Oct 23, 2026 · 16 min
Listen from 9:24
Listen at 9:24
Amazon’s idle GPUs expose cracks in the AI infrastructure boom
Monologue: Amazon Is Warehousing Billions In GPUs
Warehoused NVIDIA systems suggest the AI buildout is colliding with basic limits in power, data-center capacity, and planning.
- 1Amazon reportedly received NVIDIA systems before data centers had enough power and infrastructure to activate them.
- 2Ed Zitron argues NVIDIA’s early-chip push encouraged hyperscalers to accept hardware before deployment plans were ready.
- 3The episode frames idle GPUs as evidence of excess, poor coordination, and fragility across the AI infrastructure boom.
Don't miss
Zitron connects Amazon’s warehouses of inactive NVIDIA systems to a broader failure of planning in the AI infrastructure boom.
The brief
Ed Zitron examines reporting that Amazon is warehousing billions of dollars’ worth of NVIDIA AI systems because data centers lack the power and infrastructure to activate them.
The episode centers on Eugene Kim’s reporting about Amazon’s early rack receipt program, where hardware arrived before the facilities needed to run it were ready.
Zitron argues that NVIDIA’s push for hyperscalers to accept chips early helped turn a deployment problem into an artificial supply-chain crisis.
The result is a sharp critique of the AI infrastructure boom: excess hardware, weak planning, and physical constraints undermining the industry’s expansion story.
What was said on this episode
21 statements · 2 positive · 18 negative · 1 mixed
Billions of dollars of NVIDIA GPUs were being warehoused.
“confirming my estimate that billions of dollars of NVIDIA GPUs were being warehoused”
Listen at 2:56
Billions of NVIDIA GPUs were being warehoused before installation.
“billions of dollars of NVIDIA GPUs were being warehoused”
Listen at 2:57
NVIDIA is forcing customers to accept and pay for GPUs before they can use them.
“NVIDIA, who previously did not charge customers until they actually needed its GPUs, is now forcing them to take early receipt and pay them”
Listen at 3:52
NVIDIA appears to be channel stuffing GPU inventory.
“it appears that Nvidia is channel stuffing”
Listen at 4:20
Nvidia is prematurely pushing inventory onto customers to boost revenue.
“Nvidia is channel stuffing”
Listen at 4:21
Customers who anger Jensen Huang may lose access to NVIDIA GPUs.
“if you piss off Jensen Huang, you don't get your GPUs”
Listen at 5:08
Jensen Huang’s GPU shipment practices caused higher RAM and consumer-electronics prices.
“the increased price of RAM and consumer electronics was a result of an artificial supply chain crisis created by Jensen Huang”
Listen at 6:23
Early GPU production consumed capacity from Micron, SK Hynix, and Samsung.
“which in turn meant buying out capacity from Micron, SK Hynix and Samsung”
Listen at 6:41
NVIDIA’s advance GPU production bought capacity from Micron, SK Hynix, and Samsung.
“buying out capacity from Micron, SK Hynix and Samsung”
Listen at 6:42
NVIDIA’s advance purchasing was intended to increase its revenue growth.
“all to increase revenue growth for NVIDIA”
Listen at 6:52
Jensen Huang is responsible for worsening people’s lives.
“Jensen Huang is responsible for making life worse for everyone”
Listen at 6:56
Hyperscaler executives and related participants drove up prices broadly.
“These scumbags drove up the price of everything”
Listen at 7:17
The market did not need Jensen Huang to build all these GPUs.
“We did not need him to build all these GPUs”
Listen at 7:33
NVIDIA-related industry actions made goods more expensive and conditions worse.
“Everything is more expensive now. Everything is worse because of what these people did.”
Listen at 8:30
The generative-AI industry has worsened everything it affects.
“This destructive, pointless void of an industry has made everything it touches worse”
Listen at 9:09
Generative AI has made life more expensive and worse in various ways.
“Everything is more annoying, more expensive, and worse on some level because of generative AI”
Listen at 9:24
GPU presales raised consumer-electronics prices to support stock-market and quarterly revenue growth.
“the pre-sales of GPUs, which ran up the price of basically every consumer electronic, is a direct result of only stock market growth, and of only quarterly revenue growth”
Listen at 10:48
The AI bubble will eventually explode.
“I know it will now”
Listen at 11:18
The current AI industry is sustained by cultism and private-equity excess and will collapse.
“the AI industry as it exists today is being held together through cultism and private equity excess and is rigged to blow and is going to explode”
Listen at 12:45
The world will become difficult but better after the AI industry collapses.
“the world will be different afterwards and it's going to be difficult. But I think it will be better.”
Listen at 12:54
The world will improve after the AI industry explodes.
“But I think it will be better”
Listen at 12:58
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.