Anthropic’s Growth is Slowing: Is AI Capex Next? | The Weekly Wrap

Anthropic

Across the shows: 71 speakers in other episodes, 80 bullish and 59 bearish calls, past 30 days

• Steve Eisman · The Real Eisman PlaybookBearish

  Anthropic's roughly $800B of compute commitments leave no margin for error and could make costs explode.

  Eisman totals Anthropic's disclosed compute and equipment commitments from the S-1: Microsoft $45B, Google $115B, TPU providers including Google $150B, Amazon $110B, SpaceX $121B, AMD $20B, Nscale $45B, Lambda $35B and Broadcom $161B. These add up to about $800B of future commitments. He says this is enormous, that there is no margin for error, and that compute costs could explode starting in 2027. The argument is that fixed commitments of this size are a risk if revenue growth slows, which he sees happening.

  “It all adds up to 800,000,000,000 in future commitments and that is enormous. There is no margin for error here.”at 4:39

• Steve Eisman · The Real Eisman PlaybookBearish

  Anthropic's revenue growth has slowed sharply after token maxing ended, with weaker revenue-to-compute efficiency.

  Eisman argues Anthropic's growth is slowing now that enterprise 'token maxing' (companies telling staff to use AI regardless of cost) has ended. Sequential quarterly revenue growth fell from 147% in Q2 2026 to 45% in Q3 2026. The revenue-to-compute-expense ratio fell from 2.08x in Q2 to 1.70x in Q3, after rising from 0.69x in Q4 2025 during the token-maxing surge. He adds that Anthropic cut token prices by up to 50% in October to spark demand, and he doubts that higher OpenRouter token demand offsets those cuts. He notes the S-1 shows Q2 adjusted operating margin of 19% but calls that a peak-token-maxing figure. He concludes growth is still positive but slower, and says there are too many variables to predict whether Anthropic can deliver profits to investors.

  “Q3 twenty twenty six sequential revenue growth was 45% versus 147% in the second quarter.”at 7:18

The Boeing Company

Across the shows: 7 speakers in other episodes, 4 bullish and 3 bearish calls, past 30 days

• Steve Eisman · The Real Eisman PlaybookBearish

  Boeing is still struggling in its turnaround, with commercial aircraft revenue falling and disappointing cash flow guidance.

  Boeing's loss of 34 cents per share beat last year's $7.47 loss but missed the expected 21-cent loss. Revenue rose just 4%. The defense unit was the standout, with revenue up 15% to $8B. The commercial aircraft business is the all-important part, and its revenue fell 3% to $10.7B. Management gave disappointing cash flow guidance on the call. Eisman says the company is still in turnaround and still struggling. The stock fell 10% and is down over 20% this year.

  “This company is still in turnaround. It is still struggling and the stock was down 10% on Tuesday because of the miss and because management provided disappointing cash flow guidance on the call.”at 10:04

GE Vernova

Across the shows: 1 speaker in other episodes, 1 bullish and 0 bearish calls, past 30 days

• Steve Eisman · The Real Eisman PlaybookBullish

  GE Vernova's gas turbine orders surge will last as long as the AI story lasts, despite an earnings-day selloff.

  GE Vernova builds gas turbines that generate electricity, and only a handful of companies do this, so it is central to the AI data center power story. EPS of $4.81 was up 193% and orders of $29B were up 96%. Eisman says these results are very powerful and will last as long as the AI story lasts. The stock fell on earnings day, which he attributes to selling after strong performance rather than to any problem with the business.

  “These results are very powerful and will last as long as the AI story lasts.”at 12:01

Alphabet

Across the shows: 36 speakers in other episodes, 28 bullish and 17 bearish calls, past 30 days

• Steve Eisman · The Real Eisman PlaybookBearish

  AI is replacing Google search, which Eisman calls a high-moat business being swapped for a capital-heavy, low-moat one.

  Eisman says the fear that AI search would replace Google search has now shown up in the numbers. Search revenue grew 12%, down from 19% in Q1 2026. Free cash flow was negative $9.1B and EPS fell to $2.67 from $2.87, even as revenue grew 25%. He calls Google search a capital-light, high-moat business. He argues it is being replaced by a highly capital-intensive, low-moat business, because AI data-center and LLM spending is heavy and switching costs between LLMs are cheap. He concludes search is being replaced, slowly but surely.

  “The results last night indicate that Google search is being replaced. Yes, slowly, but replaced.”at 16:11

ServiceNow

Across the shows: 7 speakers in other episodes, 6 bullish and 2 bearish calls, past 30 days

• Steve Eisman · The Real Eisman PlaybookBearish

  ServiceNow's muted backlog guidance signals a sharp slowdown and revives SaaSpocalypse fears.

  ServiceNow's Q3 EPS was $1.00, up 4% and a slight miss. Revenue was $4.1B, up 21%. The issue was fourth-quarter guidance for current remaining performance obligations (its backlog measure), which implies 18.5% growth versus 19.7% expected by consensus and 21% in Q3. Eisman reads this as a sharp deceleration in future revenue. Software stocks had rallied after the summer 2025 SaaSpocalypse selloff, and this guidance brought those AI-disruption fears back. The stock fell 27%.

  “This points to a sharp deceleration in future revenue.”at 17:51

Verizon Communications

• Steve Eisman · The Real Eisman PlaybookBearish

  Starlink's push to become a fourth US mobile carrier is a telepocalypse threat to Verizon's stock.

  Eisman coined 'telepocalypse' for Elon Musk's stated goal of turning Starlink into a fourth terrestrial mobile carrier in the US, alongside AT&T, Verizon and T-Mobile. He says Musk's plan will be expensive, but that cost has never stopped Musk before. He expects Musk to keep announcing plans, so the narrative will keep gaining steam. He says the three incumbent carriers have seen their stocks under assault. He does not directly rebut a subscriber's view that cell towers are cheaper, but he notes Musk may try to rent tower space.

  “I call this the telepocalypse and it is not going away anytime soon.”at 18:46

AT&T

Across the shows: 1 speaker in other episodes, 1 bullish and 0 bearish calls, past 30 days

• Steve Eisman · The Real Eisman PlaybookBearish

  Starlink's push to become a fourth US mobile carrier is a telepocalypse threat to AT&T's stock.

  Eisman coined 'telepocalypse' for Elon Musk's stated goal of turning Starlink into a fourth terrestrial mobile carrier in the US, alongside AT&T, Verizon and T-Mobile. He says Musk's plan will be expensive, but that cost has never stopped Musk before. He expects Musk to keep announcing plans, and so the narrative will keep gaining steam. He says the three incumbent carriers have seen their stocks under assault. He does not directly rebut a subscriber's view that cell towers are cheaper, but he notes Musk may try to rent tower space.

  “I call this the telepocalypse and it is not going away anytime soon.”at 18:46

T-Mobile US

• Steve Eisman · The Real Eisman PlaybookBearish

  Despite good T-Mobile numbers, Starlink's fourth-carrier threat makes the telepocalypse a risk to its stock.

  T-Mobile's EPS of $3.22 was up 34% and it added 274,000 accounts, which was better than expected but below the 396,000 added a year earlier. Eisman calls the results fine but largely irrelevant because of Musk's plan to make Starlink a fourth US mobile carrier. He calls this the telepocalypse and says it is not going away soon. That is why the stock was flat on Thursday despite good numbers. He notes Musk's plan will be expensive, and says expense has never stopped Musk before.

  “I call this the telepocalypse and it is not going away anytime soon.”at 18:46

Quanta Services

Across the shows: 1 speaker in other episodes, 1 bullish and 0 bearish calls, past 30 days

• Steve Eisman · The Real Eisman PlaybookBullish

  Quanta Services is a major beneficiary of the AI capex story, with record backlog and strong results.

  Quanta builds new power plants for utilities, so Eisman sees it as a major beneficiary of AI capex. EPS of $6.06 was up 82%. Revenue of $12B rose 37% organically. Backlog reached a record $64B and the company raised guidance, and the stock rose 5%. Eisman calls it a strong result from a company with a clear link to AI infrastructure build-out.

  “It is a major beneficiary of the AI CapEx story. Unsurprisingly its results were very strong.”at 19:00

Eli Lilly and Company

Across the shows: 6 speakers in other episodes, 6 bullish and 0 bearish calls, past 30 days

• Steve Eisman · The Real Eisman PlaybookBullish

  Eli Lilly has clearly won the weight-loss drug wars, with revenue up 39%.

  Eisman says Eli Lilly has clearly won the weight-loss drug wars. Third-quarter EPS of $7.48 was up 7% and a beat. Revenue of $24B rose an amazing 39% year over year. The stock rose 7% on Thursday. He presents this as a clear winner in the obesity drug market, with strong growth that is not tied to AI.

  “Eli Lilly has clearly won the weight loss drug wars.”at 19:29

Intel

Across the shows: 9 speakers in other episodes, 3 bullish and 7 bearish calls, past 30 days

• Steve Eisman · The Real Eisman PlaybookBearish

  Intel's forecast of a prolonged PC slump and data center supply problems drove its 18% after-hours drop.

  Intel's Q3 EPS of 52 cents was up 126% and beat, with revenue slightly ahead. The problem was the forecast. Intel projects the PC slump will persist through at least the first half of 2027, with a 10% decline across 2027 after a 12% decline in 2026. Its data center component business also suffered supply-chain problems. Q4 revenue guidance of $15-16B is below the $17B consensus, and Q4 EPS guidance of $0.25 is well below the $0.43 consensus. The stock fell 18% after hours, after rising 150% this year.

  “Intel projects that the PC slump will persist through at least the first half of twenty twenty seven and decline 10% in all of 2027 following a 12% decline in 2026.”at 21:17

CoreWeave

Across the shows: 58 speakers in other episodes, 15 bullish and 55 bearish calls, past 30 days

• Steve Eisman · The Real Eisman PlaybookBearish

  CoreWeave's 15% effective cost of debt makes new AI data center builds questionable; it resembles a risky specialty lender.

  Eisman says CoreWeave's bonds have an effective yield of about 15% because of default risk, versus a 9% coupon. Total debt is $36B, or about $50B including operating leases. Because the model requires building more data centers and funding them with new debt, Eisman expects that debt to cost about 15%, and he questions whether that makes sense. He also compares CoreWeave to a specialty finance company: it borrows from fixed-income markets, builds data centers and rents out compute. That model is fragile if credit markets lose confidence. He notes CoreWeave competes with hyperscalers that have much larger balance sheets, and says the neocloud group sold off after Microsoft's capex report.

  “at current prices that debt will cost the company 15%. I'm not sure building an AI data center with such expensive debt makes sense.”at 23:54

AI infrastructure & data centers

• Steve Eisman · The Real Eisman Playbook

  Amazon holding its capex budget flat signals the AI capex growth cycle may be ending.

  Eisman says every hyperscaler had raised its capex budget every quarter until this one. Amazon kept its 2026 capex budget unchanged at $220B. He argues that this flat budget makes the market question whether AI capex growth is over. He says the effects would be broad, reaching servers, chips and every other type of equipment that goes into AI data centers. He adds that Amazon fell about 6% after hours, alongside a miss on earnings and revenue.

  “Now prior to this quarter, every hyperscaler increased its CapEx budget every quarter. Now Amazon is keeping its capex budget flat and the market is now going to question whether the growth in AI capex is over.”at 22:22

• Steve Eisman · The Real Eisman Playbook

  Power access, not GPUs or money, is the binding constraint that could slow AI data center construction next year.

  Eisman shares a viewpoint from Joseph, an electrical contractor building a data center. Joseph says power, not GPUs, high-bandwidth memory or money, is the binding constraint for 2027 and beyond. New transmission lines take 24+ months and substations take one to two years. Hyperscalers are truly building, but electrical connections are a long way out, so announced capex becomes stranded. Joseph thinks Amazon and Meta will feel the squeeze more than Microsoft and Google, and that CoreWeave and Oracle are far back in the power queue. Eisman says this is real insight into what is happening on the ground, and that if Joseph is right, AI data center construction will slow sometime next year. Eisman's view is conditional on that account.

  “If this subscriber is right, AI data center construction will slow down sometime next year.”at 26:52

Europe's economy

• Steve Eisman · The Real Eisman Playbook

  France's sovereign debt is too small to matter, so no one needs to own it and repayment talk is ridiculous.

  Eisman argues size is the key point. US sovereign debt is about $40T and French sovereign debt about $4T. Both countries have debt-to-GDP of about 120%, so US GDP is roughly ten times France's. France therefore does not count much in sovereign debt markets. BlackRock manages about $15T, which is bigger than all of France, so France is too small to threaten anyone in fixed income. He concludes no one needs to own French debt. He says rumblings from French politicians that France does not have to repay its debt are ridiculous.

  “France just does not count all that much in the world of sovereign debt. It's one more reason why no one needs to own French sovereign debt.”at 2:44

Open-source AI models

• Steve Eisman · The Real Eisman Playbook

  Enterprises switching to cheaper open-weight models after token maxing ended is a threat to closed labs' growth.

  Eisman says that in the first half of 2026 enterprises told employees to use AI with no regard for cost, which helped Anthropic and OpenAI. That ended in the summer when costs got out of control. Companies then had the option to switch to cheaper open-weight models. He links this to Anthropic's slowing growth, arguing that the open-weight alternative caps how much closed frontier labs can charge and grow.

  “But that ended this summer when costs got out of control and companies now had the option to switch to cheaper open weight models.”at 6:06

US consumer

• Steve Eisman · The Real Eisman Playbook

  Visa's 9% payment volume growth shows the top of the K is still carrying overall consumer spending.

  Eisman says the bottom of the K-shaped economy is struggling while the top benefits from a long bull market. Visa's EPS of $3.48 was up 17% with a slight beat. He says the payment volume data point matters most for consumer health. Visa's payment volume rose a strong 9%, which he reads as evidence that the top of the K is still carrying overall consumer spending. Mastercard's dollar payment volume rose 7%, slightly less strong.

  “Visa's results show that the top of the K is still carrying overall consumer spending.”at 10:29

Software & SaaS

• Steve Eisman · The Real Eisman Playbook

  The blanket SaaSpocalypse narrative is too blunt, since many software firms are deeply embedded in enterprises.

  Eisman argues the SaaSpocalypse narrative, that AI would destroy software, was too blunt. It will be hard for AI to destroy every software company, because many are deeply embedded in enterprises. The software group then rallied hard, and some commentators argued the term should be retired. He notes that the sector's fears came back after ServiceNow's weak backlog guidance and that the S&P/Moody's software businesses also sold off on AI fears. His view is that the threat is selective rather than sector-wide.

  “the problem with the SaaSpocalypse narrative was that it's too blunt. It will be hard for AI to destroy every software companies as many are embedded deeply in enterprises.”at 16:48