OpenAI CEO Sam Altman has met with about 50 senior Samsung executives as the companies continue discussing artificial intelligence, enterprise services and memory-chip cooperation. The meeting took place at Samsung’s Human Resources Development Center in Yongin, South Korea, according to Samsung and a report from Data Center Dynamics citing Bloomberg.
The discussions come as memory companies warn that shortages could continue into 2027 and beyond. The supplied reports connect the pressure partly to rising demand from AI infrastructure, while also noting that no new memory allocation or delivery terms were disclosed during the meeting.
Samsung meeting focused on AI’s business impact
Samsung said the meeting examined how AI could change corporate management and create new business opportunities. The event brought together presidents and senior executives from major Samsung affiliates, including businesses involved in electronics, finance and biotechnology.
The theme was “The Next 10 Years Shaped by AI: The Enterprise’s Choice.” Executives discussed possible applications for AI, the future direction of the technology and the leadership skills needed as companies adapt to changing business conditions.
The meeting was also part of Samsung’s wider effort to expand AI use across its businesses. Before speaking with Altman, executives completed practical training on generative AI, including ways to organize business data into personalized knowledge bases and use AI-generated insights in management decisions.
OpenAI’s existing memory-chip agreement
The meeting also sits within a broader relationship between OpenAI and Samsung. In October 2025, Samsung Electronics signed a letter of intent with OpenAI covering cooperation on global AI infrastructure, including the possible supply of advanced memory chips for OpenAI’s Stargate project.
Samsung and SK Hynix previously said that OpenAI’s projected Stargate demand could reach 900,000 DRAM wafers per month. DRAM, or dynamic random-access memory, is a type of semiconductor memory used to store data that computers and servers are actively using. The companies did not provide a timeline for meeting that projected demand.
The sources do not disclose how much memory Samsung has agreed to provide, when shipments would begin or whether the latest meeting produced a new allocation. The reported relationship therefore establishes ongoing cooperation, but not a confirmed expansion of supply terms.
Why memory supply is under scrutiny
Memory manufacturers have warned that shortages may persist into 2027 and later. The reports say consumer electronics could face pressure as manufacturers assign more capacity to AI infrastructure projects.
A January 2025 Wall Street Journal report cited TrendForce data estimating that data centers could consume up to 70 percent of worldwide memory production in 2026. That figure describes a projected share of production rather than a confirmed result, and the supplied material does not identify how it would affect any individual company’s shipments.
Altman has also commented publicly on the supply challenge. On the Cheeky Pint podcast, he argued that widespread local AI inference could leave limited DRAM sitting in desktop computers and used only part of the time. He said cloud use could provide higher utilization in a supply-constrained environment. These remarks were his stated view, not a disclosed term of the Samsung meeting.
The partnership extends beyond chips
Samsung’s relationship with OpenAI includes projects outside memory supply. Samsung SDS agreed to explore AI data-center development and enterprise AI services, while Samsung C&T and Samsung Heavy Industries pursued joint development of floating offshore data centers.
Samsung Electronics also said it would roll out ChatGPT Enterprise and Codex to employees in South Korea and across its global Device eXperience division. ChatGPT Enterprise is OpenAI’s workplace service, while Codex is a tool associated with software development tasks.
These arrangements show that the relationship covers infrastructure, workplace software and business services as well as semiconductors. The sources do not establish that all of these projects share a single commercial timetable or that they depend on a newly negotiated memory allocation.
Samsung is expanding internal AI adoption
Samsung said it began applying AI across research and development, manufacturing, marketing and administrative work in June. Each affiliate has been developing its own adoption plan, supported by teams responsible for implementation, data management and operating AI models.
The company said it had trained about 2,300 executives since June and aims to complete AI training for all employees by February 2027. It also plans to increase the use of external generative AI services in the workplace.
That internal rollout helps explain why Samsung framed the meeting around enterprise transformation, rather than only semiconductor supply. The Korea Herald described the event as part of Samsung’s broader push to accelerate AI across its operations.
What remains undisclosed
The reported meeting confirms high-level engagement between OpenAI and Samsung, but the available information leaves important supply questions unanswered. Neither company has announced a specific increase in memory allocation resulting from the discussion.
There is also no disclosed schedule for satisfying the projected Stargate requirement of 900,000 DRAM wafers per month. The sources do not state whether the figure represents a final order, a long-term target or a broader projection for future demand.
For now, the clearest verified development is that OpenAI and Samsung continue to discuss AI strategy while operating within an existing infrastructure and memory-supply relationship. The meeting reinforces the partnership’s importance, but its effect on future chip availability cannot yet be measured from the disclosed information.
