Can the last week of October confirm the frame, or only break it?
The last week of October can break the frame but cannot confirm it: only the Fed statement and a labor-cost index are sure to arrive, and the largest artificial-intelligence spenders have not dated their results.
Drafted with Claude Opus 5.5; counter-case by GPT-6 Astra · read and committed by the publisher How we work →
Bottom line
- The question is what the Fed, the largest artificial-intelligence spenders and the inflation data must show by the end of October for the frame to hold.
- Our answer: the Fed must not walk back its September position, the one labor-cost release must not contradict the inflation judgment in the Fed's minutes, and no large spender may cut its range for the year.
- Grant Bishop's view: The week can break the frame more easily than confirm it, because the October consumer-price figures fall after it and the spenders' results dates are unconfirmed, and a Fed statement that softens the September stance, or a spending range cut at third-quarter results, would change our answer.
What would change our mind
A statement at the close of the 27–28 October Fed meeting that walks back the September position, an Employment Cost Index showing labour costs easing sharply enough to undercut the September minutes' view of inflation risks, or a third-quarter results release from Alphabet, Amazon or Meta that cuts or fails to restate its spending range for the year, would change this conclusion.
The question
In the last week of October three kinds of evidence are due to meet: a Federal Reserve statement, results from the largest artificial-intelligence (AI) spenders, and inflation data. The piece bears on Spare computing capacity does not stop new power construction, The Fed under Warsh keeps raising rates into this inflation and Spending on the AI build is running ahead of its revenue. The question is what each must show for the frameThe frame: what we think is going on in the world, written down, dated, and revised when the evidence asks. The House View is the frame in full; this picture is the frame at a glance. — our working reading of the economy and AI spending, as set out in those watchpoints — to hold, and whether the week will actually produce that evidence. Our answer is that the frame survives the week if the Fed statement does not walk back its September position, the one labor-cost release does not undercut the inflation judgment in the minutes, and no large spender cuts its spending range for the year. But the week as scheduled can break the frame; it cannot confirm it.
What is known
The Fed. At its September meeting the Federal Open Market Committee (FOMCFOMC: the Federal Open Market Committee: the Fed's rate-setting body, which meets eight times a year and publishes a statement and projections.), the Fed's rate-setting body, raised the range for the federal funds rate by a quarter point, to 3.75–4.00%, with every participant in support . Most participants judged that another increase would likely be appropriate by year end, and risks to the inflation forecast were seen as skewed toward higher inflation . The committee meets on 27–28 October, ending with a press conference . The October statement has not been issued .
Inflation. The September consumer price index (CPI), already published, rose 0.6% on the month and stood 3.7% higher on the year; excluding food and energy, it rose 0.3% on the month and 2.6% on the year . The October CPI is due on 10 November , after the window closes. The only Bureau of Labor Statistics (BLS) release inside the window is the third-quarter Employment Cost Index (ECI) on 30 October , which measures what employers pay for labor, not what consumers pay for goods .
The spenders. Capital expenditure, or capexCapex: capital expenditure: money a company spends on buildings and equipment rather than day-to-day costs. — spending on data centers, servers and other long-lived equipment — ran higher in the second quarter than a year earlier at Alphabet, Amazon and Meta, each of which last reported in late July . None has announced a third-quarter results date, and the July releases do not give one , so it is not established that any of them reports inside the window.
Names that bear on this
Alphabet reported second-quarter results on 22 July: purchases of property and equipment, its measure of capital spending, of $44.9bn against $22.4bn a year earlier . That spending pace supports the frame. Against it, the company's range for the year came from its first-quarter call and was not restated in the July release . The third-quarter results release resolves it; the July release does not date it, and the company has not announced one .
Amazon reported second-quarter results on 30 July: purchases of property and equipment of $54.2bn against $32.2bn a year earlier, rising sales at Amazon Web Services (AWS), its cloud business, and free cash flow — cash from operations left after capital spending — that was an outflow over the trailing year, which the company attributed to AI-related investment . It cuts both ways: the spending is rising and so is the cash going out. Its plan for the year dates from February and was not restated in July . The third-quarter results release resolves it, including the quarter's operating-income guidance given in July ; the July release does not date it, and the company has not announced one .
Meta reported second-quarter results on 29 July: purchases of property and equipment of $30.1bn in the quarter, within year-to-date purchases of $49.1bn, and a capex range for the year narrowed to $130–145bn . The narrowing cuts for the frame; falling income from operations cuts against it . The third-quarter results release resolves it; the July release does not date it, and the company has not announced one .
The case for
The strongest case that the frame holds is that none of the tests asks for anything new. Each asks only that a position already set down in a set of minutes or a results release be repeated.
Start with the Fed. The committee raised rates in September with every participant in support, and most judged that another increase would likely be appropriate by year end . The test asks for no move at the October meeting; it asks only that the statement not walk back the September position. A committee that acted unanimously in September would need fresh evidence to back away from it, and the price data published since has not supplied any: the September CPI ran hot on the month, and the minutes had already judged inflation risks skewed higher .
The inflation bar is narrow because the calendar makes it so. The ECI is the only BLS release in the window and measures labor costs rather than consumer prices , and the October CPI comes after the window closes . The ECI has only to avoid showing labor costs easing by enough to contradict the judgment in the minutes.
On spending, the plans ask for continuity rather than a leap at Alphabet and Amazon. Alphabet's range of $180–190bn for the year needs roughly $50bn a quarter in the second half, against $44.9bn in the second quarter . Amazon planned about $200bn for the year and had spent $98.4bn by June, $54.2bn of it in the second quarter alone ; the second half at that pace, added to what was already spent, would meet the plan. The demand beneath the spending looks firm. Google Cloud revenue rose 82% and AWS sales rose 37% , and Meta narrowed its range for the year by lifting the floor , which is a commitment, not a hedge.
An honest case concedes its gaps, and conceding them does not change its core. Meta's range implies $79–94bn in the second half , which asks for acceleration rather than continuity. Neither Alphabet's range nor Amazon's was restated in the July releases . No third-quarter reporting dates are confirmed, so the Fed statement and the ECI are the only tests certain to arrive on time — and the statement is the test the September minutes make easiest to pass.
The case against
The strongest case against is that the last week of October cannot deliver the clean validation the question implies, and that what it does deliver is thinner evidence than it appears. The test has a timing problem and an evidence problem.
The inflation checkpoint is incomplete. The committee meets before the October CPI is published, and the September CPI is already out . The one scheduled BLS release at the end of the window is the ECI, a labor-cost measure, not a fresh consumer-price reading . Even a favorable ECI could not establish broad consumer-price easing by the deadline, and the CPI evidence already in hand is mixed, with the headline running well ahead of the core on both the monthly and the annual measure .
The Fed's documented starting point is tightening, not easing. Every participant supported the September increase, most expected that another would likely be appropriate by year end, and inflation risks were judged skewed higher . An October statement that merely avoids further escalation would not show that those concerns had gone. Treating softer wording as confirmation would set a lower standard of evidence than the September minutes warrant.
Spending commitments are not evidence of adequate returns. The strongest counterweight deserves its due: Google Cloud revenue grew 82%, AWS sales grew 37%, and AWS operating income was $16.6bn on AWS sales of $42.2bn . But Amazon also reported free cash flow over the trailing year that was an outflow of $7.6bn, which it attributed to AI-related investment , and Meta's revenue rose 28% while its income from operations fell 8% from a year earlier, to $18.8bn, leaving an operating margin of 31% . None of this proves the investment is failing. It does show why maintained budgets alone cannot validate its economics.
The spending hurdle is also less straightforward than the companies' stated ranges suggest. Alphabet's range came from its first-quarter call and Amazon's figure from February , and neither was restated in the July releases ; a plan that has not been repeated is weaker evidence than one that has. Meta's range implies second-half spending of about $79–94bn . Set that second-half total against $31.08bn in the second quarter : the range requires a step up, not a continuation of the second-quarter pace. That is arithmetic on the company's figures, not company guidance.
Nor are third-quarter reporting dates confirmed for Alphabet, Amazon or Meta. The honest objection is not that the frame must fail, but that the week may not bring the releases, or measures strong enough, to show it is right.
What we don't know, and when
The meeting ends on 28 October with a press conference ; until the statement is out, whether it repeats, softens or hardens the September position is unknown. The third-quarter ECI arrives on 30 October and will show whether labor costs bear out the minutes' inflation judgment, though not what consumers paid. The October CPI follows on 10 November , after the window has shut; it, not anything inside the week, is the first fresh consumer-price reading since September.
The largest unknown is the spenders. None of Alphabet, Amazon or Meta has announced a third-quarter results date, and their July releases do not give one ; each company's own announcement of a results date will settle whether the spending leg of the test falls inside the week or after it. Until those releases arrive, neither Alphabet's range nor Amazon's plan has been restated since the first half, and Meta's narrowed range has not been tested against a third quarter of spending.
What would change our mind
Our answer stands: the frame holds through the week if the Fed does not walk back its September position, the ECI does not contradict the minutes' inflation judgment and no large spender cuts its range for the year, but the week can break the frame and cannot confirm it. For now nothing changes and the frame stands, because the only evidence certain to arrive inside the window is the Fed statement and the ECI, and nothing published since September has contradicted the position the committee took then . Three things would change that. A statement at the close of the October meeting that walks back the September position would trip it at once. An ECI showing labor costs easing sharply enough to undercut the minutes' view of inflation risks would weaken it. And a third-quarter results release from Alphabet, Amazon or Meta that cuts its range for the year, or again fails to restate it, would take the spending leg away. Any of these would be weighed against Spare computing capacity does not stop new power construction, The Fed under Warsh keeps raising rates into this inflation and Spending on the AI build is running ahead of its revenue when it lands; until then the House View is unchanged.
How this piece was made
Sources. 11 numbered sources, in the Record.
Related claims
Tell the publisher
Something we should look at? Suggest a company, a topic or a feature.
Record
Written according to the Grant Bishop standard — machine-researched, machine-checked, human-judged; then read and committed by the publisher. Every number traces to a filing or a dated market record; every claim carries a date and what would prove it wrong.
- Drafted by
- Claude Opus 5.5; counter-case by GPT-6 Astra under Grant Bishop's methodology
- Verification
- multi-source verified, machine-drafted, human-gated
- As of
- 2026-10-21
- Published
- Oct 22, 2026, 14:57 ET · 2026-10-22 18:57 UTC · changelog
- Bears on
- What would change this
- A statement at the close of the 27–28 October Fed meeting that walks back the September position, an Employment Cost Index showing labour costs easing sharply enough to undercut the September minutes' view of inflation risks, or a third-quarter results release from Alphabet, Amazon or Meta that cuts or fails to restate its spending range for the year, would change this conclusion.
- Position disclosure
- The publisher holds index-level U.S. equity exposure and no direct position in any single name discussed.
- Not advice
- General commentary for a general audience. Not advice for your situation.