The numbers are stated judgments. The record began on 12 October 2026 with starting confidences that a model set in a single reading, from its general knowledge and up to ten developments for each assessment retrieved in the three weeks before (some first reported earlier), and a published rule has made every change since. They are not measured probabilities. The question and the rule are fixed before the record begins, and each change shows its evidence and its reasons. The rule and what is not claimed are on the method page.
Capital spending on AI infrastructure by the largest US cloud companies keeps rising through 2027, so demand keeps pressing on the power constraint.
67.5% stated confidence, Likely (55–80%). Starting point on the record’s first day: Likely (55–80%), 67.5 per cent. Attention level: For information. Trend: Steady.
The judgment as stated before the record began
The main assessment holds only while demand keeps pressing on the grid. The clearest public measure of that pressure is what the largest cloud companies spend on property and equipment. Those figures are filed quarterly with the SEC and cannot be revised without a public filing. The assessment would weaken if spending stopped growing year over year, if several of the largest spenders cut their outlook in the same season, or if leases they have signed but not yet started began to shrink.
How the record started this assessment
On 12 October 2026, before the record began, a model read the 8 developments below, picked from the reports retrieved in the three weeks before (some first reported earlier), and with its general knowledge chose Likely (55–80%) in a single reading. The record started the assessment at the middle of that range, 67.5 per cent. The reading is kept in the first day's record.
Its reason, machine-written by anthropic/claude-sonnet-4.6: Developments 1 and 2 show Google Cloud revenue surging and AWS leadership defending continued heavy AI spending, consistent with factor R1 and R2 pointing to sustained demand pressure. Development 7 shows a large pre-IPO financing round closing, supporting factor R3. Development 3 notes some distress in data-centre debt markets, a mild signal against the assessment, but it does not indicate a broad spending pullback, so the balance of evidence supports likely.
- 1. First reported 11 October 2026, 1 outlet. Cited by the reading.“Alphabet's Google Cloud Is Surging at 82%, and the Heavy Spending Is Starting to Make Sense” · theglobeandmail.com · 11 October 2026 · Open the report
- 2. First reported 8 October 2026, 1 outlet. Cited by the reading.“AWS CEO Matt Garman defends AI spending with a spread-out customer base” · cryptobriefing.com · 8 October 2026 · Open the report
- 3. First reported 6 October 2026, 1 outlet. Cited by the reading.“Distress Watch: Data Center Debt Selloff Snags AI Buildouts” · news.bloomberglaw.com · 6 October 2026 · Open the report
- 4. First reported 6 October 2026, 1 outlet.“The AI capex ledger: hyperscaler spending versus cloud revenue, 2024 to 2026 | BlockWest Research” · blockwest.co · 6 October 2026 · Open the report
- 5. First reported 3 October 2026, 1 outlet.“Who is paying for the AI data center boom?” · thesuperintelligencenews.com · 3 October 2026 · Open the report
- 6. First reported 30 September 2026, 1 outlet.“Data center securities flash warning signs over AI debt supply” · therealdeal.com · 30 September 2026 · Open the report
- 7. First reported 26 September 2026, 1 outlet. Cited by the reading.“Nscale Raises $3.36 Billion In Pre-IPO Convertible Financing” · pulse2.com · 26 September 2026 · Open the report
- 8. First reported 25 September 2026, 1 outlet.Headline not shown · 247wallst.com · 25 September 2026 · Open the report
Causes of uncertainty
- Quarterly spending is uneven and includes land, buildings and equipment that are not all for AI.
- A growing share of AI capacity is financed off the largest companies' balance sheets, through leases and specialized providers, so their own spending can understate demand.
- Outlooks are stated as ranges and revised each quarter, and a cut can reflect timing rather than intent.
What would change the judgment
- Combined quarterly spending by the four largest spenders falling below the same quarter a year earlier.
- Two or more of them lowering their capital spending outlook in the same reporting season.
- Leases signed but not yet started shrinking for two quarters running.
The causal factors, and which way each one cuts
| Factor | Which way it cuts | What evidence looks like | Last moved |
|---|---|---|---|
| R1. Quarterly capital spending by the largest US cloud companies Payments for property and equipment by Microsoft, Alphabet, Amazon, Meta and Oracle, as filed with the SEC each quarter. Growth means demand keeps pressing on power. Public data named for this factor, not a test until admitted: sec.xbrl.microsoft.ppe_acquisitions_usd, sec.xbrl.alphabet.ppe_acquisitions_usd, sec.xbrl.amazon.ppe_acquisitions_usd, sec.xbrl.meta.ppe_acquisitions_usd, sec.xbrl.oracle.ppe_acquisitions_usd. | Supports. A stronger factor supports the assessment. |
| Not yet |
| R2. Capital spending outlook What those companies say they will spend in the current and next fiscal year, as stated in earnings releases and on earnings calls. A raised outlook means more pressure on power, and a cut means less. | Supports. A stronger factor supports the assessment. |
| Not yet |
| R3. Financing for data centre campuses Debt, lease and joint-venture financing for data centre campuses, including financing arranged outside the spenders' own balance sheets. Financing that closes on schedule keeps construction going. | Supports. A stronger factor supports the assessment. |
| Not yet |
| R4. Leases signed but not yet started Data centre leases that the largest spenders have signed but that have not yet commenced, as disclosed in the notes to their quarterly filings. These are contracted buildout, and growth means demand is committed ahead of supply. | Supports. A stronger factor supports the assessment. |
| Not yet |
| R5. Pauses, cancellations and lease exits Announced pauses, cancellations, deferrals or exits from leases by the largest spenders. More of these means demand presses less on power. | Weakens. A stronger factor weakens the assessment. |
| Not yet |
What the assessment assumes, and what would break it
- RA1. The largest US cloud companies keep raising capital spending year over year through 2027.Linchpin.
Invalidated if the combined quarterly payments for property and equipment of Microsoft, Alphabet, Amazon and Meta, as filed with the SEC (XBRL concepts PaymentsToAcquirePropertyPlantAndEquipment for Microsoft, Alphabet and Meta, and PaymentsToAcquireProductiveAssets for Amazon, with year-to-date figures differenced into quarters and the fourth quarter taken as the annual figure less nine months), are lower than in the same calendar quarter a year earlier for two consecutive quarters.
Source: SEC EDGAR XBRL company facts. The criterion’s fingerprint was fixed on 9 October 2026 and is part of the record’s first day. - RA2. The largest spenders do not cut their spending outlooks together.Bridge.
Invalidated if two or more of Microsoft, Alphabet, Amazon and Meta lower their stated capital expenditure outlook for the current or next fiscal year within the same quarterly reporting season, as stated in the earnings release furnished on Form 8-K or on the company's earnings call, excluding changes the company attributes to accounting or to how capital expenditure is defined. A lowered midpoint counts. A narrowed range does not.
Source: SEC EDGAR, Form 8-K earnings releases. The criterion’s fingerprint was fixed on 9 October 2026 and is part of the record’s first day. - RA3. Contracted data centre buildout keeps growing.Bridge.
Invalidated if the amounts each company discloses in its 10-Q or 10-K as leases not yet commenced at the balance-sheet date (for Amazon, the “Leases not yet commenced” row of its commitments table) fall at two or more of Microsoft, Alphabet, Amazon and Meta in each of two consecutive quarters.
Source: SEC EDGAR, quarterly and annual filings. The criterion’s fingerprint was fixed on 9 October 2026 and is part of the record’s first day.
The rival explanations
Evidence moves the assessment only when it bears differently on the assessment than on at least one of these. Evidence that fits them equally carries no weight.
- RX1. Capital spending levels off in 2026 as the largest spenders absorb capacity already built.No development was recorded on this assessment in the last 7 days.Favoured by:
- Year-over-year growth in quarterly spending falls toward zero.
- Outlooks hold spending flat.
- Companies describe absorbing or digesting capacity already built.
Ruled out by:- Quarterly spending is still growing year over year through 2027.
- RX2. Spending moves off the largest companies' balance sheets to AI labs and specialized providers, so their own capital spending understates the demand on power.No development was recorded on this assessment in the last 7 days.Favoured by:
- AI labs sign large long-term capacity contracts with specialized providers.
- Campuses are financed through special-purpose vehicles.
- Spending is flat at the largest companies while contracted capacity elsewhere rises.
Ruled out by:- Contracted capacity at specialized providers falls along with the largest companies' spending.
A development is recorded when the rule applies it to the assessment or sets it aside. Each count comes from the readings sealed with the changes, and a development set aside as fitting every alternative equally tells no alternative apart.
Evidence behind changes in the last 30 days
Each item is a report or data series cited by a sealed change on this assessment. The update loop reads headlines and summaries, and no summary is shown as the evidence.
No evidence has been cited in the last 30 days.
Evidence considered and set aside
Developments the update loop read in the last 30 days and set aside because they fitted this assessment and each of its rival explanations equally. Under the rule such evidence carries no weight, so none of it moved anything.
Nothing on this assessment was set aside in the last 30 days.
Changes to this assessment
No change to this assessment yet.
The rest of the judgment
The question’s text on this page is as sealed on the record’s first day. Fingerprint . In the fingerprint of 12 October 2026, published on GitHub. Check it in your browser.