After Huang Renxun reiterated that the demand for Nvidia chips from users was unbelievably high, Nvidia's stock rose in response.
On September 12th local time, Nvidia CEO Renxun Huang gave a speech at the Goldman Sachs Communacopia technology conference, stating that the company's AI (artificial intelligence) chips are becoming the hottest commodity in the technology industry, and the limited supply has sparked fierce competition: "Our customers may be emotional now. This is understandable, and the situation is very tense. We are doing our best to respond
Huang Renxun's speech has had a strong boosting effect on the market. On the 12th, Nvidia's stock price closed at $116.91 per share, up 8.15%, the largest increase in six weeks, with a total market value of $2.87 trillion. Driven by NVIDIA, the semiconductor sector saw a general increase, with the Philadelphia Semiconductor Index rising over 5%, ARM rising over 10%, Broadcom rising over 8%, and TSMC rising over 4%.
In addition to Huang Renxun's uplifting remarks to investors, there were reports on the 12th that the US government was considering allowing Nvidia to export advanced chips to Saudi Arabia, further driving up the company's stock price.
Every $1 invested in AI infrastructure can be converted into $5
Taking the opportunity of this speech, Huang Renxun shared his positive outlook on the future of the data center market and AI infrastructure construction.
When it comes to the prospects of the data center market, Huang Renxun said that trillions of dollars worth of universal data centers worldwide need to embrace accelerated computing. "Dense" data centers will be more energy-efficient and cost-effective, while also facilitating liquid cooling processing: "So our top priority now is to accelerate and modernize data centers, increase density, and become more energy-efficient. This is what we will focus on for the next 10 years." At the same time, as the efficiency improvement speed of traditional CPUs slows down, Moore's Law is "approaching its end," and the market will see a surge in data computing costs, with accelerated computing bringing huge cost savings: "In the past 10 years, computing power has not only increased by 100 times, but by 1 million times
Huang Renxun believes that the AI revolution has created a completely new industry: "If you look back at the entire IT industry, until now, we have been manufacturing tools and instruments that people can use. And this time, what we want to create will be able to enhance people's skills. That's why people believe that AI will not be limited to trillions of dollars in data centers, but will also expand into the field of skills
During the phone conference after the latest financial report at the end of August, when asked about the ROI (return on investment) of AI capital expenditures, Huang Renxun's "promotional" response did not satisfy investors. In this speech, Huang Renxun provided a prepared 'standard answer'.
Huang Renxun explained that in the current first wave of AI, due to the surge in global demand for AI technology, AI infrastructure participants and all cloud service providers can receive "amazing returns". Every $1 invested in AI infrastructure can be converted into $5 worth of rent: "(After using accelerated computing) you can save a lot of computing time, although the computing cost will slightly increase, the overall return on investment is very considerable. This is the direct ROI brought by acceleration
In addition, AI also plays a significant role in improving productivity, especially in the field of software engineering: "The era where every line of code was written by software engineers is completely over. In the future, every software engineer will have a digital engineer as an assistant to assist them 24 hours a day. Our company has 32000 employees, but we hope to increase this number by 100 times through digital engineers
Nvidia can transfer TSMC OEM orders to other suppliers at any time
After sharing his positive outlook on the sustainability of the AI boom, Huang Renxun also introduced Nvidia's competitive advantages.
Huang Renxun emphasized that AI is not only about hardware, but also about infrastructure: "Building an AI computer is not simply assembling chips, but building a complete data center. For example, our Blackwell architecture consists of seven different types of chips, and Blackwell chips are just one of them After more than 30 years of persistence, Nvidia's GPU architecture is ubiquitous and has become the first choice for entrepreneurs.
Huang Renxun also revealed the secret behind Nvidia's high-speed innovation pace: "Our innovation pace is based on this fundamental approach - remember, our architecture contains seven different chips. The update cycle for each chip is about two years, and we can make a mid-term upgrade to them every year. If you can launch a new architecture every two years, it's like moving at the speed of light. Now, we have seven different chips that contribute to the overall performance of the architecture. Therefore, we can launch a better AI cluster or super cluster than the previous generation every year
In addition, Huang Renxun confidently stated that although its partner TSMC has a strong advantage in chip foundry, Nvidia can "transfer orders to other suppliers at any time" if necessary.
Huang Renxun pointed out, "Last year, Nvidia's revenue experienced explosive growth, which could not have been achieved without the support of the supply chain. TSMC and the supply chain's ability to respond quickly is incredible. In less than a year, we have significantly increased production capacity and will continue to expand next year... If we have to transfer from one factory to another, we have the ability to do so. Perhaps their process technology is not as good, perhaps we cannot achieve the same level of performance or cost, but we can maintain supply
Finally, Huang Renxun emphasized again that Nvidia will mass produce the latest chip Blackwell and expand production capacity in the fourth quarter of this year, and continue to expand production next year: "The demand for Blackwell is really huge, everyone wants to be the first company to have it, everyone wants to have the most production capacity, and everyone wants to be ahead
Huang Renxun said that Nvidia's advantageous position also makes him nervous: "We now have cooperation with every AI company in the world, with every data center in the world... With this comes enormous responsibility. We have many people's responsibilities on our shoulders, and everyone is counting on us. In this situation, the delivery of our components, technology, infrastructure, and software is really exciting because it will directly affect their revenue and directly impact their competitiveness." Huang Renxun also joked that "sleeping three hours a day is enough" for this.
Previously, Nvidia's second quarter financial report for the fiscal year 2025 ending on June 28 showed that the company achieved revenue of $30.04 billion during the period, a year-on-year increase of 122%, higher than market expectations; It is expected that the sales revenue for the third quarter of fiscal year 2025 will reach 32.5 billion US dollars, with a fluctuation of 2%. However, due to design flaws in the mask, the production of the latest chip Blackwell has been postponed to the fourth quarter of this year. After the release of the financial report, Nvidia's stock price plummeted by over 8% and continued to decline for the next 4 trading days.