AI and Chip Stocks: Stock Market Nervous Ahead of Nvidia's Quarterly Results

Contents

Will it continue to go downhill or soon head downwards for artificial intelligence? Unfortunately, AI cannot see into the future either, and so global stock markets are currently nervous. Because this evening, local time, Nvidia's Jensen Huang will announce the results for the world's most valuable company for the third calendar quarter of 2025. For Nvidia itself, it is the third quarter of fiscal year 2026.

Actually, the market for AI data centers, where Nvidia generates the lion's share of its revenue, looks excellent. Numerous large AI projects are under construction worldwide, now even in Germany.

But there are reports that analysts and some investors are reacting nervously to

Prominent Non-Investors Anymore

Yesterday it became known that prominent investor Peter Thiel has sold all his Nvidia shares. This could be classified as taking profits in itself, or one could speculate that he needs a lot of money for other projects.

However, just a few days earlier, SoftBank CEO Masayoshi Son had also completely exited Nvidia. Son is a relevant business partner for Nvidia because he is involved in the gigantic AI project Stargate in the USA. Furthermore, Son had already sold all of SoftBank's Nvidia shares in 2019.

Further Pessimism

Further doubts about continued AI growth were raised by none other than the CEO of Google and Alphabet, Sundar Pichai. In an interview with the BBC, he stated that the current AI boom also has “irrational elements.”

And analysts from the investment bank J.P. Morgan had published a calculation that increases the pressure on the AI industry to finally demonstrate larger revenues. According to this, J.P. Morgan's model assumes that AI would have to generate global revenues of around 650 billion US dollars annually by 2030 to deliver an approximately 10 percent return on the gigantic investments.

Stronger Competition

Some experts in the AI hardware industry also believe that Nvidia's computing accelerators, which have so far dominated, especially in generative AI models, are increasingly facing competitive pressure. For example, more and more AI-inferencing benchmarks for locally executed Large Language Models (LLMs) are also delivering decent results with chips from AMD, Apple, and Intel.

However, AI computing accelerators for data centers from Nvidia's competitors are also catching up. AMD, for instance, has raised great expectations for its chip generation coming in 2026 and the “Helios” racks equipped with them. Google's latest TPU generation and Amazon AWS's in-house AI chips are also said to be performing well.

Furthermore, it is speculated that the important AI company OpenAI might rely less on Nvidia chips in the future. Although OpenAI maintains a close partnership with Nvidia, also cooperates with Broadcom in the development of customized AI accelerators.

Broadcom is also considered a development partner of Google for their TPU chips. Amazon AWS, in turn, is said to be cooperating with Marvell on its in-house AI chips, Trainium and Inferentia.

However, more competition for Nvidia would not indicate that the AI hype is overblown. Rather, more choice in AI hardware and consequently tending to falling prices would be positive signals for AI development.

(ciw)