· via Hacker News – Front Page (native)
Nvidia enters post-Jensen Huang era facing AMD, custom silicon and an eroding CUDA moat
Nvidia has ousted CEO Jensen Huang and installed CFO Colette Kress just as AMD, custom cloud silicon, AI code-porting and Chinese rivals chip away at its dominance, WIRED reports.

Nvidia has ousted co-founder and CEO Jensen Huang, and CFO Colette Kress has taken the reins while it remains unclear who will hold the job permanently, WIRED reports. The leadership change lands at an awkward moment: the company that supplies most of the hardware behind the AI boom is facing competitive pressure on several fronts at once.
Huang co-founded Nvidia in 1993 and, as WIRED recounts, saw AI's promise in the mid-2010s — funding academic research, building supporting software and giving hardware to labs and startups. That bet paid off when ChatGPT sent demand soaring: revenue climbed from $26.91 billion in January 2022 to $215.94 billion in January 2026, according to figures cited in the report.
Rivals are gaining ground
Nvidia still captures about 80 percent of AI accelerator sales, according to Silicon Analysis data cited by WIRED. But AMD's share has climbed from 2 percent in 2022 to between 7 and 10 percent in 2026, and Google, Amazon and Microsoft are all designing more of their own AI chips to cut costs and loosen their dependence on Nvidia. Specialized startups add further pressure: Cerebras sells oversized chips aimed at very large models, and AMD recently bought Taalas, a company that hard-bakes model weights into silicon for efficiency. Nvidia made its own countermove, acquiring the specialized chip startup Groq in December 2025, according to the report.
Patrick Moorhead, chief analyst at chip-industry tracker MoorInStrat, told WIRED that the rise of AI agents makes inference — running models that have already been trained — increasingly important relative to training, an area where hyperscalers and startups compete with Nvidia more effectively. He also pointed to growing demand for general-purpose CPUs as assistants and agents spread, a market where Nvidia barely competes.
China cost Huang his job
According to WIRED, Huang's exit was driven by his China dealings. He had a romantic relationship with Huang Le, also known as Alice Huang, an executive at the Singaporean company 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 billion and $7.5 billion worth of Nvidia chips into southeast Asia, and reports suggest the company may have given Chinese AI firms access to them.
China has also shrunk as a market. US restrictions bar Nvidia's most powerful chips there, and China now accounts for about 8 percent of company revenue, down from 20 percent in 2024, according to filings cited by WIRED. Meanwhile, Huawei has developed new supercomputers that harness many more chips than usual, and Chinese AI firm Z.ai — which WIRED describes as making one of the world's best open-source models — says it trained its models entirely on Huawei silicon. Babak Minovi, a senior analyst at the Eurasia Group, told WIRED that Huang was right that cutting off China would empower domestic rivals: "Huang was taking a pragmatic line, and he knows competition."
The CUDA moat is thinning
Nvidia's strongest lock-in has been CUDA, the software layer that made it easy to build for its chips. But WIRED notes that AI models are now adept enough at reading and writing code that they can port models onto competitors' silicon quickly. Startup Wafer, which ports models to AMD hardware and then serves them more cheaply, can have a major new open-weight model running on AMD within hours at 80 percent of the efficiency, according to CEO Emilio Andere, who says his engineers barely need to review the output. The report also flags AI's growing role in chip design itself — OpenAI and Anthropic have both touted models that can design electronics — though Moorhead argues Nvidia is riding the same wave.
An AI-company CEO who spoke to WIRED anonymously put the strategic problem bluntly: "Everything sold today was designed two-to-three years ago." Kress, or whoever replaces her permanently, must chart the course the chip industry takes two years from now.
Why it matters
Nvidia is the infrastructure backbone of the AI industry, so how this transition plays out will shape computing costs and hardware choices across the sector. Veteran chip analyst Dan Hutchinson told WIRED the situation echoes Intel's 2019 replacement of engineering-focused CEO Brian Krzanich — who, like Huang, resigned after a relationship, and WIRED notes Huang is alleged to have had multiple — with CFO Bob Swan, after which the company drifted toward sales over technical progress and revenue sank. Hutchinson's verdict: Nvidia may be entering an "Intel swan moment," and "Nvidia is now an Intel waiting to happen." None of the individual challenges may be existential on their own, but together they arrive just as the company loses the leader whose foresight built its lead.
- #nvidia
- #jensen-huang
- #ai-chips
- #semiconductors
- #amd