📊 Full opportunity report: Seoul Spotlights Memory As The Key Limiting Factor In AI Growth on ThorstenMeyerAI.com — validation score, market gap, and execution plan.
TL;DR
Seoul officials, led by SK Group chairman Chey Tae-won, warn that AI memory demand will grow 50-60% by 2027, but supply capacity remains insufficient, posing risks to industry and geopolitics.
Seoul’s top industry leader, Chey Tae-won of SK Group, has publicly warned that the global demand for AI memory will increase by at least 50-60% in 2027, while supply remains stagnant, creating a significant bottleneck that could impact AI development and geopolitical stability. This warning underscores the critical role of memory capacity in the AI industry and signals potential risks for supply chain security and international relations.
During a press briefing at the Korea Chamber of Commerce and Industry’s Jeju Forum, Chey Tae-won stated that customers are requesting 60 to 100% more AI memory in 2027 than current levels. He emphasized that AI now accounts for over half of total semiconductor consumption, with demand growth expected to be at least 50-60%. Despite this, he noted that no significant new capacity is expected to come online in 2026, exacerbating the supply-demand imbalance.
Chey highlighted that the shortage is most acute in high-bandwidth memory (HBM), which is critical for AI accelerators, and warned of a potential escalation in geopolitical tensions surrounding memory access. He described current lobbying efforts and government interventions as chaotic and predicted that, soon, governments might pressure each other over access to memory resources, which could threaten global AI development and industry stability.
Counterpoint Research reports that SK hynix held 58% of the global HBM revenue in Q1 2026, with Samsung and Micron each holding about 21%, indicating a highly concentrated market. SK hynix has announced plans to accelerate capacity expansion, including a new HBM-focused facility scheduled for February 2027 and additional investments, but these will not address the immediate capacity shortfall in 2026.
Models get the headlines.
Memory is the chokepoint.
SK Group’s chairman at the Jeju Forum, per The Korea Herald: customers want 60–100% more AI memory in 2027, governments now treat memory access as economic security — and no company has meaningful new capacity arriving next year.
The gap, in his own numbers
customer requests to SK hynix vs this year. AI already consumes over half of all semiconductors; total demand growth floored at 50–60%.
“No company has meaningful new capacity coming online next year.” The gap year is already locked in — fabs don’t move faster than physics.
Result, per Chey: near-chaotic lobbying — no longer just from companies. Foreign governments are intervening for domestic industries; next, governments pressure governments.
Tighter than the chokepoints you worry about
SK hynix’s race against its own warning
Company figures and projections as announced — none of it lands in 2026.
Half true: unified-memory Apple Silicon doesn’t queue for HBM — a fleet you own is insulated from allocation politics, and owned hardware converts supply-chain risk into sunk cost.
The other half: LPDDR and HBM share DRAM wafer economics — chipflation reaches workstation memory too, and training compute stays fully hostage. Local inference changes who feels the shortage, not whether it exists.
Week tie-in: if memory demand grows into capacity that doesn’t exist, doing the job in 3B parameters on memory you already own isn’t aesthetics — it’s engineering under constraint.
high bandwidth memory (HBM) modules
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Implications of Memory Shortage for AI and Geopolitics
This development is significant because it highlights a critical bottleneck in AI advancement: memory capacity. The shortage could slow down AI research, deployment, and innovation, especially for large-scale models that require substantial high-bandwidth memory. Additionally, the concentration of memory supply among a few companies raises concerns about market dominance and geopolitical risks, as access to memory becomes a matter of economic security. The warning from SK hynix’s chairman signals that industry and government stakeholders may soon face increased pressure over resource allocation, potentially leading to new geopolitical conflicts and supply chain disruptions.
AI memory storage solutions
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Memory Market Concentration and Industry Dynamics
Memory, especially high-bandwidth memory (HBM), is concentrated among three major players: SK hynix (58% of global HBM revenue in Q1 2026), Samsung, and Micron. The industry has experienced sustained demand growth, with SK hynix projecting a 33% compound annual growth rate in HBM demand through 2030. However, new capacity investments are not expected to match this growth in 2026, creating a looming supply crunch.
Chey Tae-won’s remarks come amid broader concerns about the geopolitical implications of semiconductor supply chains, which have traditionally focused on TSMC and other foundries. The current situation underscores the fragility of the memory supply chain, which is less discussed but equally critical, especially as demand for AI accelerators surges. The market’s tight concentration and lack of immediate capacity expansion options heighten the risk of supply disruptions and geopolitical tensions.
Furthermore, the high prices driven by demand have led to what Chey described as “chipflation,” which could slow AI adoption and increase costs for device manufacturers and consumers. SK hynix’s recent investments aim to mitigate this, but the capacity gap remains significant through 2026.
“No company has meaningful new capacity coming online next year.”
— Chey Tae-won, SK Group Chairman
enterprise GPU memory upgrade
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Unconfirmed Details and Future Capacity Outlook
It remains unclear whether SK hynix and other suppliers can accelerate capacity expansion sufficiently to meet the projected demand growth by 2027. The timeline for new facilities coming online suggests a capacity shortfall will persist through 2026, but the exact impact on global supply and prices remains uncertain. Additionally, the potential for increased government intervention or geopolitical conflicts over memory access is still developing and not yet fully defined.
high-performance RAM for AI servers
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Next Steps in Capacity Expansion and Policy Responses
Industry leaders and governments are expected to prioritize capacity expansion in the coming months, with SK hynix’s new facilities scheduled for 2027. Monitoring of geopolitical developments and potential resource allocation conflicts will be crucial. Additionally, the industry may explore alternative memory architectures or supply chain diversification to mitigate risks. Further official statements and investment announcements are anticipated to clarify the capacity outlook and policy responses.
Key Questions
Why is memory capacity so critical for AI development?
Memory capacity, especially high-bandwidth memory (HBM), is essential for training and deploying large AI models. Insufficient memory can slow down AI processing, increase costs, and limit the scale of models that can be handled efficiently.
What are the geopolitical implications of the memory shortage?
The concentration of memory supply among a few companies makes access a matter of economic security. Countries may intervene or restrict exports, potentially leading to geopolitical tensions and affecting global AI competitiveness.
When will new memory capacity come online?
SK hynix plans to complete its new HBM-focused facility by February 2027, but existing capacity shortfalls are expected to persist through 2026. The timeline for other suppliers remains uncertain.
How might this shortage impact AI innovation and deployment?
The shortage could slow down the development and deployment of large AI models, increase costs, and push innovation toward smaller, more efficient models or alternative architectures.
Are there any solutions to address the memory bottleneck?
Potential solutions include increasing capacity investments, diversifying supply sources, developing alternative memory technologies, and optimizing AI models for efficiency to reduce memory demands.
Source: ThorstenMeyerAI.com
