As we read it, the week of 26 October to 1 November 2026 showed the AI build running ahead of the cash that pays for it. Amazon, Alphabet, Microsoft, Meta and Oracle spent about $31.8 billion more on property and equipment in their latest quarter than their operations brought in, by our arithmetic from their 10-Q filings, against about $61.4 billion left over a year earlier. Our count of the commitments eight companies keep off their balance sheets reached about $3.94 trillion.
Around those two lines, markets repriced the biggest builders, Anthropic’s S-1 opened its books, power and permits set the pace of new capacity, and the labs and their backers turned to money and politics. The six sections below are our grouping of the week’s FourWeekMBA coverage; each item links to the full piece, where the primary source is cited.
Business Pill · FREE CASH FLOW
An evergreen Business Engineer explainer on free cash flow: profit on paper versus cash in the bank. It teaches the general idea only and says nothing about any company in this roundup.
The key insight: As we read it, the week’s coverage showed the AI build being paid for in three ways at once: out of operating cash, which no longer covered the capital spending of three of the four biggest builders; with new debt and equity, about $325 billion lined up at Alphabet, Amazon and Meta in 2026 by our arithmetic; and off the balance sheet, where our count of commitments reached about $3.94 trillion.
As we read it, the earnings week put the same pattern in four filings. Alphabet said Google Cloud revenue rose 103% to $30.7 billion in the third quarter, while its purchases of property and equipment reached $51.4 billion and free cash flow was negative $9.1 billion by Alphabet’s own reconciliation. Microsoft reported Azure revenue up 48%; by our arithmetic its free cash flow fell to $18.6 billion from $25.7 billion. Read: the $93 billion swing, Alphabet’s quarter and Microsoft’s quarter.
Amazon said AWS segment sales rose 39% to $46.0 billion. Its cash flow statement shows $77.3 billion of purchases of property and equipment against $42.4 billion of operating cash flow, about 1.8 times by our arithmetic, and $48.1 billion of proceeds from long-term debt in the quarter. Meta’s free cash flow went from $12.39 billion in the first quarter, by our arithmetic from its filings, to negative $2.12 billion in the third, as Meta reported on 21 October. Read: AWS and Amazon’s capex, Amazon’s $48.1 billion of borrowing and Meta’s cash flow.
Alphabet, Amazon and Meta lined up about $325 billion of new debt and equity in 2026, by our arithmetic from their filings. Microsoft finance chief Amy Hood said calendar-2026 capital spending “should be approximately $170 billion due to the 3P capacity delays noted earlier”, and that Microsoft “will access additional sources of funding when we believe it is attractive to do so.” Read: the $325 billion and Hood’s remarks.
Alphabet, Amazon, Microsoft and Meta spent about $201.9 billion on property and equipment in the quarter to 30 September 2026 against about $175.5 billion of operating cash flow, a gap of about $26.4 billion, by our arithmetic from their earnings releases as reported in our piece on the $325 billion the builders lined up. Microsoft was the only one of the four with operating cash above capex.Operating cash flow against purchases of property and equipment (Microsoft: additions) in the quarter to 30 September 2026: Amazon $42.4 billion against $77.3 billion, Alphabet $42.3 billion against $51.4 billion, Meta $32.8 billion against $33.9 billion and Microsoft $58.0 billion against $39.4 billion, as reported in our piece on the $325 billion the builders lined up. Capex is gross and excludes finance leases; Amazon’s 1.8x multiple of capex to operating cash is our arithmetic.
Markets Repriced the Builders
Meta, Alphabet, Microsoft and Amazon lost about $665 billion of market value in the first trading session after each released its results, by our arithmetic from daily closing prices and 10-Q share counts, and Alphabet accounted for $496.5 billion of it, about 75%. Google Search & other brought in 49.7% of Alphabet’s revenue, by our arithmetic the first quarter below half in the 32 quarters of figures we read.
In credit markets, the cost to protect against a debt default at Amazon, Alphabet, Microsoft, Meta and Oracle widened to the highest levels going back to 2021, The Wall Street Journal reported, citing Solve. S&P Global Ratings revised its outlook on Meta to negative from stable and kept the AA- rating. Gartner’s Hype Cycle for Artificial Intelligence, 2026 places generative AI in the Trough of Disillusionment, and the firm said on 14 October that more than $3.5 trillion of AI infrastructure spending could be exposed to impairment by the end of 2030. Read: the debt insurance, S&P on Meta and Gartner’s trough.
Market value of the world’s ten most valuable companies at the 30 September and 30 October 2026 closes, from daily closing prices and the share counts on each company’s latest SEC filing cover (Saudi Aramco: the Yahoo Finance share count and its 29 October close), as reported in our piece on the $1.52 trillion October fall; Alphabet lost $780.2 billion. All values and changes are our arithmetic.Animated: combined market value of Alphabet, Amazon, Meta and Microsoft against the 23 October close, by trading day to 30 October, from daily closes and 10-Q share counts; our arithmetic. This calendar-week view (minus $662 billion) differs from the first-session-after-earnings total in the text because Meta reported on 21 October.
The Bills Moved Off the Balance Sheet
Eight companies list about $3.94 trillion of commitments that are not on their balance sheets, by our arithmetic, and Alphabet’s lines reached $1.06 trillion at 30 September. A chart credited to ZeroHedge puts the same eight at $4.11 trillion; most of the $161.7 billion gap is Nvidia’s guarantees, capped at $105 billion, for an OpenAI affiliate’s leases at an Ohio campus. Microsoft said it had leases, primarily for datacenters, not yet commenced of $364.6 billion. Read: our $3.94 trillion count, the two $4 trillion counts and Microsoft’s leases.
Alphabet and Meta disclose up to $133.4 billion of contingent support for the companies that build and power their data centers, by our arithmetic from their 10-Qs; Microsoft, Amazon and Oracle disclose no comparable arrangement in a full-text search. When Meta put its Hyperion campus into a joint venture, it contributed $4.30 billion of assets and received $2.55 billion in cash, about 59% by our arithmetic, and the $27.3 billion bond that financed the campus carries a 6.581% coupon, IFR reported. Read: the $133 billion of backstops, Hyperion’s 59% and the 6.581% bond.
A Bloomberg Intelligence chart shared on X projects a gross AI funding gap of about $330 billion in fiscal 2027. By our reading of its bars, Oracle and SpaceX account for about 52% of the roughly $790 billion gap it shows across fiscal 2026 to 2030. Read: the funding gap.
Anthropic Opened Its Books
Anthropic’s S-1, filed on 26 October, puts its revenue for the three months ended June 30, 2026 at $11,555 million and its cash, cash equivalents and short-term investments at $131.7 billion, by our arithmetic. On a like-for-like basis rebuilt from SEC filings, Anthropic’s 2025 revenue was about 112 times its 2023 revenue, by our arithmetic, against about 40 times for Nvidia. Read: the S-1’s numbers and the 112x.
The filing says Anthropic “entered into new, and amended existing, lease and commercial agreements primarily to secure additional compute and infrastructure capacity, with aggregate future minimum payment obligations of approximately $245 billion”, about 1.9 times its cash by our arithmetic. Blackstone has told investors that Anthropic may not secure an investment-grade credit rating after its IPO, the Financial Times reported. Read: the $245 billion and the rating.
We also counted the warnings. Anthropic’s S-1 lists 88 risk factors, and the longest, at 4,066 words or 6.1%, is about chips, electrical power and data-centre capacity, by our count. Across nine filings, leaving aside a catch-all Other group, cyber, safety and outage risks get the most words in 4 of the 9, and Alphabet’s quarterly report adds: “During model testing, we have sometimes observed unexpected behavior, including autonomous agents taking actions outside of their intended scope”. Read: the S-1’s risk factors, what the filings fear and Alphabet’s agents.
Power Set the Pace
Alphabet, Amazon, Meta and Microsoft have guided 2026 capital spending of about $730 billion at the midpoints, about 1.19 times the roughly $615 billion of all US energy investment in the IEA’s figures, by our arithmetic. BEA data put private investment in US data centers at $83.0 billion at an annual rate in the third quarter, the highest since the series began in 2020. Read: the $730 billion and the $83 billion pace.
Elon Musk wrote on X that there will be “a total of around 1.2M Blackwells at Colossus 2 by end of year”; in GB300 NVL72 racks at HPE’s published spec, they would draw about 2.2 GW before cooling, by our arithmetic. Chamath Palihapitiya, quoting that post, wrote that “Power is THE binding constraint in AI/SI.” A federal judge in Mississippi denied a motion by X.AI Corp. and MZX Tech LLC to dismiss an NAACP lawsuit over the gas-turbine plant behind Colossus 2. Read: the 2.2 GW, Chamath on power and the NAACP suit.
Google’s head of data center energy said data centers in space could cost up to 100 times as much as those on Earth, The Information reported. South Korea exported $56.44 billion of semiconductors in October, up 259.4%, according to its trade ministry; by our arithmetic chips were 49.0% of its exports. Read: space data centers and Korea’s chip exports.
Labs, Money and Politics
The labs raised and spent on different terms. OpenAI “Raises Less Than Half of $30 Billion Target in Latest Round”, The Wall Street Journal reported, in a round that “values it at $1.25 trillion, less than its $1.4 trillion target”. Aramco has taken a 49% stake in HUMAIN, Saudi Arabia’s state AI company, the Public Investment Fund said. Read: OpenAI’s round and Aramco and HUMAIN.
Usage kept climbing as buyers looked for cheaper ways in. Amazon said Bedrock token volume grew “nearly 17x year-over-year”, and Revolut says moving 2,000 engineers onto self-hosted open models cut its AI coding costs 90%, StartupHub.ai reported from a talk by Revolut’s David Abitbol. Screenshots on X showed a model labelled gpt-6.2-astra and the name Claude Mythos 5.5; OpenAI has not announced a GPT-6.2 model, and when we fetched Anthropic’s Mythos page it read 5.1. Read: Bedrock’s 17x, Revolut’s switch, the GPT-6.2 screenshots and the Mythos 5.5 trace.
Two days before the US midterms, data centers were on the ballot in the ads. Campaigns and outside groups spent $173 million on almost entirely negative ads about data centers in the general election, according to AdImpact data reported by Axios, and OpenAI plans to argue that its infrastructure should be called “AI compute campuses”, Axios reported. By our arithmetic, donors from tech, AI and crypto account for 43.0% of the money on a New York Times list of the 20 biggest midterm donors. Read: the $173 million of ads, OpenAI’s rebrand and the donors.
What the Week Established
On the documents read for our coverage: the five builders’ 10-Qs show, in total, about $31.8 billion more spent on property and equipment than their operations brought in, by our arithmetic, and Amazon’s shows $48.1 billion of new long-term debt in the quarter. Alphabet’s and Meta’s filings disclose up to $133.4 billion of backstops and guarantees for their partners, and Microsoft’s $364.6 billion of leases not yet commenced. Anthropic’s S-1 puts its quarterly revenue at $11,555 million and its new compute obligations at approximately $245 billion.
In markets: about $665 billion of market value lost by four builders in the first session after each one’s results, by our arithmetic, debt-insurance costs at their highest since 2021 per Solve, and a negative outlook on Meta from S&P.
The Structural Read
The cash side comes from the filings: Amazon’s cash flow statement shows $77.3 billion of purchases of property and equipment against $42.4 billion of operating cash flow, and Alphabet’s own reconciliation puts its free cash flow at negative $9.1 billion.
The off-book side comes from the notes: Alphabet and Meta disclose up to $133.4 billion of contingent support for their data-center partners, and Microsoft says it had $364.6 billion of leases, primarily for datacenters, not yet commenced.
The open question is the price: markets took about $665 billion off four builders in the first session after their results, by our arithmetic, and Solve’s data, as the Journal reported, put their debt-insurance costs at the highest since 2021.
Microsoft finance chief Amy Hood, on the company’s 28 October 2026 earnings call, as quoted in our piece
“to complement cash generated by the business and preserve financial flexibility, we will access additional sources of funding when we believe it is attractive to do so.”
Three Implications
LENDERS AND BOND BUYERS Amazon took $48.1 billion of new long-term debt in the quarter, the Hyperion bond pays 6.581%, IFR reported, and S&P revised Meta’s outlook to negative while keeping its AA- rating.
PARTNERS AND LANDLORDS Meta received $2.55 billion in cash for the $4.30 billion of assets it put into Hyperion, by our arithmetic about 59%, and Alphabet’s backstops and power-equipment guarantees reach $91.9 billion.
MODEL LABS Anthropic’s S-1 lists approximately $245 billion of new compute obligations, and the Journal reported that OpenAI raised less than half of its $30 billion target.
The Business Engineer Lens
This story maps onto the Business Engineer framework The AI Capex Map.
The framework puts it this way: “the capex race is no longer a corporate spending cycle. It is a sovereign-scale financing problem dressed in tech-company clothing — and the financial architecture is starting to invert.”
As we read it, this week’s coverage shows that architecture in documents: Alphabet, Amazon and Meta lined up about $325 billion of new debt and equity in 2026, by our arithmetic, Alphabet and Meta disclose up to $133.4 billion of backstops and guarantees for their partners, and our count of off-balance-sheet commitments reached about $3.94 trillion.
What Is Not Established
This roundup adds no new reporting. Every figure above is taken from the linked FourWeekMBA story, each of which was built from its own primary source. The six sections, and the selection within them, are our grouping, not a finding; the week’s other stories are not covered here.
We do not claim that any of these stories caused another, that the spending will or will not pay off, or that any guarantee will be called. Several items are reports citing unnamed people, charts shared on X that we could not read in full, screenshots we could not verify, or estimates that their authors present as their own, and none of them is a forecast by FourWeekMBA.
This Week In The AI Supercycle
Anthropic’s Reality Check, The Capital Squeeze, & The New World Order
This week’s issue: for two years the AI build-out was priced as a growth story. This week it started being priced as a credit story.
FourWeekMBA’s coverage of 26 October to 1 November 2026 showed, as we read it, the AI build running ahead of its cash. The five biggest builders spent about $31.8 billion more on property and equipment than their operations brought in, by our arithmetic, and our count of off-balance-sheet commitments reached about $3.94 trillion.
In the same week, the four biggest builders lost about $665 billion of market value after their results, Anthropic’s S-1 listed approximately $245 billion of new compute obligations, and power, permits and politics shaped the terms for new capacity.
95,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. This weekly roundup is built only from FourWeekMBA’s own news coverage of 26 October to 1 November 2026, linked above; every story was built from its own primary source, which it cites and links. We re-read each piece on 1 November 2026 and took every figure here from it. The six sections are our own grouping. Nothing here is a forecast, and nothing here is financial or investment advice.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
