AI & Tech · National
Nvidia faced growing competition from AMD, cloud giants and startups even before Jensen Huang's exit, Wired reports
Wired reports that Nvidia, now led by CFO Colette Kress, was already contending with AMD, custom chips from Google, Amazon and Microsoft, shrinking China sales and AI tools that ease switching to rival hardware.
Listen to this brief · 5 min
Read aloud by an AI voice from the text below. Download MP3
What happened
Wired reported that Nvidia was facing a growing field of competitors before the controversy that led to Jensen Huang's departure as CEO. Chief financial officer Colette Kress has now taken the reins of the company, and Wired described the timing as difficult.
The CEO of a large AI company, who spoke to Wired anonymously, said "Everything sold today was designed two-to-three years ago." That person told Wired that Kress, or whoever permanently replaces her, will need to set the direction the chip industry takes two years from now. Our earlier coverage looked at how Huang's exit affects AI startup funding.
Where Nvidia stands today
Nvidia still holds 80 percent of sales in the market for AI accelerators, according to Silicon Analysis, as cited by Wired. The company's revenue rose from $26.91 billion in January 2022 to $215.94 billion in January 2026, after ChatGPT brought AI into the mainstream in 2022.
Wired credited much of that lead to Huang, who cofounded Nvidia in 1993 to make graphics chips for PCs. In the mid-2010s he funded academic AI research, built software for AI engineering and supplied hardware to labs and startups.
Who is gaining ground
AMD's rising share
AMD has made a concerted push into AI chips, Wired reported. Its share of the market grew from just two percent in 2022 to between 7 and 10 percent in 2026. AMD also recently bought Taalas, a company that builds model weights directly into silicon to make computing highly efficient.
Cloud companies and startups
Google, Amazon and Microsoft are each developing specialized chips for training and running AI models, aiming to cut costs and depend less on Nvidia. Startups such as Cerebras, which sells very large chips for large models, also offer alternatives. Wired noted that using many of these chips has required rewriting a lot of code.
Inference and general processors
Patrick Moorhead, chief analyst at MoorInStrat, told Wired that the rise of AI agents makes inference, or running already trained models, more important. Nvidia has less of an edge there, though it acquired a specialized chip startup called Grok in December of 2025. Moorhead also pointed to growing demand for general purpose CPUs, an area where Nvidia barely competes.
China and the Huang controversy
Nvidia cannot sell its most powerful chips in China but can sell older generations. Sales of those and non AI chips doubled over the past year, yet China now makes up only eight percent of company revenue, down from 20 percent in 2024, according to company filings cited by Wired. Meanwhile, Huawei has built supercomputers that link many more chips than usual, and some Chinese AI firms such as Z.ai say they trained models entirely on Huawei chips.
Wired reported that Huang's dealings with China played a major role in his downfall, including a romantic relationship with an executive at Singapore-based Megaspeed whom the US government suspects of being an asset of China's Ministry of State Security. Megaspeed and its subsidiaries have reportedly imported between $4.6 and $7.5 billion in Nvidia chips into southeast Asia. Lawmakers have since looked at tighter AI chip export rules.
AI tools weaken Nvidia's software advantage
For years, Nvidia's CUDA software has been seen as a moat keeping developers on its chips. Moorhead told Wired that AI models can now rewrite code quickly enough to move a model to a rival's hardware. Emilio Andere, CEO of the startup Wafer, said his company can run major new open weight models on AMD hardware at least 80 percent as efficiently within hours.
Veteran chip analyst Dan Hutchinson told Wired that Nvidia may be entering its "Intel swan moment," referring to Intel's 2019 replacement of its CEO with its CFO, after which Wired said the company's focus shifted toward sales.
What to watch next
- Who will permanently replace Colette Kress as Nvidia's leader, a question Wired described as still open.
- How many chips exported to other countries end up in China or are used remotely by Chinese firms, which Wired said remains unclear.
- Whether Chinese open-weight models gain traction abroad and create demand for Chinese chips outside China.

