guest
AI financing · Part 03 · settled Sep 8

The next $50B-$60B issuance will show whether new-issue concessions keep widening or start to normalize.

How it settles

The reading on October 2, 2026 (the last observation on or before that day) against the mean of the daily observations from August 1 to August 31, 2026: confirmed if it is 3 basis points or more below the August mean, refuted if it is 3 basis points or more above it. FRED, ICE BofA US Corporate Index option-adjusted spread (BAMLC0A0CM)

The exact question and the full rule

Did the investment-grade corporate bond spread (ICE BofA US Corporate Index option-adjusted spread, FRED series BAMLC0A0CM) read tighter on October 2, 2026, the end of the four weeks after Labor Day, than its August 2026 average?

The reading on October 2, 2026 (the last observation on or before that day) against the mean of the daily observations from August 1 to August 31, 2026: confirmed if it is 3 basis points or more below the August mean, refuted if it is 3 basis points or more above it. The split between AI-related issuers and the rest of the market, as the press reported it on September 22, 2026, is read into the Result as evidence and does not decide the test.

If the source is late or silent: A series that is late, revised after the reading or silent for the window is inconclusive, reviewed by the owner, and the Result says what was read, what was missing and the next step. A difference inside the 3 basis point band is inconclusive and closes the test.

If no

1999 capex plus 2008 credit

read as the fuse

If yes

The borrowing is rational

read as the fuel

Nosettled Oct 10

Confirmed: the investment-grade spread ended the window five basis points wider than August's average, as the Call expected.

See what it movedThe Result
— You
— Call
— Machinesnot run
— Readersshows from 5

The trail · since Oct 9
the Call — machines 0 runs readers show from 5
The case · 5 people · 2 minutes
1999 capex plus 2008 credit2 people argue noThe borrowing is rational3 people argue yes
Stronger if no · 2 people

1999 capex plus 2008 credit

Michael BurryAug 12 · 2 claims

Nebius extended its server depreciation schedule from four years to five this year even though its own deal pricing implies roughly 50 percent annual decay; customers that would not quit depreciating are what extended the bust twenty-five years ago.

Stronger if yes · 3 people

The borrowing is rational

Market SentimentAug 8 · 2 claims

Capacity is sold before it is built: AWS backlog reached 496 billion dollars, Microsoft closed June 2026 with 678 billion of commercial RPO up 84 percent year over year, Google Cloud's backlog hit 514 billion; CoreWeave's 2020-era A100s are fully…

Strongest against the CallOguz Erkan · Aug 6

Rental prices for six-year-old A100 GPUs are up more than 20 percent year-to-date, a direct read on AI model providers' willingness to pay and, by extension, end-customer demand.

What would change the Call
  • Nebius’s short-term premium collapses toward long-term rates.
  • H100 re-lets stay near original prices and the queued $50–60 billion of issuance clears without wider concessions.
The Call · Part 03 · 1999 capex plus 2008 credit

The machines will earn. The credit stack is where this buildout breaks first.

Open the dossierThe thesis
The dossier · 5 minutes
The machines' reasons

The machines have not run this one yet; they run when the Call is recorded.

The Part

The Machines Will Work. The Debt Might Not.

AI demand may be real. The more dangerous question is who is financing the machines, and how quickly lenders can change their minds.

Chart: The cloud revenue hurdle · Data Desk

Read the PartAll 5 questions