As we read it, the week of 2 to 8 November 2026 showed the AI build securing its own power while credit markets put a price on it. Anthropic bought commitments for 1.7 gigawatts of heavy-duty gas-fired generators for $3.4 billion, the Financial Times reported, and Goldman Sachs’s credit strategy team tallied hyperscaler debt supply at $296 billion so far this year against its own $250 billion forecast for 2026, MarketWatch reported, about 18% over by our arithmetic. Read: Anthropic’s generator purchase and the $296 billion tally.
Around those two lines, voters, towns and lawmakers weighed in on data centers and advanced AI, AI stocks split between memory makers and the rest, Anthropic’s amended S-1 showed whose chips it has committed to, and Apple and China each drew a closer look. The six sections below are our grouping of the week’s FourWeekMBA coverage; each figure links to the story it comes from.
Business Pill · CREDIT DEFAULT SWAPS
An evergreen Business Engineer explainer on credit default swaps: what the spread tells you. 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 reaching for two scarce inputs at once: power it can own, as in Anthropic’s $3.4 billion of generator commitments reported by the FT, and money it has to borrow, as in the $296 billion of hyperscaler debt supply Goldman tallied, per MarketWatch. Credit markets put prices on both, from roughly 139 basis points on Nvidia’s swaps to around 1,155 on CoreWeave’s, as CNBC reported.
The FT calls Anthropic’s deal the largest known single purchase of power-generation equipment by an AI lab. Anthropic bought the commitments from OnePWR Solutions, an equipment supplier based in San Antonio, Texas, according to the FT; by our arithmetic, $3.4 billion for 1.7 gigawatts works out to about $2.0 million per megawatt. The deal is about 99% of the $3.421 billion of Power Generation sales Caterpillar reported for the third quarter, by our arithmetic, and Caterpillar’s filing shows firm backlog of approximately $78.1 billion at 30 September. Read: the Caterpillar comparison.
A public hearing notice posted by the City of Starbase, Texas says SpaceX “proposes to construct, own, and operate a gas-fired power plant and super-compute facility” on an approximately 444-acre tract, with estimated capital investment for initial phases of approximately $24 billion; the city commission will consider a property tax abatement at a meeting on 9 December, the notice says. Read: the Starbase notice.
In China, SemiAnalysis says Envision’s Galaxy AI campus in Ulanqab has two wind farms of its own with a combined 420MW of installed capacity, about 1.75 times the 240MW load of its first two phases. By our arithmetic, Anthropic’s 1.7GW of generators is about 4 times Galaxy’s 420MW of wind. Read: Galaxy’s wind and the two ways to buy power.
Hyperscaler debt supply so far in 2026 against Goldman Sachs’s $250 billion forecast for the year, as Goldman’s credit strategy team tallied it in a 5 November report reported by MarketWatch; the $46 billion gap and 18% are our arithmetic. Goldman expects $440 billion in 2027, according to MarketWatch.
Voters, Towns and Lawmakers Weighed In
On unofficial counts after the 3 November elections, limits on data centers were ahead in all 35 local votes with results posted, according to our correction of Ballotpedia’s table against local reports. Cleanview says nine states now have a governor or governor-elect who has promised or imposed a pause on new data centers, and those states hold 158.6 GW of planned capacity, 41% of the US pipeline; it also wrote that “OpenAI’s entire planned pipeline is located in states with a new governor who backed a pause on the campaign trail”. Read: the local votes and the pause states.
Across the UK, 315 data centres are queueing to connect to the power grid, representing 73 gigawatts of demand compared with the country’s peak energy demand of 45GW, the Guardian reported; a coalition of 75 organisations is set to march in Edinburgh on 14 November, Sky News reported. Growth minister Dan Tomlinson said the UK will not slow down on AI, according to the FT. Read: the UK queue and the minister.
The pushback also reached advanced AI. The Bloomington, Indiana, Common Council voted on 4 November to ban the distribution and deployment of superintelligent AI inside the city, and New York Assemblymember Alex Bores wrote on X: “Congress should ban RSI, this year if possible”. Read: Bloomington’s ordinance and Bores’s post.
In the midterms, 97.5% of the money the Public First network of AI-safety super PACs reported spending went into AP-called races that ended its way, against 52.6% for the Leading the Future network, by our count from FEC filings and AP race calls. California state Sen. Scott Wiener, the author of a state AI safety law, defeated San Francisco Supervisor Connie Chan for Nancy Pelosi’s House seat, KTVU reported, citing the AP. Read: the PAC scorecard and Wiener’s race.
Credit Put a Price on the Build
Goldman now expects another $440 billion in hyperscaler debt supply in 2027, according to MarketWatch, 1.49 times this year’s tally so far by our arithmetic. Investment-grade AI corporate bonds traded about 148 basis points above Treasuries on 5 November, against 77 for the rest of the market, according to BofA Global data cited by MarketWatch. Meta priced $15.0 billion of notes in five series, according to the pricing term sheet it filed with the SEC. Read: Meta’s bond sale.
Credit default swaps on Nvidia debt are up 76% in the last month, CNBC reported, at roughly 139 basis points, and IFR reported, citing S&P Global Market Intelligence data, that they trade at levels “more associated with a Double B rating – one notch below investment grade – than the Double A rating it holds”. CoreWeave’s five-year swaps traded around 1,155 basis points, according to Deutsche Bank data cited by CNBC, about 8.3 times Nvidia’s by our arithmetic. Read: Nvidia’s swaps and the spreads side by side.
Six US-listed companies that rent out or host AI computing capacity have $32.1 billion of debt principal falling due through 2030, according to the repayment schedules in their latest quarterly filings, against $16.1 billion of cash and cash equivalents, about $2 of principal due for every $1 of cash by our arithmetic. Read: the neocloud maturities.
AI Stocks Split
The 12 AI stocks we tracked closed a combined $1,447.4 billion lower on 5 November than the day before, by our arithmetic from Nasdaq’s official closes, and Nvidia closed at $181.60, down 9.26%. Intel’s shares closed at $56.28 on 6 November, down 53.2% from 30 September, a fall in market value of roughly $340 billion by our arithmetic. Read: the 5 November sell-off and Intel’s quarter.
For the year, Micron’s shares were up 141.76% and AMD’s 109.68% at the 6 November close, according to Nasdaq.com closing prices, while the simple average change for Amazon, Microsoft, Meta and Alphabet was +1.4%, by our arithmetic. Over the week from 29 October, the three biggest memory makers fell between 18.7% and 26.1%, by our arithmetic, while Counterpoint put third-quarter DRAM revenue at nearly $200 billion. Read: the year-to-date split and the memory makers.
Equity long-short hedge funds made their largest active reduction of positions in more than a decade in October, according to Morgan Stanley data cited in republished accounts. Bank of America strategists led by Michael Hartnett say the US midterm results mark peak AI equity concentration and favour tech bonds over tech stocks, Investing.com reported. Read: the hedge funds and the BofA note.
Animated: Intel’s daily closing price from 30 September to 6 November 2026, from Nasdaq.com historical closes, as in our piece on Intel’s quarter: $120.23 to $56.28, −53.2% by our arithmetic, including a 28.6% one-day fall on 30 October.Share-price change from the 31 December 2025 close to the 6 November 2026 close, from Nasdaq.com closing prices as our year-to-date piece reports them: Micron +141.76%, AMD +109.68%, Oracle −40.43%, and +1.4% for the simple average of Amazon, Microsoft, Meta and Alphabet. Indexing and the average are our arithmetic.
Anthropic Readied Its Listing
Anthropic is reported to be seeking $80 billion to $100 billion in its IPO, a range that sits on either side of SpaceX’s $86.25 billion June offering, by our arithmetic from SpaceX’s SEC filings; the range comes from the FT, as summarised by Techmeme. Read: the SpaceX yardstick.
Google TPUs account for $282.5 billion, or 43.9% by our arithmetic, of the $643.8 billion in compute contracts with stated dollar amounts in the amended S-1 Anthropic filed on 2 November. The same filing says Anthropic expects to pay SpaceX’s AI unit approximately $3.1 billion a month for computing capacity once all of it has been delivered. Read: the chip split and the SpaceX bill.
Apple Was Pressed, China Was Measured, Nvidia Was Asked In
Apple CEO John Ternus pushed software chief Craig Federighi to accelerate the company’s AI road map, Bloomberg’s Mark Gurman reported in the Power On newsletter. Apple’s annual report shows research and development spending of $45.765 billion, up 32%. Gurman wrote that the company “could make Siri more like an agent as early as next year”, and that Mike Rockwell, who led the development of Siri AI, “continues to seriously consider stepping down”.
SemiAnalysis says the Galaxy campus delivered its first 120 MW phase within a year at a fifth of US costs, that China’s data center rents have roughly halved over the past five years to about $40/kW/month excluding power, and that a live 5MW hall already runs 800V DC power to the rack. A trial sent 64 Tbps in each direction over a 1,007.48-kilometer hollow-core fiber link, the Global Times reported, citing CCTV; China Telecom’s research institute, YOFC and ZTE ran it, Sina Finance reported. Read: Galaxy, the rents, the 800V hall and the fiber trial.
Three House Republican committee chairs asked Nvidia CEO Jensen Huang to brief their committees on 18 November, the day after Nvidia reports third-quarter results, according to a 5 November release from the House Select Committee on China; the request is for a joint, closed-door briefing, not a subpoena or a public hearing. Read: the briefing request.
What the Week Established
On the documents read for our coverage: Caterpillar’s filing shows $3.421 billion of third-quarter Power Generation sales, the Starbase notice puts about $24 billion on SpaceX’s initial phases, Anthropic’s amended S-1 lists $643.8 billion of compute contracts with stated amounts, and six AI-hosting companies list $32.1 billion of principal due through 2030.
In markets: hyperscaler debt supply about 18% over Goldman’s full-year forecast, by our arithmetic from MarketWatch’s report, Nvidia’s swaps up 76% in a month per CNBC, Micron up 141.76% for the year and Intel down 53.2% for the quarter on Nasdaq closes.
Credit default swap spreads as CNBC reported them on 6 November 2026, with CoreWeave’s five-year spread from Deutsche Bank data cited by CNBC, as our piece on the spreads side by side reports them. The 8.3 times (1,155 / 139) is our arithmetic.
The Structural Read
The power side comes from documents: the FT’s report of 1.7 gigawatts of generators bought from OnePWR, Caterpillar’s 10-Q with $3.421 billion of third-quarter Power Generation sales, the City of Starbase notice with about $24 billion for SpaceX’s initial phases, and SemiAnalysis’s count of 420MW of wind at Galaxy.
The money side comes from reported tallies and filings: Goldman’s $296 billion against a $250 billion forecast, per MarketWatch, Meta’s $15.0 billion term sheet, and six AI-hosting companies with $32.1 billion of principal due through 2030 against $16.1 billion of cash.
The pushback side comes from votes and texts: limits ahead in all 35 local data center votes with results posted, on unofficial counts, Cleanview’s 158.6 GW in nine pause states, Bloomington’s ordinance and Bores’s post on X.
Deutsche Bank’s Jim Reid, in a note cited by CNBC, as quoted in our piece
“We’ve had plenty of warnings from the credit space”
Three Implications
LENDERS AND SWAP TRADERS CNBC reported Nvidia’s swaps up 76% in a month to roughly 139 basis points, and BofA Global data put investment-grade AI bonds about 148 basis points over Treasuries against 77 for the rest of the market, per MarketWatch.
EQUIPMENT MAKERS Caterpillar reported $3.421 billion of third-quarter Power Generation sales and firm backlog of approximately $78.1 billion at 30 September, according to its filing.
TOWNS AND STATES Cleanview says nine states with a pause-backing governor or governor-elect hold 158.6 GW of planned capacity, 41% of the US pipeline, and the City of Starbase will consider a tax abatement for SpaceX on 9 December.
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 financing in prices: hyperscaler debt supply about 18% over Goldman’s full-year forecast by our arithmetic from MarketWatch’s report, and swap spreads that CNBC reported from roughly 139 basis points at Nvidia to around 1,155 at CoreWeave.
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 ban, pause or briefing will change a company’s plans. Several items are reports citing unnamed people, newsletter or paywalled articles summarised by other outlets, or posts on X; each linked story says which.
This Week In The AI Supercycle
The Great Convergence of Capital, Power, and Geopolitics
This week’s issue: for two years the AI build-out was priced as a supply story, and last week as a credit story. This week it was priced as a permission story, and the answer arrived as a cost, not a ban.
FourWeekMBA’s coverage of 2 to 8 November 2026 showed, as we read it, AI builders buying power equipment of their own while lenders and swap traders put a price on the build. Anthropic bought commitments for 1.7 gigawatts of gas-fired generators for $3.4 billion, the FT reported, and hyperscaler debt supply reached $296 billion against a $250 billion forecast, per Goldman via MarketWatch.
In the same week, local votes, state pause pledges and a city ordinance pushed back on data centers and advanced AI, memory stocks and the hyperscalers split, and Anthropic’s filing showed 43.9% of its stated compute contracts tied to Google TPUs, by our arithmetic.
A note on sourcing. This weekly roundup is built only from FourWeekMBA’s own news coverage of 2 to 8 November 2026, linked above; every story was built from its own primary source, which it cites and links. We re-read each piece on 8 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.
