Michael Burry Goes After Nvidia Again: Bullwhip, Bezzle, and the Cisco Moment

Burry's Verdict: 'Not Enron. Clearly Cisco.'

Michael Burry of Scion Asset Management published a new Substack post — 'The Heretic's Guide to AI's Stars Part III: Tracepalooza and the Bezzle' — that his most specific attack yet on the AI infrastructure trade. More than 200,000 subscribers read it, and Nvidia took the unusual step of sending a memo to sell-side analysts to push back on his arguments around stock-based compensation and depreciation. Burry responded directly: 'I stand by my analysis. I am not claiming Nvidia is Enron. It is clearly Cisco.'

Three Mechanisms: Buyer Concentration, Bullwhip, and the Bezzle

Burry's warning centers on three structural mechanisms. First, buyer concentration: hyperscalers — Microsoft, Google, Amazon, and Meta — account for approximately 50% of Nvidia's data center revenue. Burry calculated that if Microsoft alone cuts its Nvidia chip spending by 20%, Nvidia takes an approximately 4.2% revenue hit. A small change in one customer's capex plans produces outsized effects.

Second, the bullwhip effect: when supply chain buyers over-order due to fear of missing out, that distortion amplifies backward through the chain. Nvidia has locked in $119 billion in non-cancellable supply commitments with TSMC. Data center financing has expanded to accommodate the buildout. The entire chain is betting demand is permanent. Burry argues it is not.

Third, the bezzle — John Kenneth Galbraith's term for the gap between what people believe they own and what actually exists. Burry argues that current AI revenue flowing through Nvidia reflects a training and benchmarking phase that will eventually give way to an inference and deployment phase with a fundamentally different — and lower — demand profile. 'That distorted demand is working like a bullwhip into NVIDIA's own supply chain,' Burry wrote.

The Cisco Parallel — And Why It Matters Now

Cisco supplied the networking infrastructure for the dot-com internet buildout. Its revenue was real, its customers were real, and the technology was genuinely transformative. What wasn't real was the assumption that the buildout phase would continue indefinitely. When telecom companies stopped buying, Cisco fell more than 80% from its peak. Current top 10 AI stocks have surged 784% in 12 months — versus 622% at the peak of the pre-dot-com period. That 162 percentage-point gap forms the quantitative basis for Burry's comparison. ⚠️ Note: Burry has been early — and wrong — on Nvidia for three years. His framework may prove correct while his timing remains off by years.