"Sellers of shovels" are going to make a fortune.
Technology giants are engaged in a fierce "arms race" in AI, and more chip manufacturers, including NVIDIA, are benefiting from it.
On Tuesday, AMD released its second-quarter earnings report, and during the earnings call, AMD CEO Lisa Su stated that "the number of AMD's AI data center chip customers has grown by more than seven times this quarter."
According to the financial report, AMD's revenue and profit both declined in the second quarter of this year, with a YoY decrease of 18% in revenue and a YoY decrease of 44% in adjusted net profit. However, the decline was not as significant as expected by Wall Street.
At the same time, AMD expects that its data center and embedded business will grow throughout the fiscal year.
During the earnings call, Lisa Su stated that they anticipate significant growth in the data center business in the second half of the year, with a focus on the fourth quarter, where they expect a 50% increase in performance for this business.
With multiple customers initiating or expanding projects supporting the future large-scale deployment of AMD Instinct MI250 and MI300 software and hardware, the number of AI customers has grown by more than seven times.
She mentioned that AMD is increasing its R&D investment in AI and has developed an AI strategy, including the development of AI-specific chips and software, with the goal of making AI a "significant growth driver" for the company. She also mentioned that AI could potentially help boost PC chip sales.
In addition to AMD, another chip manufacturer, Intel, is also expanding its presence in AI.
Last week, during Intel's second-quarter earnings call, CEO Gelsinger stated that the company's investment in AI accelerators has exceeded $1 billion, and in the second quarter alone, the scale of this product line has expanded sixfold.
In addition to their own AI accelerator product, Gaudi2, Intel can also benefit from the increasing demand for AI accelerators through their foundry business, manufacturing custom AI chips for large cloud companies like Alphabet and Amazon.
Intel has previously stated that part of its 18A manufacturing process foundry line is dedicated to companies designing their own AI accelerators.
As an established chip manufacturer, Intel has faced challenges due to declining PC demand, but it has returned to profitability in the second quarter, indicating that the worst period may be over.
NVIDIA: Quarterly target achievable with four customers
As the industry leader, NVIDIA's performance expectations are naturally optimistic.
On Monday, analysts at Bernstein pointed out that only four companies, Microsoft, Meta, ByteDance, and Google, are enough to help NVIDIA achieve its guidance for this quarter. They also expect other large tech companies and AI startups to have a more positive demand for NVIDIA. Bernstein analysts said that the current factor limiting Nvidia's revenue is the advanced chip packaging capability of semiconductor foundries. With the improvement of packaging capability next year, Nvidia is expected to generate $75 billion to $90 billion in data center and AI chip revenue in 2024, while the consensus expectation of Wall Street analysts is only $42 billion.
The Bernstein team believes that Nvidia's performance will continue to grow and may continue to rise for a considerable period of time, at least in the next 12-18 months. This analysis reflects an optimistic outlook for Nvidia's future development.
Previously, Mizuho analyst Vijay Rakesh also pointed out that AI has brought Nvidia an unimaginable upside potential in stock price, and raised Nvidia's target price from $400 to $530.