SK Group and Kioxia: A Semiconductor Alliance That Could Reshape Global Memory Markets

A Deal on the Horizon? SK and Kioxia Are Testing the Waters

SK Group chairman Chey Tae-won did not mince words during a September interview with Japan’s Asahi Shimbun: if SK Hynix fits into Kioxia’s future strategy, the Korean conglomerate is ready to partner — across R&D, shared supply chains, and even co-investing in manufacturing.

The statement is notable not just for what it says, but for what it signals. SK Hynix and Kioxia already sit in a uniquely entangled relationship. Through a special-purpose company, SK Hynix holds the majority of voting rights at Kioxia — a stake reached after displacing Toshiba as the Japanese memory maker’s largest shareholder in August. Under existing agreements, SK Hynix cannot exercise more than 15 percent of Kioxia’s voting power directly or indirectly until 2028, but the conversion of its convertible bonds would shift that calculus.

In practice, the market already treats SK Hynix as Kioxia’s de facto controlling investor. The question now is whether SK will press further — moving from silent majority to active operational partnership.

Why Japan? The Memory Gap Is Forcing SK’s Hand

The strategic logic is clear. SK Hynix faces a structural gap between AI-driven demand and its ability to fill it. Chey told reporters that data-center memory is currently short by 20 to 30 percent globally. That figure is not a forecast — it is a current diagnosis. Generative-AI workloads have multiplied the memory footprint per server, and Korean domestic capacity alone cannot close the deficit in time.

Japan, with its dense cluster of semiconductor equipment and materials suppliers, represents the fastest route to additional NAND capacity. Local governments have already made factory-lease proposals, Chey confirmed, and SK Hynix plans to announce a specific investment direction before year-end. This is not exploratory chatter; it is a timeline.

Chey has previously praised Japan’s semiconductor ecosystem in interviews with Nikkei. He has also floated the idea of a Korea-Japan economic community under the banner of the Korea Chamber of Commerce and Industry — a vision of institutionalized tech cooperation that runs parallel to, and somewhat ahead of, diplomatic thawing between Seoul and Tokyo.

The NAND Market Would Shift Overnight

The competitive implications are significant. According to Counterpoint Research, SK Hynix held 22 percent of global NAND shipments in the second quarter of 2026, while Kioxia held 14 percent. Combined, the two would eclipse Samsung Electronics at 25 percent — a clean sweep of the world’s largest NAND producer.

Such a consolidation would not be a merger. It would be a coordinated partnership with shared supply chains, joint R&D, and potentially co-located production — a structure that could align incentives without triggering antitrust scrutiny in most jurisdictions. For Kioxia, it offers a path to scale. For SK Hynix, it unlocks capacity without the capital intensity of building greenfield fabs from scratch.

An industry source in Tokyo noted that NAND is a sector where many players exist but few can survive on technology alone, and that a pure-play NAND company like Kioxia would be an attractive partner for SK. Kioxia itself confirmed that collaboration on MRAM research is already underway and said it expects the relationship to strengthen.

What Wins, What Loses

SK Hynix stands to gain the most: faster access to Japanese supply-chain depth, additional NAND output, and leverage over Kioxia that goes beyond equity ownership. Kioxia gains manufacturing expertise and a backstop from its largest shareholder at a time when the memory business is becoming increasingly capital-intensive and technologically demanding.

Samsung loses visibility and margin control in NAND. Its nearest competitor — already the second-largest player — would be moving closer to the market leader’s output level, potentially with cost advantages from shared R&D and supply chains. Intel’s foundry ambitions also look dimmer by comparison, since SK Hynix would be deepening ties with an existing Japanese ecosystem rather than outsourcing to a third party.

For Japan, the investment would reinforce its position as a semiconductor materials and equipment hub — a role it has long occupied but never fully leveraged into final-node production. The trade-off is that Japan gains capacity and jobs while ceding some strategic influence over memory chip pricing and supply to a Korean parent.

The Bigger Picture: Korea-Japan Tech Alignment in a Fracturing World

Chey’s proposals are not isolated. They are part of a broader move by Korean industrial leaders to decouple economic cooperation from diplomatic friction. With US-China tensions driving supply-chain fragmentation and export controls tightening across multiple jurisdictions, the practical answer from Seoul and Tokyo has been to talk industry first and politics second.

SK Group’s potential M&A activity among Japanese semiconductor materials and equipment firms — reported as underway alongside the Kioxia talks — would deepen that integration. The result would be a vertically aligned Korea-Japan memory cluster with fewer dependencies on Chinese or Southeast Asian supply nodes, better insulated against the kind of geopolitical shocks that have already disrupted electronics manufacturing in recent years.

The timing matters. If SK Hynix announces a concrete Japanese investment before the end of 2026, it will set a precedent for other Korean companies looking for alternative capacity locations outside Korea and China. If the talks stall, the memory shortage will simply intensify — and Samsung will face less competitive pressure from a weakened Kioxia rather than a partner.

Chey made that exit clause explicit: if no cooperation path opens, SK will wind down its Kioxia investment. That is not empty rhetoric. It is a negotiating posture designed to give Tokyo and Kioxia urgency — and the Korean economy an off-ramp if the deal does not land on favorable terms.