Post

Ed Zitron
@edzitron.com

Free newsletter: Outside of major hyperscaler backstops, I estimate that Anthropic and OpenAI would get a speculative B or CCC credit rating, which would be insufficient for the $50bn-$100bn a year they need, in an era of historically high interest rates. www.wheresyoured.at/credit-crunch/

wheresyoured.athttps://www.wheresyoured.at/credit-crunch/If you liked this piece, you should subscribe to my premium newsletter, and you can subscribe on the following links: $70 a year, $18 a quarter, or $7 a month. In return you get a weekly premium…

101 reposts565 likes

The Infinitely Prolonged
@wowbaggert.bsky.social

"all that data center compute will be used by someone else one day!" Another way that the dot com fiber overbuild analogy is wrong: the usable lifetime of fiber infra is 20-40 years. The usable lifetime of a gpu is ... 2-3 years.

Ed Zitron
@edzitron.com

To be clear, this isn't something affecting just those exposed to the AI bubble. All corporate debt is influenced by the cost of ten-year treasuries, the cost of which has soared recently, making junk-heavy data center debt much more expensive to raise. www.wheresyoured.at/credit-crunch/

Ed Zitron
@edzitron.com

For a company like Oracle, that must continuously raise debt, this is disastrous. Not only is the baseline cost of borrowing higher, but also, its previous debt looks shaky, meaning that investors will demand more yield to cover their risk. www.wheresyoured.at/credit-crunch/

Howl The Wizard
@howl--the--wizard.bsky.social

@edzitron.com, where do you reckon that 22 billion of non-frontier compute spend is going? I was thinking at first some smaller fish trying to stand-up/expand inference providers or regular old data center spend but 22 billy seems awfully small for that purpose.

Howl The Wizard
@howl--the--wizard.bsky.social

Hope your take on debt markets being a little more practical with numbers rings true. It seems like every level of the business world just sees the potential $$$ signs and ignores due diligence. My lack of familiarity means I don't have a very strong reason to assume they wouldn't do the same.

Alot
@alot.bsky.social

As a subscriber, I find it utterly reprehensible that you served up something as powerful as "Goatse-level spreads" in a free article. Proper belly laugh.

Jack Kliever
@silverhorseman.bsky.social

What people believe is mostly based on what they wish was true and what their peers believe. Facts and reason are secondary influences at best. Altman, Musk and their ilk know this instinctively and are seeing how far they can take it. Farther than I would have guessed.