Elon Musk answered Sam Altman on Sunday with a post on X that read “The people deserve to have their RAM too!”, according to Benzinga and India Today. It came three days after the OpenAI chief executive said that, “from a global allocation perspective”, it is “quite terrible” for the world’s limited supply of DRAM to sit in personal computers.
Musk’s own company has already cut how much of the same memory its chips use. On 1 and 2 October he wrote on X that Tesla had cut the memory on its AI5 chip to 96GB and on its AI6 chip to 144GB, calling it “the only way to get enough volume for Optimus production”. Both chips ended one-third below their earlier memory, by our arithmetic.
Business Pill · The Memory Wall
A short explainer on memory as a constraint in AI systems. It teaches the general idea only and says nothing about Tesla, Musk or Altman.
The key insight: Sam Altman said on 29 October that keeping the world’s limited DRAM in personal computers is “quite terrible” from an allocation view, and Elon Musk replied on Sunday that people deserve their RAM, as reported; a month earlier Musk said Tesla cut its AI5 and AI6 chip memory to 96GB and 144GB for Optimus volume, a third less each, by our arithmetic.
Altman made the point on Stripe co-founder John Collison’s Cheeky Pint podcast, in an episode published on 29 October. He was answering the idea that open-source models running on home computers would remove the need for data centres.
“A lot of people want local inference,” Altman said, according to YouTube’s auto-captions of the episode. “But from a global allocation perspective, actually, it’s quite terrible to have everybody have a bunch of the world’s limited supply of DRAM sitting in their own computer.”
He said that memory is used “5% of the time” and added: “You want maximum utilization in the cloud if you’re trying to deal with this very supply constrained environment.” The passage runs from 29:11 to 29:41 in the episode.
Memory per chip on Tesla’s AI5 and AI6, before and after the cuts Elon Musk announced on X on 1 and 2 October 2026. The before values are our arithmetic from his ‘cut in half’ (AI5) and ‘1/3’ (AI6); AI5 was first cut to 72GB, then raised to 96GB.
What Musk Posted
Musk’s Sunday post shared a meme based on the opening scene of the film Inglourious Basterds, in which an officer asks, “You are withholding DRAM from the cloud, are you not?”, India Today reported. Benzinga described the post as Musk pushing back against Altman’s view.
On 26 October, six days before that post, Musk replied on X to another user’s post: “Silly analysis. Robot RAM needs will exceed everything else on Earth, as you need the robots to be on physical substates in order to shape atoms.” He added that superintelligence needs will “vastly exceed that”.
The Tesla Cuts
“We cut our RAM in half for the Tesla AI5 chip (now 72GB of LP5) and 1/3 for AI6 (now 144GB of LP6),” Musk wrote on 1 October. He said the cut “greatly reduces cost” and should have “a negligible effect on Optimus performance, as memory bandwidth is a bigger limiting factor than total memory storage”.
The next morning he revised AI5: “Actually, we decided to nudge AI5 up a little to 96GB, as Tesla would otherwise be the only company using the min RAM version of LP5”.
Halving to 72GB implies AI5 started at 144GB, and a one-third cut to 144GB implies AI6 started at 216GB, by our arithmetic. At 96GB and 144GB, each chip now carries 33% less memory than before.
The Price Backdrop
The average price of 8-gigabit DDR4 DRAM used in PCs rose 5.77% to $27.50 in October, its 19th consecutive monthly gain, TheElec reported on Monday, citing DRAMeXchange. TrendForce forecasts PC DRAM contract prices will rise 13% to 18% in the fourth quarter from the previous quarter, according to the same report.
Connecting the Coverage
The same Cheeky Pint episode was the source of our 29 October piece in which Altman said memory makers are each “going to make like $100 billion of free cash flow in a year”: Altman sees $100B memory cash flow. What is new here is his case for keeping DRAM in the cloud, and Musk’s reply.
As we read the two positions, they describe the same scarce chips from opposite ends: Altman argues for pooling DRAM in data centres where it is used more of the time, while Musk’s October posts describe Tesla cutting memory per chip so that more chips can be built for Optimus.
As we read Musk’s 26 October post together with the Tesla cuts, he expects robot memory demand to grow while, for now, Tesla is fitting less memory into each AI5 and AI6 chip, which he said greatly reduces cost.
“This was the only way to get enough volume for Optimus production and greatly reduces cost.”
Three Implications
FOR MEMORY MAKERS TheElec reported, citing DRAMeXchange, that the average price of 8-gigabit DDR4 DRAM rose 5.77% to $27.50 in October, its 19th consecutive monthly gain.
FOR TESLA Musk wrote that the cuts should have a negligible effect on Optimus performance because memory bandwidth, which was held constant, is a bigger limiting factor than total memory storage.
FOR PC BUYERS TrendForce forecasts PC DRAM contract prices will rise 13% to 18% in the fourth quarter from the previous quarter, according to TheElec.
The analysis argues: “Memory is becoming the new compute.”
As we read it, Altman’s case for keeping DRAM in the cloud and Tesla’s cuts to per-chip memory are two responses to that scarcity. The framework is the owner’s argument and does not discuss Altman’s or Musk’s October remarks.
What Is Not Established
We could not open Musk’s Sunday post directly; its wording and the meme are as reported by Benzinga and India Today. Altman’s words come from YouTube’s auto-captions. Musk did not say how much memory Tesla saves in total, and Tesla has not published Optimus production volumes for the AI5 chip.
This Week In The AI Supercycle
Anthropic’s Reality Check, The Capital Squeeze, & The New World Order
This week’s issue: for two years the AI build-out was priced as a growthstory. This week it started being priced as a credit story.
Altman argues that scarce DRAM is better used in data centres than in home computers. Musk replied that people deserve their RAM, a month after he said Tesla cut the memory on its AI5 and AI6 chips to get enough volume for Optimus, which leaves each chip with a third less memory, by our arithmetic.
A note on sourcing. Musk’s October words come from his own posts on X. Altman’s words come from YouTube’s auto-captions of the Cheeky Pint episode. The wording of Musk’s Sunday post is as reported by Benzinga and India Today. DRAM prices are DRAMeXchange’s, as reported by TheElec. The 144GB, 216GB and 33% figures are our arithmetic.
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.
