Could AI Demand Reshape The World’s Memory Market?

🔍 Read the full analysis: Could AI Demand Reshape The World’s Memory Market? on ThorstenMeyerAI.com

Age 18–24?Offer from Amazon

Prime made for students and young adults

• Fast, free delivery for dorm and study essentials
• Prime Video and Amazon Music included
• Member-only deals

Try Prime for Young Adults →Free trial for eligible 18–24 year olds

As an affiliate, we earn on qualifying purchases.Hide

OpenAI CEO Sam Altman met about 50 senior Samsung executives in Yongin, South Korea, on October 31, as demand for memory chips used in AI data centers strains supply. Samsung confirmed the meeting, while Bloomberg reported it was intended to deepen cooperation on memory and AI infrastructure; no new supply contract or chip volumes were announced.

Samsung said executives from its electronics, finance and biotech affiliates met Altman at the company’s Human Resources Development Center. It described the session as part of a leadership program on AI, with discussion of the technology’s direction, its effects on management decisions and potential new markets. Samsung did not publish an attendee list or announce a commercial agreement.

Bloomberg characterized the visit as an effort to deepen ties on memory chips and AI data center infrastructure amid tightening global supply. The Seoul Economic Daily reported that Richard Ho, OpenAI’s vice president responsible for its in-house chip work, accompanied Altman. That outlet also reported that an OpenAI accelerator could use Samsung’s next-generation HBM4 memory, citing industry sources. Neither company has confirmed that report.

Altman’s visit came days after remarks on the podcast Cheeky Pint, published October 29. Asked about high memory prices and the pace of new supply, he said the market’s “invisible hand is failing right now.” He argued that shared data centers could use memory more intensively than personal computers, which he estimated are using memory “5% of the time.” That figure was Altman’s estimate, not a published measurement.

AI Data Centers Compete for Memory

The meeting comes as memory has become a major constraint for companies building and operating AI systems. Data centers need large quantities of memory to serve AI workloads, while the same manufacturers also supply phones, computers and other devices. When capacity is tight, competition between those uses can affect chip availability, prices and product specifications.

Altman’s remarks frame that competition as an allocation question: whether memory should sit in individual devices, where it may not be used continuously, or be concentrated in shared infrastructure. That is an argument about how to deploy limited resources, not evidence that personal computers will lose memory or that OpenAI has secured priority access. Altman also leads a company that sells cloud AI services, giving his argument a direct business context.

The pressure may reach consumers without empty store shelves. The source reports that TrendForce expects conventional DRAM contract prices to rise 10% to 15% in the fourth quarter and warns PC memory supply could decline in 2027 as more capacity shifts to servers. Those are forecasts, not confirmed outcomes. If costs rise, manufacturers may pass them on through higher prices or offer less memory and storage in some mainstream devices.

Popular right now

KitchenAid Artisan 5-Quart Tilt-Head Stand Mixer, Almond Cream

AmazonView Latest Price

Jura E6 Automatic Espresso Machine, Piano Black (15622)

AmazonView Latest Price

Ninja Foodi Smart Dual Heat Air Fry Oven, 13-in-1, Silver (SP351)

AmazonView Latest Price

As an affiliate, we earn on qualifying purchases.

Existing Ties and Rising Chip Demand

The meeting took place against an existing relationship. In October 2025, OpenAI and four Samsung companies signed a letter of intent. Samsung said then that OpenAI’s projected demand could reach up to 900,000 DRAM wafers per month. That was a projection in the letter of intent, not a confirmed delivery schedule or a new commitment announced at the October 2026 meeting.

OpenAI said in June 2026 that Samsung Electronics would introduce ChatGPT Enterprise and Codex to employees in Korea and its global device division. The Korea Herald reported that Samsung executive chairman Lee Jae-yong visited OpenAI in San Francisco in July. The source also says SK Group chairman Chey Tae-won met Altman in Seoul on October 31 alongside SK hynix executives; details of that discussion were not disclosed.

Samsung is one of the world’s major memory producers, and its chip business has become central to its financial results. The source reports that in the third quarter Samsung’s chip division recorded 106.5 trillion won of 107.4 trillion won in company operating profit. It also reports a 0.6 trillion won operating loss for the phone and network business, partly linked to cost pressures. The figures highlight the commercial stakes, but do not establish that the meeting produced a supply agreement.

““From a global allocation perspective,” it is “quite terrible” that so much memory sits in people’s computers, “being used 5% of the time.””

— Sam Altman, on the Cheeky Pint podcast

No Chip Deal or Volumes Disclosed

The public record does not establish that OpenAI secured additional memory supply during the meeting. No contract, prices, delivery dates or volumes were announced, and Samsung’s account focused on AI leadership training rather than a chip negotiation. Bloomberg’s description of the visit as an effort to deepen cooperation is a report about its purpose, not confirmation of a deal.

It is also unclear whether OpenAI’s planned accelerator will use Samsung HBM4. The claim came from industry sources cited by the Seoul Economic Daily and has not been confirmed by either company. The status of any discussion with SK Group or SK hynix is similarly unknown. More broadly, forecasts for memory prices and PC supply may change as manufacturers adjust production and demand develops.

Watch for Supply and Pricing Updates

The next concrete signals will be any company announcements about memory supply, accelerator components or manufacturing plans. Until then, the October 2025 letter of intent is the clearest disclosed indication of the scale OpenAI may seek, but it should not be treated as a purchase order. No follow-up agreement or announcement has been reported in the source material as of November 1.

For buyers and device makers, the near-term indicators are contract-price updates and whether manufacturers shift more production toward server memory. TrendForce’s fourth-quarter price outlook and warning about possible 2027 PC supply declines are forecasts to track, not settled results. The available information does not show that consumer devices are facing shortages, but it points to a risk of higher prices or reduced memory configurations if supply remains constrained.

Key Questions

Did OpenAI and Samsung announce a new memory supply deal?

No. Samsung confirmed the October 31 meeting, but neither company announced a new contract, supply volume or price.

What did Altman mean by memory in personal computers being used 5% of the time?

He argued that memory in individual computers may sit idle while memory in shared data centers can serve more workloads. The 5% figure was Altman’s estimate; the source provides no measurement method or independent verification.

Could the reported OpenAI accelerator use Samsung HBM4?

The Seoul Economic Daily reported that possibility, citing industry sources. OpenAI and Samsung have not confirmed it, so the chip design and supplier remain unverified.

Could AI demand make computers more expensive?

It could add pressure if memory supply remains tight, but the source does not establish a direct price increase for specific consumer products. TrendForce forecasts conventional DRAM contract prices will rise 10% to 15% in the fourth quarter; that is a market forecast, not a confirmed retail-price change.

Source: ThorstenMeyerAI.com