Bar chart: AI suppliers booked into 2030 — Micron 2030, Lumentum 2029, AWS 2028, ASML 2028; Broadcom says capital might be the biggest constraint

AWS 2027 Is Largely Sold Out; Broadcom’s CEO Flags Capital

Amazon chief executive Andy Jassy said of AWS computing capacity, Reuters reported on 23 October 2026: “2027 is largely sold out.” Broadcom chief executive Hock Tan named a different limit: capital will be the biggest constraint facing the company as AI development continues, Reuters reported from a panel in Singapore on 12 October. Both appear in a list of 15 business leaders on AI capacity that Shanu Mathew posted on X on 25 October 2026.

We checked the quotes in that list against the companies’ own statements, the outlets that reported them, and our own sourced pieces. Nine could be traced to a primary or a named outlet; six could not in the time we had, and they are left out below.

Business Pill · LEAD TIME

A short Business Engineer explainer of lead time: why the wait for a product is part of its price. It teaches the general idea only and says nothing about any company or statement in this piece.

The key insight: As we read it, the new element is where the limit sits: of the nine statements we could trace, seven describe capacity or supply as the limit, from Jassy’s “2027 is largely sold out” to Lumentum’s “almost sold out for 2029,” while Broadcom’s Hock Tan put capital as the biggest constraint, Reuters reported on 12 October.

Nine Leaders, Checked

The statements below were made between 30 September and 23 October 2026. Each uses its own measure: a year booked, a supply shortfall, demand ahead of supply, or, for Lam, higher demand driven by AI.

WhoWhat they saidWhere and when
Andy Jassy, Amazon“2027 is largely sold out,” of AWS computing capacity; a good chunk of AWS’s 2028 is sold out, as our sellout piece reportedReuters, 23 October
Sundar Pichai, AlphabetGoogle is “capacity constrained as we are serving it”Big Technology Podcast, published 20 October
Jensen Huang, Nvidia“Capacity is the constraint”GTC Berlin, as reported by SDxCentral, 21 October
Lisa Su, AMDExpects chip demand to stay “very high” for the next several yearsTaipei, 6 October, per Insider Monkey
Sanjay Mehrotra, Micron“We do not have line of sight to when supply and demand will return to balance”Fiscal fourth-quarter call, per Computerworld
ASMLSupply as currently planned for 2027 and 2028 “will not fully meet the demand of our customers”Third-quarter report, per Evertiq, 15 October
Michael Hurlston, LumentumThe optics maker is “almost sold out for 2029”Bloomberg Television, 9 October
Tim Archer, Lam Research“With AI driving higher semiconductor demand and greater device complexity”Results release, 21 October
Hock Tan, BroadcomCapital will be the biggest constraint facing the company, Reuters reportedPanel in Singapore, 12 October

Sources as listed. ASML’s line is from the company’s quarterly report; the Hock Tan line is Reuters’ paraphrase, not a quotation.

The furthest year four AI suppliers named in October 2026 as booked or short of demand: Micron's agreements co
The furthest year four AI suppliers named in October 2026 as booked or short of demand: Micron’s agreements cover DRAM and NAND through 2030, Lumentum’s chief executive said it is almost sold out for 2029, and ASML and Andy Jassy (for AWS) pointed to 2028. Measures differ by company; bar length is years beyond 2026, our arithmetic.

What Was Posted on X

Shanu Mathew (@ShanuMathew93) lists 15 business leaders on AI infrastructure demand and capacity from the past month or so, from Andy Jassy’s “2027 is largely sold out” to Hock Tan saying capital “might be the biggest constraint.”

Where the List and the Record Differ

Two wordings in Mathew’s post differ from what we found. The post quotes Pichai as saying “We are capacity constrained on a daily basis”; the transcript passage our Pichai piece quotes says Google is “capacity constrained as we are serving it.” The post has Lumentum’s Hurlston saying “completely sold out”; the Bloomberg Television interview our sellout piece cites has “almost sold out for 2029.”

Six quotes we could not trace to a primary or a named outlet, so they are not in the table: those attributed to Matt Garman of AWS, C.C. Wei of TSMC, Mark Zuckerberg of Meta, Josh Payne of Nscale, Satya Nadella of Microsoft and Lip-Bu Tan of Intel. That is a statement about our search, not about whether they said them.

The Bill on the Other Side

The same weeks brought numbers from the other side of the ledger. Apollo chief economist Torsten Slok wrote in The Daily Spark of 24 October that artificial intelligence “has delivered zero financial payoff to corporate America outside of tech,” as our Slok piece reported.

Meta’s information statement, filed with the SEC on 21 October, says an adviser put the most Meta could raise in equity without Mark Zuckerberg’s voting power falling below 50% at about $700 billion over a number of years, as our Meta filing piece reported. That figure is the adviser’s estimate, and Meta has not filed or announced an offering. And Fireside Alpha relayed Citrini warning that if combined AI lab ARR lands at $300 billion instead of the $500 billion being underwritten, “that’s a disaster,” in a post we embedded in our Gartner piece.

The Structural Read

As we read it, the suppliers are naming years, not quarters: Micron’s agreements cover DRAM and NAND through 2030, Lumentum points to 2029, and ASML says its planned supply for 2027 and 2028 falls short of demand.

As we read it, the buyers say the same from the other end: Pichai said Google is “capacity constrained as we are serving it,” and Huang said “Capacity is the constraint.”

As we read it, the bill is the open question: Meta’s filing puts its equity ceiling at about $700 billion, an adviser’s estimate, and Slok wrote that AI “has delivered zero financial payoff to corporate America outside of tech.”

Andy Jassy, Amazon, on AWS computing capacity, as reported by Reuters, 23 October 2026

“2027 is largely sold out.”

Three Implications

FOR AI BUYERS ASML says its planned supply for 2027 and 2028 “will not fully meet the demand of our customers,” and Micron says it has no line of sight to when supply and demand will return to balance.

FOR SUPPLIERS Lisa Su expects chip demand to stay “very high” for the next several years, Insider Monkey reported from Taipei.

FOR INVESTORS Fireside Alpha relayed Citrini warning that if combined AI lab ARR lands at $300 billion instead of the $500 billion being underwritten, “that’s a disaster.”

The Business Engineer Lens

This story maps onto the Business Engineer framework The AI Capex Map & The State of AI Hyperscalers.

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”.

As we read it, the October statements fit both halves of that line: suppliers say capacity is booked for years, and Broadcom’s Hock Tan named capital as the biggest constraint, Reuters reported. The framework was published in May 2026 and does not discuss these October statements.

What Is Not Established

The nine statements use different measures and different dates, so they cannot be added up or ranked against each other. Sold-out capacity describes orders and plans, not profits for the buyers. Hock Tan’s line is Reuters’ summary of a panel remark. We did not see the full transcripts of every call, and we did not contact the people or companies named, or Shanu Mathew.

This Week In The AI Supercycle

Execution Fault Lines, Concentrated Demand, and Geopolitical Firewalls

This week’s issue: the AI build is not short of demand for its parts. It is short of cheap money, and its end revenue rests on a narrow base of large, often venture-funded buyers.

Read the full issue →

Business Engineer Framework

The AI Capex Map & The State of AI Hyperscalers

A Business Engineer map of who is spending on AI infrastructure, how it is financed, and where it can crack.

Read the Map of AI →

The Bottom Line

Of the 15 statements in Shanu Mathew’s list, nine trace to a primary or a named outlet. Amazon, Alphabet, Nvidia, AMD, Micron, ASML and Lumentum describe capacity or supply as the limit, from Jassy’s “2027 is largely sold out” to Lumentum’s “almost sold out for 2029.” Broadcom’s Hock Tan put capital as the biggest constraint, Reuters reported on 12 October, the same month Slok wrote of “zero financial payoff” outside tech.

95,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.

A note on sourcing. Shanu Mathew’s X post was read on 25 October 2026. The Hock Tan line is Reuters’ report as published by The Star; ASML’s statement is from Evertiq’s report of the company’s quarterly results; Micron’s is from Computerworld; Lisa Su’s from Insider Monkey. The Jassy, Pichai, Huang, Hurlston, Archer, Slok, Citrini and Meta figures come from our earlier pieces, which name their sources. We did not contact the people or companies named.

Sources: Shanu Mathew on X, 25 Oct 2026 · Reuters via The Star, Hock Tan, 12 Oct 2026 · Evertiq on ASML’s Q3 report, 15 Oct 2026 · Computerworld on Micron’s FQ4 call · Insider Monkey on Lisa Su, 11 Oct 2026

Scroll to Top
FourWeekMBA