Amazon can deliver almost anything in a day. Its new AI racks are waiting on electricity
SummaryBusiness Insider reports that AWS is taking delivery of billions of dollars of Nvidia GB300 server racks earlier than before, and that some sit in warehouses because the data centers don't have enough power to turn them on yet.
· MODEL Editors · 2 min read
Some links may be affiliate links. Disclosure

Amazon has built the world's most famous delivery machine. Its newest AI hardware is arriving fine. Plugging it in is the hard part.
What Business Insider found. In a report published Tuesday, Business Insider's Eugene Kim says AWS set up a new "Early Rack Receipt" process for Nvidia's GB300 server racks. Under it, AWS expected to receive roughly $5 billion of GB300 racks at US facilities before final deployment, according to August internal emails the outlet reviewed, and projected roughly $3 billion more each month "for the foreseeable future."
The catch is power. Some AWS sites had room for more racks but not enough available power to turn them on, people familiar with the effort told Business Insider, so finished equipment was staged at intermediary warehouses until the facilities were ready. At least one AWS team did not have "a clear forecasted path" for getting the incoming racks into data centers at the time, one email said.
Amazon's side. An Amazon spokesperson told Business Insider the company routinely stages finished hardware near deployment sites so it can install it once capacity is available, calling it "best business practice," and disputed that the broader company lacked a path for getting racks into data centers.
Why it matters. Previously AWS took ownership of a finished rack only when it reached a data center; under the new approach it can pay for and record some racks earlier in the supply chain. That secures scarce hardware before rivals do, but ties up cash before the gear earns cloud revenue. Business Insider says Amazon expects roughly $220 billion of capital spending this year and that its free cash flow turned negative last quarter for the first time since 2023. Separately, the Financial Times reported this month that Amazon is exploring moving about $8 billion of Grace Blackwell chips into an investor-funded vehicle that would lease them back.
Our take
The AI buildout's bottleneck keeps moving: first it was chips, now it's the wall socket. If you build on cloud GPUs, this is a reminder that "capacity coming soon" depends on power grids and substations as much as on Nvidia's shipping schedule, so plan for regional shortages and keep a second provider in your back pocket.
What we did and didn't do: this story is based on Business Insider's report linked below, which rests on internal emails and unnamed sources we could not see ourselves; the dollar figures are Amazon's projections from August, not a count of what is in warehouses today. The FT leaseback detail comes from Bloomberg Law's summary of the FT report. This is not investment advice.
MODEL Editors
Sources
Not investment advice.
FREE · EVERY WEEK
The top AI tools this week, tested.
Every week MODEL's editors pick the AI tools worth your time, try them hands-on, and tell you what's good, what's not, and who it's for. Free, in your inbox.
Free. Unsubscribe anytime. Our picks are editorial and can't be bought.
The weekly three is editorial. A subscription does not buy a place in it.
Continue

Markets
Microsoft walks away from Britain’s biggest planned AI data centre
Nscale’s Loughton site in Essex was meant to hold 23,040 Nvidia GPUs for Azure. UK reports say its grid connection may not arrive until the 2030s. Microsoft says the decision covers a single site.
· MODEL Editors

Markets
Anthropic wants $60 billion for chips. The lender wants 10%.
The Financial Times reports Anthropic is lining up the largest chip financing yet, to pay for about 1.5 gigawatts of Google TPUs, and offering Blackstone a double-digit return to anchor it.
· MODEL Editors

Markets
Chip stocks crashed on a Saturday. The stocks weren't trading.
With Wall Street closed, about $51.6 million in leveraged crypto bets on chip stocks were wiped out in ten minutes. The shares themselves never moved.
· MODEL Editors