Test report DSG-6676 · Rev A · tested November 16, 2026

AI Datacenter InfrastructureDevice under test

GPU Rental Data Undercuts Michael Burry's Depreciation Warning

Older NVIDIA GPUs are fetching rising rental prices, data from Silicon Data show, challenging Michael Burry's claim of $176 billion in understated depreciation through 2028.

Read
3 min
Words
694
Node
3nm
Operator
Amara Osei

Spec summary

  1. Burry estimated the industry could understate depreciation by $176 billion between 2026 and 2028.
  2. NVIDIA shares closed at $162.85 on November 13, down 26.41% over the past month.
  3. NVIDIA posted Q2 revenue of $96.221 billion, with Data Center revenue of $89.023 billion, up 117% year over year.
  4. NVIDIA supply commitments reached $279 billion, with AI cloud guarantees capped at $108.5 billion.
  5. Next-quarter guidance stands at $108.0 billion revenue, plus or minus 2%, with 74.0% gross margins.
Michael Burry Warned AI Chips Would Lose Value Fast. New Data Suggests He Was Wrong
Fig. AMichael Burry Warned AI Chips Would Lose Value Fast. New Data Suggests He Was Wrong — heladodementa / pixabay

By Amara Osei

Rental prices for certain older NVIDIA GPUs are rising rather than falling, according to Silicon Data, undercutting a central assumption in Michael Burry's $176 billion depreciation-fraud thesis. The investor has spent the past year arguing that the AI boom rests on stretched server accounting. New market data suggest the hardware holds commercial value longer than he assumed.

The stakes are immediate for NVIDIA (NASDAQ:NVDA). Shares closed at $162.85 on November 13, down 26.41% over the past month. Jensen Huang resigned as CEO amid disclosures about personal relationships, and the board appointed finance chief Kress interim CEO this week. Investors now question whether AI capital spending can hold up without him.

What is Burry's case?

In his Substack, Cassandra Unchained, Burry called stretching the assumed lives of servers and GPUs "one of the more common frauds of the modern era." He estimated the industry could understate depreciation by $176 billion between 2026 and 2028.

The mechanics are simple. Buyers of NVIDIA processors spread each chip's cost over its estimated useful life. A longer schedule means smaller yearly expense and higher reported profit. Microsoft (NASDAQ:MSFT) and Alphabet (NASDAQ:GOOGL) have moved certain server classes to four to six years.

Burry counters that NVIDIA's 12- to 18-month product cycle shortens the real economic life of each generation, so today's earnings borrow from tomorrow's losses. His argument concerns accounting lives and reported earnings — a thesis that still leaves room for older GPUs to keep commercial value.

What does the rental data show?

In a Yahoo Finance interview with Deirdre Bosa, Silicon Data founder and CEO Carmen Li said older GPUs are holding value better than expected, with rental prices for certain older chips rising rather than collapsing as newer generations arrive. Silicon Data publishes daily GPU benchmarks for on-demand rental rates, giving it a direct view of compute costs.

NVIDIA's own figures point the same way:

  • Revenue reached $96.221 billion in the fiscal second quarter, with Data Center revenue of $89.023 billion, up 117% year over year.
  • Kress said on the August 26 call that NVIDIA compute is "fully utilized across every cloud we serve" and that supply should stay a constraint "at least through the end of fiscal year '28."
  • The company still ships Hopper parts, now two generations back.

Why do older chips keep earning?

Frontier model training makes headlines, but inference, fine-tuning, research, and smaller workloads can run economically on older hardware. Huang said on the call that an AI agent needs 15 to 100 times the compute of a human user, and that a deployed system "will be preserved and useful and productive for a lot longer." When demand runs ahead of supply, last-generation chips stay in service.

Where does Burry still have a point?

Rents measure current earning power. Resale value, book value, and lifetime profitability are separate questions. Today's pricing also reflects a shortage that could ease.

Kress said Vera Rubin delivers 35 times lower token costs relative to Grace Blackwell Ultra. Once that supply arrives in volume, demand for older chips could weaken. NVIDIA has also absorbed more risk to keep the expansion moving:

  • Supply commitments reached $279 billion.
  • Guarantee obligations for AI cloud partners are capped at $108.5 billion.
  • Days sales outstanding stretched to 60 days from 45 days.

Which numbers settle the debate?

The accounting question will be decided by data that has not yet arrived. Three indicators matter:

  • Rental indexes: whether older-chip rents hold once Vera Rubin supply catches up.
  • Hyperscaler disclosures: any change to server useful-life estimates in quarterly filings.
  • NVIDIA's next report: results against guidance of $108.0 billion in revenue, plus or minus 2%, gross margins of 74.0%, and any comment on utilization under interim leadership.

The stock trades at a P/E of 33, so the market already prices in durable demand. Sharp rent declines after the next product ramp would give Burry's warning weight; rents that hold would strengthen the case for longer depreciation schedules.

The data so far lean against the short-life thesis.

via 247wallst.com (Source)

Filed under

  • nvidia
  • gpu-rentals
  • server-depreciation
  • ai-infrastructure
  • michael-burry
Share this article:

More from Amara Osei

Amara Osei

Show full bio

Staff writer covering money, funding and market shifts around hardware layer of artificial intelligence at Die Signal.

698 articles

Same lot · LOT-362C

« Previous articleNext article »