The Calls

Every Call, with its number and its next test.

Each Part of The Buildout takes a side on one question, gives the yes answer a number before the test’s date, and keeps the score here.

The Docket Results Scorecard The questions

  1. Part 13Energy and infrastructure

    Who pays for the AI power bill?

    SplitThe ratepayer pays first.

    Data centers pay for their own wires. The higher capacity price their demand sets lands first on every PJM household's bill.

    Next testThe federal price data for August: whether the median twelve-month rise in household prices across PJM's nine jurisdictions stays at least four points above the median of the other 42, from 11.7% against 5.6% in July.Call · 82% yesOn the Docket

    3 tests on the Docket · none resolved yet

  2. Part 12Labs and models

    Does the frontier model still command a price?

    SplitThe price has left the model.

    A frontier model no longer commands a price per token for long. The lab still does, and only a prospectus says what that price is worth.

    Next testAlphabet, Microsoft and Meta report, and the Fed decides the same day: whether the TPU system sales inside Google Cloud's backlog get a size, and whether Cloud growth holds near the second quarter's 82% without them.Call · 20% yesOn the Docket

    5 tests on the Docket · none resolved yet

  3. Part 11Chips and suppliers

    Do the orders turn into profit?

    SplitRevenue yes, margin not yet.

    The orders will ship. The profit comes later, and a margin that lags the ramp will be reported as a demand miss.

    Next testMicrosoft reports: the share of backlog due within twelve months, 30% last quarter, and whether the near book's growth closes on the far book's 113%; at Alphabet, whether the TPU contracts inside cloud backlog get a size; at Meta, whether the $10.8 billion of restricted cash stays restricted.Call · 45% yesOn the Docket

    5 tests on the Docket · none resolved yet

  4. Part 10Markets and rates

    Does the hardware hold its value?

    SkepticalCheap is conditional.

    The market is not cheap. It is capitalizing AI earnings before anyone knows who owns them.

    Next testAnthropic filing or listing reveals audited revenue, losses and concentration, testing the $400B revenue milestone.On the Docket

    6 tests on the Docket · none resolved yet

  5. Part 09Distribution and applications

    How fast do habits break?

    SplitSell habit, buy the rails.

    Agents will drain weak consumer habits first. Back the rails (trust, records, distribution, compute), not the forgotten subscription.

    Next testMeta reports its September quarter; users, transaction volume and fees would show whether downloads are becoming a business.Call · 55% yesOn the Docket

    5 tests on the Docket · none resolved yet

  6. Part 08Capital and credit

    Who writes the checks for the buildout?

    SkepticalLenders are mispricing scarcity.

    Lenders are right to finance capacity and wrong to assume it stays scarce. Scarcity is a season; debt is a contract.

    Next testSeptember-quarter results test cash flow toward $2 trillion and whether maintenance costs enter accounting.Call · 80% yesOn the Docket

    5 tests on the Docket · none resolved yet

  7. Part 07Markets and rates

    Why did the long end climb after the hike: growth, debt, or a missing buyer?

    SkepticalSupply sets the long end.

    The bond market is pricing the AI bill, not a recession. The Fed cannot hike its way to a lower 10-year.

    Next testTreasury publishes August foreign holdings; renewed Japanese buying despite poor hedged returns would undercut the missing-buyer thesis behind the long-end selloff.Call · 35% yesOn the Docket

    4 tests on the Docket · none resolved yet

  8. Part 06Energy and infrastructure

    Is power scarcity a warning, or a moat?

    ConstructiveLong access, not hardware.

    This is a power-access trade before it is a chip trade. Own the permit and the interconnect, not the loudest fleet.

    Next testERCOT is due to verify its first large-load batch study; the cleared list shows whether Texas weeds out speculators or rations power.On the Docket

    5 tests on the Docket · none resolved yet

  9. Part 05Labs and models

    What does a frontier lab sell: a commodity, or a premium?

    SplitSplit the trade.

    Stop pricing OpenAI and Anthropic as one bet. One revenue line holds under pressure; the other has to be bought.

    Next testReturn on invested capital should be visible by Romero’s deadline; failure would turn today’s extraordinary economics into a first-year illusion.On the Docket

    5 tests on the Docket · none resolved yet

  10. Part 04Capital and credit

    Is the capex spending defense, or is the return not there?

    SkepticalProve the return.

    Capital intensity is not a moat. Every new AI dollar now has to out-earn its financing cost.

    Next testAlphabet discloses its higher 2027 budget; a rise in the stock revives the defense case, another fall validates the return test.Call · 45% yesOn the Docket

    4 tests on the Docket · none resolved yet

  11. Part 03Capital and credit

    The debt behind the machines: 1999 capex, 2008 credit, or neither?

    SkepticalBearish on the credit.

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

    Next testShort-term premiums collapsing toward long-term rates would weaken the backwardation warning; persistence would make it harder to dismiss.Call · 60% yesOn the Docket

    1 of 5 tests resolved · 1 held · 0 missed

  12. Part 02Capital and credit

    SpaceX after the IPO: exit liquidity, or the next ten-bagger?

    SkepticalAhead of the evidence.

    SpaceX’s AI build is credible. The stock has already paid for 10 gigawatts it has not built.

    Next testQuarterly deployed gigawatts will test the roughly 10 GW site plan against 0.4 GW operating now.On the Docket

    4 tests on the Docket · none resolved yet

  13. Part 01Compute and cloud

    Is the marginal buyer of AI compute real?

    SplitBull on demand. Bear on the financing.

    Nvidia’s orders are real. Price its revenue as financed, not recurring.

    Next testNo dated test yet. 3 open questions on the Docket.

    3 tests on the Docket · none resolved yet

The Call’s number is the chance, from 0 to 100, that the Call gives the yes answer to a test’s question, recorded before the test’s date; it is DeepStack’s own reading, not a market price and not advice. The readers’ mark on the same line is the share of yes among the anonymous clicks on the test, shown from 25 clicks. Analysis and opinion, not investment advice.