AI (Artificial Intelligence) chip giant Nvidia is confident that the company is at the starting point of the next wave of growth.
On May 22nd local time, after the release of the fully unexpected first quarter results for 2025, Nvidia CEO Huang Renxun and CFO Colette Kress released a lot of "hot news" during the performance conference, including the launch plan of the new chip Blackwell, explanations for possible "window periods" between product cycles, and the company's development after Blackwell.
Regarding the performance of the fiscal quarter, Huang Renxun said: "The next industrial revolution has begun... In addition to cloud service providers, generative AI has also expanded to customers in consumer Internet companies, enterprises, sovereign AI, automobiles and healthcare, creating multiple vertical markets worth billions of dollars. We are at the beginning of the next wave of growth."
After the financial report was released, Nvidia's stock price rose by over 7% after the market, breaking through $1000 per share. On the 22nd, NVDA's stock price fell 0.46% to close at $949.50 per share, with a total market value of $2.34 trillion. According to Wind data, Nvidia's stock price has risen by over 230% in the past 2023, and has risen by 91.74% so far this year.
Blackwell will bring in significant revenue this year
For Nvidia's unexpected performance this time, the market generally believes that Nvidia, as a representative of the AI craze, looks unstoppable and shows no signs of slowing growth.
Ryan Detrick, Chief Market Strategist at Carson Group, said, "Despite facing heavy expectations, Nvidia has once again stood up and delivered excellent results. The company's historically important data center performance has been strong, and its guidance for the future is equally impressive."
At the performance meeting, Huang Renxun pointed out that the workload of inference is "significantly increasing", electronic computing is "switching from information retrieval to production skills", and the Blackwell chip launched by the company in March this year is already in "full production".
Huang Renxun has provided a detailed timeline for the supply of Blackwell chips, and it is expected that the sales of new chips will be reflected in the data center business in the fourth quarter of the 2025 fiscal year: "Blackwell chips will be shipped in the second quarter of this year, increased production in the third quarter, and deployed to data centers in the fourth quarter... It is expected that Blackwell will bring a large amount of revenue to the company this year.".
However, Huang Renxun also acknowledges that customers are "putting a lot of pressure on the company" regarding delivery issues. Large cloud service providers are the most important customers of Nvidia data centers, including Amazon Cloud Services AWS, Microsoft Azure, Google Cloud, and Oracle Cloud. Nvidia stated that these cloud service providers accounted for 40% -45% of the company's record high data center revenue in the first quarter, reaching $22.56 billion.
Kress pointed out that cloud service providers have a "rapid and significant" return on investment in Nvidia hardware. She said that for every dollar invested by cloud service providers in Nvidia hardware, they can rent it out and earn $5 in revenue over the next four years.
In addition to large manufacturers, demand from startups is also very strong. Huang Renxun said, "Currently, there are approximately 15000 to 20000 generative AI startups waiting to become customers of Nvidia in various fields, using Nvidia's devices to train models."
Huang Renxun also mentioned that the demand for GPUs in the automotive market is still soaring, and car companies will become the "largest vertical enterprise in data centers this year.".
Previously, there were rumors that Amazon Cloud Service AWS would suspend purchasing Hopper chips while waiting for Blackwell chips, causing some investors to worry that demand for Nvidia products would decline between two product cycles. Subsequently, AWS responded that the company has not suspended any chip orders from Nvidia.
Regarding this issue, Huang Renxun emphasized that it is expected that the market demand for the company's previous generation Hopper chips will continue to increase this quarter. Blackwell chips can be "backward compatible" with systems that already use Hopper, allowing customers to smoothly transition: "Every customer is eager to bring their infrastructure online, they want to save money and make money, and they want to achieve this goal as soon as possible."
Huang Renxun added that in order to launch the new platform, Nvidia has been helping customer ecosystems prepare them for using liquid cooling systems.
Kress also pointed out that the demand for GH200 chips and Blackwell chips far exceeds the supply, and this situation may continue until next year. Although the supply of chip GH100 released in 2022 has improved, the supply of GH200 released by the company in August last year is still in short supply.
Implementing annual chip upgrades, Nvidia continues to accelerate
Wall Street expects Nvidia to drive the entire US stock market on its own and inject new momentum into the already hot AI concept.
Since the beginning of this year, Nvidia's stock price has performed significantly better than the S&P 500 index and NASDAQ 100 index, with a weight of over 5% in the S&P 500 index and 6.5% in the NASDAQ 100 index.
Jay Woods, Chief Global Strategist at Freedom Capital Markets, a financial institution, pointed out, "Few stocks can have such a huge impact on the overall market. But Nvidia has already won this position."
What's even more surprising is that Huang Renxun revealed that there are new chips after Blackwell, and the company's technology research and development pace is "very fast", able to achieve "one new chip per year". Huang Renxun repeatedly emphasized at the meeting, "We are accelerating.".
Undoubtedly, this reflects the strong demand in the AI chip market. Previously, Nvidia's chip architecture update cycle was approximately every two years: the Ampere architecture was launched in 2020, the Hopper architecture was launched in 2022, and the Blackwell architecture was launched in 2024.
Previously, Guo Minggui, an analyst at Tianfeng International, predicted that Nvidia's next generation AI chip, the Rubin series/R100, is expected to begin mass production in the fourth quarter of 2025.
Not only AI chips, Nvidia has also shown greater ambition. Huang Renxun stated that while accelerating the development of AI chips, Nvidia will also accelerate the production of all other types of chips: "We will comprehensively advance at a very fast speed. A large number of chips are on the way, including new CPUs, new GPUs, new network cards, new switches..."
At the meeting, Nvidia also revealed its outlook for the future of the AI field, stating that sovereign AI and new multimodal AI capabilities will become driving factors for the company's further development.
Previously, Huang Renxun emphasized the importance of "sovereign AI capabilities" in multiple interviews. According to NVIDIA, sovereign AI refers to the ability of a country to produce AI using its own infrastructure, data, labor, and business networks. Nvidia believes that as data security issues become the focus, sovereign AI will become more attractive.
Kress said, "The importance of AI has attracted the attention of every country. As countries around the world deploy sovereign AI, data center revenue continues to diversify... We believe that this year's revenue from sovereign AI will rise from zero revenue last year to a high of billions of dollars."
Multimodal AI, which can generate various types of data such as text, photos, videos, and audio, has also been a hot topic in recent years. In recent weeks, OpenAI、 Google and Microsoft have successively released a series of innovative achievements in multimodal AI, fully reflecting the fierce competition in the AI field. As an AI arms dealer, Nvidia has also been highly sought after by major technology giants.
When an analyst asked about the chips that the company had previously launched into the Chinese market, Huang Renxun said, "Our business in China is indeed significantly lower than in the past. Due to limitations on our technology, competition in China is becoming more intense now."
According to the Global Times, the US Department of Commerce issued a series of new chip export restrictions on October 17, 2023, expanding the definition of advanced artificial intelligence chips and imposing additional licensing requirements on chip products shipped to over 40 countries to avoid resale to China. The new restrictive measures are scheduled to take effect on November 16, 2023.
Foreign Ministry spokesperson Mao Ning stated at a regular press conference on January 8th this year that the United States has continuously increased its chip export control measures against China under the pretext of "national security", unreasonably suppressing Chinese semiconductor companies, and is a genuine economic bullying act. "The facts clearly indicate that the US is deliberately suppressing the development of China's chip industry not for the sake of 'national security', nor is it any legitimate competition behavior, but rather a unilateral bullying behavior without any principles or bottom line, depriving emerging markets and developing countries of the right to pursue a happy life." Mao Ning said that the US's actions seriously undermine the stability of the international supply chain, poison the atmosphere of international cooperation, and promote division and confrontation. This selfish approach is bound to lift a stone and hit oneself in the foot.