Investors are eager to see if the high demand for artificial intelligence (AI) products can help make up for the decline in global sales for computer hardware. Nvidia, the world’s most valuable chipmaker, is set to release its quarterly results on Wednesday, and its success may hinge on the increasing demand for its processors used in training large language models.

Nvidia reported in its previous earnings that revenue would be driven up by nearly two-thirds and earnings per share would quadruple in the three-month period ending in July due to the demand for its processors for AI applications. In fact, the company plans to triple the production of its top H100 AI processor in 2024, compared to this year’s expectations.

The demand for Nvidia’s chips is so high that AI processors are already sold out through 2024. This massive thirst for chips is unfortunately impacting the broader market for computing equipment, as businesses prioritize investment in AI over general-purpose servers. This shift is causing a decline in overall server revenues for companies like Foxconn and Lenovo, as they struggle to keep up with the demand for AI servers.

Cloud service providers like Microsoft, Amazon, and Google, who dominate the global server market, are redirecting their focus towards building AI infrastructure. However, this comes at a cost as the weak economic environment and GPU supply constraints force them to cannibalize their spending on other areas.

Although the demand for AI server chips is projected to grow by nearly 50% annually for the next five years, the overall tech slump caused by the economic downturn presents a downward pressure. The global CSP capital expenditure is expected to grow by only 8% this year, compared to the nearly 25% growth seen in 2022.

Analysts attribute the delays and challenges in meeting the demand for AI servers to supply chain bottlenecks. These bottlenecks include a capacity shortage in advanced packaging and high-bandwidth memory (HBM), both of which are essential for AI processor production. Efforts to increase capacity are underway, but the situation may not improve until the end of 2024.

In addition to these challenges, Chinese companies face export controls on Nvidia’s H100, making the less powerful H800 a more affordable alternative. However, even when the shortages and challenges are resolved, companies in the server supply chain could benefit massively. Shipments of servers for AI algorithms are expected to triple next year, and the share of AI servers in the overall market is projected to reach 20% by 2027.

The shift towards AI servers will bring multiples of everything in the supply chain. The demand for baseboards, cooling systems, power supplies, and larger racks will all increase significantly compared to general-purpose servers. Companies like Foxconn and its affiliate, Foxconn Industrial Internet, are well-positioned to capitalize on this shift, as they offer a comprehensive range of components and assembly services.

Ultimately, the success and resilience of the computer hardware market may rely on the booming demand for AI products. As businesses increasingly invest in AI infrastructure, the market for AI servers is projected to grow significantly in the coming years.

Sources:
– [source 1]
– [source 2]
– [source 3]

The post Can Artificial Intelligence Demand Offset Slump in Computer Hardware Sales? appeared first on TS2 SPACE.