Investor Gavin Baker says a chip buyer that walks away from a memory long-term agreement (LTA) risks its business: “you might blow up your entire business and your franchise by breaking an LTA. And that was never the case before”, Baker says in a clip posted on X by Fireside Alpha on November 8. Some of the same remarks were posted on X on August 4, the day Invest Like the Best released the episode “Gavin Baker – AI Market Jitters”, and Baker wrote that day that “Game theory suggests memory LTAs are likely to be durable”.
Jukan, an account on X quoted in the Fireside Alpha post, points to the chipmaker onsemi as a precedent. onsemi’s long-term supply agreements (LTSAs) held about $16.6 billion of remaining performance obligations at the end of 2022 and about $7.1 billion at the end of 2025, according to its 10-K filings, a 57% drop by our arithmetic, while annual revenue fell 28%, from $8,326.2 million to $5,995.4 million.
Business Pill · Lead Time
A short Business Engineer explainer on lead time, the gap between an order and its delivery. It teaches the general idea only and does not discuss memory LTAs or onsemi.
The key insight: Investor Gavin Baker says breaking a memory long-term agreement could cost a chip buyer its future supply allocations, and wrote on August 4 that game theory suggests memory LTAs are likely to be durable. onsemi’s long-term supply agreements, which its CEO called legally binding, held $16.6 billion of remaining performance obligations at the end of 2022 and $7.1 billion at the end of 2025, its 10-Ks show, a 57% drop by our arithmetic; its 2025 10-K says certain LTSAs were modified that year in response to changes in demand.
Baker frames it as game theory among four buyers: “There’s Amazon with their Trainiums, there’s Google with their TPUs, there’s AMD, and then there’s NVIDIA who’s much bigger than everybody else combined”. Baker says memory is “the single most important thing you could do to increase kind of token output per unit of compute”.
Baker says that “market shares are, I think for the next several years, are going to be determined by supply chain allocations and what you have pre-purchased”. Baker continues: “If you break your LTA, and then in the next two or three years, for any reason, leverage shifts back to the memory guys, you’re out of business. It’s over”.
onsemi LTSA remaining performance obligations were about $16.6 billion at end-2022, $16.5 billion at end-2023, $11.9 billion at end-2024 and $7.1 billion at end-2025; fiscal-year revenue was $8,326.2 million, $8,253.0 million, $7,082.3 million and $5,995.4 million (onsemi 10-K filings). The 57% and 28% declines are our arithmetic. Obligations also fall as products ship; the 10-Ks do not split the decline between shipments and modifications.
What Was Posted on X
Gavin Baker on memory LTAs: “You might blow up your entire business and your franchise by breaking an LTA. And that was never the case before”
Talking about memory makers moving buyers onto long-term agreements (LTAs), he works through the game theory of what happens to a buyer who walks away from one:
“Let’s just think about the game theory of breaking an LTA. So there’s four companies that matter at scale. There’s Amazon with their Trainiums, there’s Google with their TPUs, there’s AMD, and then there’s NVIDIA who’s much bigger than everybody else combined.”
“Let’s just say it’s 2027, it’s very important to realize memory is, the more memory you put with FLOP for a given unit of compute, the more tokens you get out. It’s the single most important thing you could do to increase kind of token output per unit of compute. And then that obviously definitionally actually lowers costs, which is why the demand hasn’t responded at all negatively. There’s been no elasticity just because it’s kind of the only, it’s the axis that is dominating all others. And this is at some level, a giant Game of Thrones or emperors between these companies.”
“And okay, it’s 2027 or ’28, you’re vaguely tempted to break one of these LTAs, try and get a lower price. But to a large degree, market shares are, I think for the next several years, are going to be determined by supply chain allocations and what you have pre-purchased. So if you break the LTA and you, and this is assuming we’re not in a severe oversupply situation, the game theory even holds in a severe oversupply situation. If you break your LTA, and then in the next two or three years, for any reason, leverage shifts back to the memory guys, you’re out of business. It’s over.”
“Let’s just say Google breaks an LTA. There’s an oversupply. I’m making this up, in ’28, ’29. They break their LTAs. Well, if they’re breaking their LTAs, it probably means, your oversupply, prices are coming down, and then, capacity naturally contracts. Well, what do you think is going to happen to Google’s allocations? And then, this is a cyclical industry, and oversupply is followed by undersupply. What do you think they think is going to happen to their allocations next time?”
“So I just think given that this is the axis around which kind of everything is revolving, man, you might blow up your entire business and your franchise by breaking an LTA. And that was never the case before.”
Fireside Alpha posted this clip of investor Gavin Baker on 8 November. In it, Baker says a buyer that breaks a memory long-term agreement risks its allocations if leverage later shifts back to memory makers. Matching quotes were posted on X on 4 August, the release date of the Invest Like the Best episode with Baker.
I’d like to believe LTAs are durable too. But take one example.
Analog chipmaker onsemi said this in 2023. On January 5, 2023, at JPMorgan’s CES forum, CEO Hassane El-Khoury was asked how much LTSAs would still hold once the shortage eased. His answer: “They are a legally binding agreement ’cause they’re a liability for the customer because they’re take or pay.” He also said he would not accept a customer reopening price talks just because it could get the same parts 10% cheaper elsewhere.
On October 30, 2023, on the Q3 earnings call, onsemi cut its SiC shipment forecast for the year from $1 billion to more than $800 million after one automotive OEM reduced demand. An analyst asked how this could happen when the company had been saying its LTSAs were legally binding. The same CEO reaffirmed that they were binding, but said he did not want to enforce the contracts just to push inventory onto a customer whose demand had dropped. If it worked for both sides, terms could be adjusted customer by customer, whether by a quarter, a year, or longer.
In its 10-K filings, onsemi also states that it amended LTSAs with some strategic customers by mutual agreement in 2024, and changed delivery schedules and committed volumes on some LTSAs in 2025 as demand shifted.
Microchip is another case. Customers in its Preferred Supply Program (PSP) originally had to place 12 months of orders and could not cancel or reschedule them unless Microchip raised prices. On its November 2025 earnings call, however, the CEO said the company had accepted some cancellations and pushouts so customers could adjust their backlog. On the same call, the CFO separately addressed the multiyear LTSAs the company had signed. Some of those contracts were still in effect, he said, but customers had been allowed to defer taking delivery of committed volumes by one to two years if they needed to.
Every time I see cases like these, I doubt how much memory makers’ LTAs will protect earnings in a downcycle. Right now the most reliable thing is the prepayments and deposits themselves. That cash can go into capacity expansion today.
I can picture hyperscalers with no DRAM inventory continuing to buy at current prices. But once inventory has piled up, will they keep taking every committed unit at the same price? I have a hard time picturing memory makers holding customers to the high contract price all the way through.
The LTA itself may never be broken and may stay in place. But if both sides agree to push this year’s volumes into next year or two years out, this year’s earnings estimates come down. If price gets renegotiated too, margin estimates have to come down as well.
Again, I agree LTAs are sturdier than they used to be. What I question is whether, after demand turns, memory makers can still earn 80% margins on contracts signed during the upturn and keep collecting hundreds of billions of dollars in profit because of those LTAs.
The account Jukan (@jukan05) posted this on 7 November, citing onsemi’s long-term supply agreements and Microchip’s supply programme as cases where binding contracts were adjusted when demand fell. The post says the prepayments and deposits are the most reliable part today; we checked onsemi’s 10-K language, not the call remarks the post describes.
The onsemi Precedent
onsemi introduced LTSAs in May 2021, according to a July 23, 2026 order by the U.S. District Court for the District of Arizona in an investor class action, Lew v. ON Semiconductor. The order says that in August 2022 onsemi CEO Hassane El-Khoury affirmed that LTSAs are “legally binding” and that customers are “liable” for agreed prices and quantities.
At a September 7, 2023 Citi conference, the order quotes El-Khoury saying an LTSA is a legally binding agreement that both parties in a win-win “will renegotiate an amendment”, adding: “we’re not going to be left holding the bag”. The court granted onsemi’s motion to dismiss, calling such statements inactionable puffery, and noted that the plaintiff did not explain how an LTSA loses its legal enforceability because a customer “decommits”.
The filings changed their wording over time. The 10-K for 2022 describes LTSAs as multi-year agreements that include “firmly committed amounts”. The 10-K for 2024 says the timing, pricing or amounts delivered under LTSAs “may be modified or canceled in circumstances upon mutual agreement”. The 10-K for 2025 says “During 2025, certain LTSAs were modified in response to changes in demand”, with modifications primarily related to delivery schedules and volume commitments.
Some LTSAs carry cash up front. onsemi recorded capacity payments of $162.9 million in 2022 and $206.3 million in 2023 within contract liabilities, its 10-Ks say, describing them as non-cancelable payments that secure production availability or prepay part of a customer’s product obligation.
Timeline: onsemi’s CEO called LTSAs legally binding in August 2022 and said at a September 7, 2023 Citi conference that both parties will renegotiate an amendment (as quoted in the July 23, 2026 order in Lew v. ON Semiconductor); onsemi’s 10-K filed in February 2025 says LTSAs may be modified or canceled upon mutual agreement and its 10-K filed in February 2026 says certain LTSAs were modified in 2025; Gavin Baker wrote on August 4, 2026 that memory LTAs are likely to be durable.
The Counterpoint Quoted With the Clip
The Fireside Alpha post quotes a November 7 post by the account Jukan (@jukan05), which cites onsemi as a precedent. Jukan writes that onsemi cut its silicon carbide shipment forecast for 2023 from $1 billion to more than $800 million on its October 30, 2023 call after one automotive customer reduced demand, and that Microchip’s CEO said on its November 2025 call that the company had accepted some cancellations and pushouts. Jukan writes: “Right now the most reliable thing is the prepayments and deposits themselves”.
Jukan also writes: “I agree LTAs are sturdier than they used to be”, while questioning whether memory makers “can still earn 80% margins on contracts signed during the upturn” after demand turns.
As we read Baker’s argument, the claim rests on allocations: a buyer that breaks an LTA during a glut risks losing supply in the next shortage, which Baker says did not use to be the case.
As we read onsemi’s 10-Ks, the wording moved from firmly committed amounts in the 2022 filing to amounts that may be modified or canceled upon mutual agreement in the 2024 filing, and to certain LTSAs modified in 2025.
As we read the two cases together, they differ on who holds leverage: Baker describes memory makers holding it through allocations, while Jukan’s post argues that prepayments and deposits are the most reliable part of the contracts today.
Gavin Baker, in the clip posted by Fireside Alpha on November 8
“you might blow up your entire business and your franchise by breaking an LTA. And that was never the case before”
Three Implications
FOR MEMORY MAKERS onsemi’s filings show LTSAs that its CEO called legally binding were later modified in response to changes in demand, and its LTSA book fell 57% from 2022 to 2025 by our arithmetic.
FOR CHIP BUYERS Baker says market shares for the next several years will be determined by supply chain allocations and what buyers have pre-purchased.
FOR WHAT TO WATCH onsemi’s capacity payments, $162.9 million in 2022 and $206.3 million in 2023 per its 10-Ks, are the cash part of such contracts, the part Jukan’s post calls the most reliable today.
The essay says “HBM growth is outpacing accelerator demand, making memory the rate-limiting factor in the AI scaling equation.”
As we read it, that is the condition under which Baker expects memory LTAs to hold: when memory is the binding input, a buyer’s allocation is worth more than a lower price. The essay does not discuss LTAs or onsemi.
What Is Not Established
onsemi’s 10-Ks do not split the fall in remaining performance obligations between products shipped and contracts modified, so the 57% figure is not a measure of renegotiation. The memory makers’ own LTA terms are not in the documents we read. We did not verify Jukan’s accounts of the January 2023 CES remarks, the October 2023 call or Microchip’s calls. The episode transcript is behind a login; the Baker quotes are from the posted clips. The court granted the motion to dismiss without ruling that the LTSAs were unenforceable.
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.
Baker argues that breaking a memory LTA could cost a buyer its future allocations, and wrote in August that memory LTAs are likely to be durable. onsemi’s CEO called its LTSAs legally binding. Its 2025 10-K says certain LTSAs were modified that year in response to changes in demand. Its LTSA book fell 57% from 2022 to 2025 by our arithmetic, while revenue fell 28%.
95,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. Gavin Baker’s remarks are quoted from the clip post by Fireside Alpha on X (8 November 2026); matching quotes were posted on X on 4 August 2026, the release date of the Invest Like the Best episode, whose transcript is behind a login and was not read. onsemi’s figures and contract language are from its Form 10-K filings for fiscal 2022 to 2025 on SEC EDGAR, read in full for the LTSA notes.
The quotes of onsemi’s CEO are as quoted in the July 23, 2026 order of the U.S. District Court for the District of Arizona in Lew v. ON Semiconductor, which we read in full. Jukan’s claims about onsemi’s 2023 call and Microchip are from the post and were not independently verified. The 57% and 28% figures are our arithmetic. We did not contact Baker, onsemi or Jukan.
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.
