not only exceeding expectations but also crushing them in all aspects. "In the past three quarters, NVIDIA's revenue growth has consistently exceeded 200%. All of this has led investors to set their expectations higher, raising the bar for NVIDIA to beat expectations."