Michael Burry has spent much of the AI boom looking for cracks in Nvidia Corporation (NASDAQ:NVDA). Now, the “Big Short” investor thinks he may have found one.
Burry recently called Etched, a four-year-old AI chip startup, “serious competition” for Nvidia. The timing makes the warning harder to dismiss. Etched just raised $700 million in a Jane Street-led round that valued it at $21 billion, more than double its valuation less than a month earlier. Reuters says the company has already secured more than $1 billion in customer contracts.
Why Burry May Be Right
Etched is not really trying to beat Nvidia at everything. It is attacking AI inference, where cost per token, latency and power efficiency increasingly matter as AI companies move from training enormous models to serving them at scale.
Michael Burry of Scion Asset Management
Burry cited an industry source who believes Etched could deliver roughly 10 times the performance at lower cost. There is also an uncomfortable talent angle for Nvidia: about 15% of Etched’s 400-plus employees previously worked there. Etched’s leadership includes Brian Loiler, a 22-year Nvidia veteran who helped build the HGX and DGX systems.
If those economics hold up as deployments scale, Etched does not need to dethrone Nvidia. Taking meaningful share in inference could be enough to pressure Nvidia’s extraordinary pricing power and margins.
Nvidia’s own actions suggest specialized inference is worth taking seriously. Earlier this year, the company entered a major non-exclusive inference technology licensing and talent agreement with Groq for about $17 billion. The deal strengthens Burry’s broader argument that specialized inference hardware matters, while also showing Nvidia is hardly standing still.
Why Nvidia Bulls Aren’t Panicking
Etched still has to prove it can manufacture and deploy its systems at Nvidia-like scale. Its performance claims have not yet been independently demonstrated across the enormous range of workloads Nvidia supports. Nvidia also sells far more than chips: CUDA, networking, rack-scale systems and software create switching costs that a faster specialized chip does not automatically erase.
And the financial machine is hardly sputtering. Nvidia’s latest quarter produced $81.6 billion in revenue, up 85% year over year, including $75.2 billion from data centers. BofA analyst Vivek Arya recently maintained a $350 target, implying more than 50% upside from recent levels.
Institutional positioning is also nowhere near panic mode, as shown in Insider Monkey’s hedge fund database. Fisher Asset Management held 90.9 million Nvidia shares in Q2, while AQR increased its position 18% and Arrowstreet added 10%. Nvidia short interest actually fell 9.7% by July 31 to about 292.7 million shares, roughly 1.2% of shares outstanding.
Etched therefore looks like a credible threat, not yet an Nvidia killer. The bigger danger is that specialized inference chips slowly chip away at Nvidia’s dominance and compress its margins. With Nvidia still growing at breakneck speed and trading around 20 times forward earnings, the stock looks increasingly attractive if meaningful weakness creates a better entry point. But Burry has identified a competitive threat Nvidia investors can no longer comfortably ignore.
While we acknowledge the risk and potential of NVDA as an investment, our conviction lies in the belief that some AI stocks hold greater promise for delivering higher returns and doing so within a shorter time frame. If you are looking for an AI stock that is more promising than NVDA and that has 10,000% upside potential, check out our report about this cheapest AI stock.
READ NEXT: NVIDIA (NVDA): What Foxconn and Super Micro Are Telling Us about the AI Boom and Pony AI Is Scaling Robotaxis Fast—Can the Stock Reach BofA’s $17 Target?
Disclosure: None. Follow Insider Monkey on Google News.
