Michael Burry Points to Etched, an AI Chip Startup Backed by Jane Street, as Potential Competition for NVIDIA

A high-profile bet on the artificial intelligence chip market has resurfaced with investor Michael Burry highlighting a young startup he says could pose meaningful competition to NVIDIA, the dominant supplier of accelerated computing chips used for AI training and inference.

The company at the center of the renewed attention is Etched, an AI chip startup founded in 2022. According to the report circulating via Yahoo Finance, Etched has attracted at least $700 million in funding from Jane Street and other investors, and it has been valued at roughly $21 billion in the process.

Burry’s characterization of the startup as “serious competition” underscores how intensely the AI chip landscape is being watched. NVIDIA’s business has been closely tied to demand for data-center compute, and any credible alternative platform, even one still early in scale, can shift expectations among suppliers, customers, and capital markets.

Funding and valuation alone do not determine competitive impact, particularly in semiconductors where development cycles, software ecosystems, and manufacturing execution can take years to mature. The report does not outline what specific chips Etched is building, what performance metrics it targets, or whether customers have deployed the technology at meaningful volume.

In its own right, Etched’s backing by Jane Street indicates that sophisticated trading and market-structure investors are taking interest in specialized compute infrastructure, a theme that has shown up across the AI hardware supply chain as algorithmic firms seek faster, cheaper ways to run models.

NVIDIA, meanwhile, has built an extensive platform around its GPUs, including software tooling and system integration designed to make large-scale AI workloads practical. That broader ecosystem is often as important as raw chip specs, because developers generally want predictable performance, support, and portability across environments.

What remains unclear from the reported claims is whether Etched has reached production in a way that meaningfully challenges NVIDIA in current deployments. The circulating post does not provide details on shipping timelines, partnerships with cloud providers or OEMs, or the maturity of compilers and runtime software needed to translate hardware capabilities into day-to-day AI use.

For investors and technology buyers, the near-term question is not whether Etched exists, but whether it can demonstrate repeatable throughput at scale, earn software adoption, and secure supply. The market will likely focus on proof points such as customer deployments, performance benchmarks under real workloads, and any public evidence of volume manufacturing readiness.