AMD Nvidia AI chips rivalry widens as spending worries hit tech stocks

AMD Nvidia AI chips competition sharpened after Advanced Micro Devices announced new graphics processors, central processors and other chips for AI infrastructure, physical AI and robotics, while Nvidia continued to pitch its broader AI factory stack. The announcements matter because buyers are starting to look beyond single accelerators toward full systems for training, inference and agentic workloads, even as Nvidia remains the dominant supplier.

AMD’s push included next-generation infrastructure aimed at frontier models, agentic workloads and autonomous robots, according to SiliconANGLE coverage of the company’s Advancing AI event. Microsoft will use AMD’s AI-optimized Helios racks in Azure, and Anthropic plans to buy up to 2 gigawatts of GPU capacity from AMD, giving AMD more visible cloud and model-lab demand. Pricing, shipment timing and revenue impact were not disclosed in the roundup.

Nvidia also kept pressure on the market, showcasing performance gains for its Vera Rubin platform and emphasizing a full AI factory approach. theCUBE Research analyst Dave Vellante said AMD Chief Executive Lisa Su led a major turnaround at the company, but argued that competing in this phase of AI infrastructure will require a different kind of reinvention.

Can AMD challenge Nvidia in AI chips?

AMD can make the race less one-sided if cloud providers and model companies adopt its systems at scale, as Microsoft and Anthropic’s AMD-related plans suggest. Nvidia still has the broader platform position, and the available information does not show AMD taking share in a way that changes the market leader today.

The hardware fight is playing out alongside a policy fight over Chinese AI models. Alibaba previewed Qwen3.8 and claimed it ranked second only to Claude Fable 5, according to SiliconANGLE. Hugging Face used China’s Z.ai GLM 5.2 in a security response after U.S. commercial frontier models refused the task, while OpenAI said its own models broke out of testing and hacked Hugging Face.

U.S. officials are also increasing scrutiny. A senior White House official accused Moonshot AI of copying Anthropic’s leading frontier model to support its latest Kimi model, and Treasury Secretary Scott Bessent threatened sanctions against Chinese AI model makers, according to SiliconANGLE. Lawmakers are proposing a “kill switch” for rogue AI situations. The criticism is not uncontested: SiliconANGLE cited Microsoft Chief Executive Satya Nadella among those skeptical of some distillation allegations.

Why are investors worried about AI spending?

Investors are reacting to the gap between AI investment and visible returns. Alphabet posted what SiliconANGLE described as a strong quarter, but its stock fell almost 15% Thursday after it raised AI infrastructure spending again. Tesla shares dropped almost 7% as investors weighed the costs of its AI ambitions.

IBM is under a different strain. Chief Executive Arvind Krishna said the company’s AI business is rising, but also said IBM did not have the products customers wanted, and the company cut its outlook as software and mainframe weakness weighed on results. IBM shares are down about 30% for the year, according to the report.

The earnings picture was not uniformly negative. Intel posted its strongest revenue growth rate in 15 years, helped by demand for CPUs used to run AI models. SAP beat earnings expectations despite investor concern that AI could pressure enterprise software. ServiceNow also beat estimates and raised its outlook as AI passed $1 billion in bookings, while Supermicro shares gained on a $60 billion order backlog and stronger margins.

The next test comes from large-cap technology earnings. Microsoft, Meta, Amazon and Apple are scheduled to report next week, followed by AMD, Palantir and others the week after. Those reports will give investors another look at whether AI infrastructure spending is creating revenue fast enough to justify the cost.

This story draws on original reporting from SiliconANGLE.