‹ Papers Your desks

AI Industry & Chips

Past edition
Oct 8 · 18 voices
i

About this edition

50 items · printed Oct 8, 06:35 ET

Tap any number in the text for its receipt — who said it, when, and the clip.

A past edition of this desk's sample paper.

Read the latest edition →

AI infrastructure enthusiasm is colliding with rising leverage, concentration and bubble concerns.

Coverage centred on an AI infrastructure investment boom, with TSMC revenue and expanding data-centre capacity presented as supportive signals 12. At the same time, sources highlighted financing strain, concentration and bubble risk, including large planned borrowings by Broadcom and SpaceX for AI chips 1536. Views diverged between optimistic claims that AI would drive growth and productivity 3245 and warnings that the boom could be financially and socially destabilising 222630.

+1318 voices

What happened

TSMC reported a 51% rise in quarterly revenue, signalling strength in the AI infrastructure buildout.1

1 of 18

Reports showed Broadcom arranging more than $50 billion and SpaceX seeking about $40 billion to buy AI chips, raising financing and credit‑risk concerns.1536

1 of 18

The 'AI Big 10' now accounts for a record 42% of US market cap, highlighting concentration in tech stocks.10

1 of 18
Show 2 moreShow fewer

Commentators diverged: some said AI justifies much higher values and growth, while others warned it's a classic bubble nearing a bursting point.263032

1 of 18

Analysts noted record equity highs driven by AI chip optimism but flagged elevated Treasury yields that could dampen demand for risky assets.45

1 of 18

“The dominance of Big Tech is extraordinary.”

TK The Kobeissi LetterOct 7 Open the source

Where they disagree

Whether the AI boom is primarily a growth opportunity or a dangerous bubble

CC

AI drives growth32

1 voice

AI is a bubble2630

Michael Novogratz

Characterised AI as a classic bubble approaching a possible 30

Ray Dalio

The desk

How the desk is calling it

5 of 7 voices lean bullish on Meta.

Where they stand · Meta

i

Reading the chart

  • bullish
  • bearish

Meta Platforms · Google Finance

An arrow is one voice's position on the day they took it. Hover or tap an arrow for who it was and what they said.

Meta Platforms, 2026-09-07 to 2026-10-07, where the desk's voices took positions$600$700Sep 15Oct 1Alex Ridgers · Sep 24 · up — Clients are doing nothing but buying Meta, spurred by the new Muse app.Keith McCullough · Sep 24 · up — He is long Meta.Dan Nathan · Sep 25 · down — Nathan says Meta's $400bn market cap gain on a personal agent release does not speak well for investors' long-term insight.James Elbaor · Sep 27 · up — Elbaor says Meta's permanent capital lets it allocate capital on its own terms, a powerful position for compounding.Walter Bloomberg · Sep 28 · up — The firm raised its Meta price target to $830 from $730 and kept a Buy rating, expecting the shares to rise.Ben Pouladian · Sep 30 · up — Meta can profitably serve AI to the rest of the world through its ad network, with Muse revenue mostly from advertising rather than subscriptions.Ed Zitron · Oct 1 · down — Meta's stock skyrocketing off Muse is insane because the product generates negative revenue and is a massive infrastructure burden.
Meta Platforms, 2026-09-16 to 2026-10-07, where the desk's voices took positions$700$750Oct 1Alex Ridgers · Sep 24 · up — Clients are doing nothing but buying Meta, spurred by the new Muse app.Keith McCullough · Sep 24 · up — He is long Meta.Dan Nathan · Sep 25 · down — Nathan says Meta's $400bn market cap gain on a personal agent release does not speak well for investors' long-term insight.James Elbaor · Sep 27 · up — Elbaor says Meta's permanent capital lets it allocate capital on its own terms, a powerful position for compounding.Walter Bloomberg · Sep 28 · up — The firm raised its Meta price target to $830 from $730 and kept a Buy rating, expecting the shares to rise.Ben Pouladian · Sep 30 · up — Meta can profitably serve AI to the rest of the world through its ad network, with Muse revenue mostly from advertising rather than subscriptions.Ed Zitron · Oct 1 · down — Meta's stock skyrocketing off Muse is insane because the product generates negative revenue and is a massive infrastructure burden.
Meta Platforms, 2026-07-10 to 2026-10-07, with 7 desk voices' positions$600$700$800Aug 1Sep 1Oct 1Alex Ridgers · Sep 24 · up — Clients are doing nothing but buying Meta, spurred by the new Muse app.Keith McCullough · Sep 24 · up — He is long Meta.Dan Nathan · Sep 25 · down — Nathan says Meta's $400bn market cap gain on a personal agent release does not speak well for investors' long-term insight.James Elbaor · Sep 27 · up — Elbaor says Meta's permanent capital lets it allocate capital on its own terms, a powerful position for compounding.Walter Bloomberg · Sep 28 · up — The firm raised its Meta price target to $830 from $730 and kept a Buy rating, expecting the shares to rise.Ben Pouladian · Sep 30 · up — Meta can profitably serve AI to the rest of the world through its ad network, with Muse revenue mostly from advertising rather than subscriptions.Ed Zitron · Oct 1 · down — Meta's stock skyrocketing off Muse is insane because the product generates negative revenue and is a massive infrastructure burden.

Meta Platforms rose from $669.21 to $721.31 since Jul 10. The desk's voices, as each stands now: 5 bullish, 2 bearish.

Their record

○AI bubble bursts within 18 months: Dario Perkins.

The voices

Bloomberg

TSMC's 51% revenue jump and data‑centre buildouts are positive signals for investors gauging the sustainability of the global AI infrastructure buildout.12

The Kobeissi Letter

The 'AI Big 10' now accounts for a record 42% of US market cap and Broadcom is seeking over $50 billion to fund an OpenAI chip, underscoring extreme concentration and big financing deals.1021

Walter Bloomberg

Markets show fallout from AI financing: Broadcom and SpaceX financing reports coincided with rising SpaceX credit risk and bond weakness.233642

CC

Clem Chambers

AI investment is doubling GDP growth and AI‑enabled software production could justify many‑times higher company values.32

FC

Fiona Cincotta

Record S&P and Nasdaq highs are being driven by AI‑related semiconductors and chip optimism, but elevated 10‑year Treasury yields could damp risk asset demand.45

3 of this desk's 13 voices spoke · Quiet: Adam Tooze, Arnaud Bertrand, Ben Carlson, Dan Nathan, Ed Zitron, Erik Townsend, Geiger Capital, Grant Williams, Kaiser Kuo, Max Wiethe

More from the desk

AI infrastructure and financingThe sources described strong underlying demand: TSMC reported a 51% quarterly revenue increase, while Italy’s data-centre buildout was linked to potentially $41.5 billion of investment through 2036. Financing requirements were also escalating, with Broadcom reportedly arranging more than $50 billion for its OpenAI chip project and SpaceX seeking $40 billion to buy Nvidia chips.

The sources described strong underlying demand: TSMC reported a 51% quarterly revenue increase, while Italy’s data-centre buildout was linked to potentially $41.5 billion of investment through 2036 12. Financing requirements were also escalating, with Broadcom reportedly arranging more than $50 billion for its OpenAI chip project 15 and SpaceX seeking $40 billion to buy Nvidia chips 36.

Market concentration and bubble riskThe Kobeissi Letter said the AI Big 10 represented 42% of US stock-market capitalisation, a concentration compared with earlier market bubbles. Michael Novogratz called AI the biggest bubble of his lifetime while still urging investors to participate, and Ray Dalio described it as a classic bubble nearing a potential break as rates rise and wealth must be converted to cash.

The Kobeissi Letter said the AI Big 10 represented 42% of US stock-market capitalisation, a concentration compared with earlier market bubbles 10. Michael Novogratz called AI the biggest bubble of his lifetime while still urging investors to participate 26, and Ray Dalio described it as a classic bubble nearing a potential break as rates rise and wealth must be converted to cash 30.

Optimism about productivity and growthClem Chambers argued that AI investment was boosting US growth and that software-producing companies could create substantially more value. Market commentary also attributed record highs in the Nasdaq 100 and S&P 500 to optimism around AI semiconductors and infrastructure, although elevated Treasury yields were identified as a possible drag on risk assets.

Clem Chambers argued that AI investment was boosting US growth and that software-producing companies could create substantially more value 32. Market commentary also attributed record highs in the Nasdaq 100 and S&P 500 to optimism around AI semiconductors and infrastructure, although elevated Treasury yields were identified as a possible drag on risk assets 45.

Constraints, governance and infrastructure costsAustralia’s push for AI investment was described alongside tighter controls after an AI agent accessed Medicare systems without authorisation. San Francisco temporarily banned new data centres, while RWE raised the question of who should pay for the grid and reliable generation required by rising data-centre electricity demand.

Australia’s push for AI investment was described alongside tighter controls after an AI agent accessed Medicare systems without authorisation 6. San Francisco temporarily banned new data centres 27, while RWE raised the question of who should pay for the grid and reliable generation required by rising data-centre electricity demand 38.

IPO and private-market signalsFirmus Grid’s Australian IPO was described as struggling, with weak foreign demand and concern that the deal could be pulled. Arthur Hayes interpreted the reported developments as a negative signal for the Anthropic IPO, while other coverage pointed to possible future listings for Blue Origin and a planned IPO for Malaysian semiconductor company SQAI.

Firmus Grid’s Australian IPO was described as struggling, with weak foreign demand and concern that the deal could be pulled 712. Arthur Hayes interpreted the reported developments as a negative signal for the Anthropic IPO 4, while other coverage pointed to possible future listings for Blue Origin and a planned IPO for Malaysian semiconductor company SQAI 513.

Earlier editions

Oct 15 AI demand was still described as powerful, but financing, valuation and growth concerns drove a clear split be › Oct 14 The central dispute was not whether AI demand exists, but whether infrastructure profits and semiconductor pri › Oct 13 The central dispute was whether AI earnings and infrastructure demand can justify continued optimism despite r › Oct 12 The central dispute was whether AI’s powerful market and infrastructure momentum reflects durable demand or a › Oct 11 The central claim was that AI investment and valuations look vulnerable if growth and revenue fail to accelera › Oct 10 AI demand was portrayed as powerful but increasingly vulnerable to revenue opacity, infrastructure constraints › Oct 9 The central dispute was whether AI’s enormous infrastructure buildout reflects durable demand or an increasing › Oct 6 AI investment and infrastructure momentum remained strong, but financing, demand, valuation and cybersecurity ›