Nomura argues that recent massive investment announcements by South Korean chipmakers and Meta's data-center monetization strategy do not signal an AI memory oversupply. The brokerage contends the market remains severely short of memory products, noting that new semiconductor manufacturing capacity typically takes more than a decade to come online, making near-term supply constraints likely to persist.
Summaries like this, in your inbox every morning.
Sign up free →What happened
Nomura rejected investor concerns that recent South Korean chipmaker and Meta investment announcements signal oversupply in AI memory. South Korean memory manufacturers announced approximately 4.8 quadrillion won ($3.5 trillion（約560兆円）) in long-term investment programmes, with around 3.7 quadrillion won dedicated to memory production.
Why it matters
The market remains severely short of memory products driven by rapid AI adoption. Large semiconductor manufacturing complexes require years to build—Nomura cites the Yongin Semiconductor Cluster, originally launched nine years ago, as unlikely to begin limited production before late 2027. This timeline means announced investments will not ease supply constraints for several years.
What to watch
Nomura also defended Meta's plan to commercialize excess data-center computing capacity, arguing it reflects sound business strategy (improving returns on invested capital) rather than weakening AI infrastructure demand. The brokerage sees lower computing costs as likely to encourage broader AI adoption rather than reduce hardware demand.
Investor concerns about oversupply in AI memory rest on the assumption that announced capital investments by major chipmakers and Meta signal weakening demand. However, Nomura's analysis highlights a critical mismatch between announcement timing and implementation reality. The semiconductor industry operates on decade-long cycles between investment decision and production capacity, meaning the $3.5 trillion（約560兆円） in announced programmes will not materially affect supply for several years.
Nomura also reframes Meta's strategy of monetizing unused data-center capacity. Rather than a sign that Meta has overbuilt, the brokerage views this as a rational move to improve returns on invested capital while expanding access for AI-focused customers such as OpenAI and Anthropic. The analysts further suggest that lower computing costs could deepen AI adoption rather than suppress demand for hardware infrastructure. This interpretation shifts the narrative from a demand-destruction story to one where infrastructure efficiency gains reinforce long-term AI investment thesis.
No discussion yet for this article
Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started FreeFree · takes 30 seconds · unsubscribe anytime
1 minute a day. The AI essentials.
200+ sources · Email / LINE / Slack
