The statement was categorical: SK Hynix denied that it is in talks to acquire Intel’s semiconductor campus in Ohio, responding to an exclusive report from Korea JoongAng Daily. The denial was issued the same week Korean media outlets including KED Global and Aju Press, citing people familiar with the matter, reported that negotiations are in fact continuing — and that SK Hynix aims to move a significant portion of its memory chip manufacturing to the United States within five years.
The contradiction is the story. Companies deny deals they are negotiating all the time, and reporters cite sources who contradict denials all the time. What makes this one interesting is the scale of what is at stake: the Ohio campus represents one of the largest semiconductor construction programs in American history, a project that Intel has been shrinking, delaying and rethinking since its foundry ambitions ran into financial reality.
Intel’s Ohio plans have been in flux for two years. The company broke ground on the two-fab complex with great ceremony and CHIPS Act backing, then slowed construction, pushed out timelines and scaled back its foundry strategy as losses mounted. The site’s value lies in what is already built: land, shells, utilities and a workforce pipeline assembled with government subsidies. For a memory maker seeking a fast path into US manufacturing, it is a rare asset — and for Intel, it is a way to recover capital from a bet that did not pay off.
SK Hynix’s interest would make strategic sense. The company is the dominant supplier of high-bandwidth memory, the component at the center of the AI boom, and it has been running its existing plants at capacity to keep up with demand. American customers — including the hyperscale cloud providers building data centers by the dozen — have been pressing memory suppliers for US production, and both the CHIPS Act and the political climate reward domestic manufacturing. A US base for HBM production would hedge against tariffs, shipping risk and the concentration of the world’s advanced memory capacity in South Korea.
The timing adds urgency. SK Hynix is about to report second-quarter earnings that analysts expect to show record profits, powered by the AI memory supercycle — the pricing surge in HBM and DRAM that has made memory the most profitable corner of the semiconductor industry. A company with record cash flow is a natural buyer of a distressed asset, and the stock market is already pricing in its success: the company’s US-listed shares trade at a premium of nearly 33 percent to their Seoul listings, a gap that reflects overseas investors’ appetite for the HBM leader.
The denial leaves room for ambiguity. SK Hynix said it is not in talks — it did not say it has no interest, and it did not rule out future negotiations. Korean media reporting the talks are continuing is consistent with a deal that is alive but not yet formal, or with a company that wants to negotiate without the glare of public attention. In deal-making, the denial itself is often a tell; whether it is a tell that a deal is close or a tell that it is already dead is the question the market cannot answer.
For Intel, selling Ohio would be a retreat dressed as a realignment. The company’s foundry strategy — the plan to manufacture chips for other companies — was its most ambitious attempt to reclaim relevance, and the Ohio campus was its physical centerpiece. A sale to SK Hynix would concede that the foundry dream, at least in that form, is over. But it would also hand Intel billions in cash at a moment when the company needs every dollar for its product turnaround, and it would preserve jobs at a site the government subsidized with that expectation.
For the United States, the deal would be a mixed outcome. A memory fab in Ohio under Korean ownership delivers domestic production and jobs, but it does not deliver the domestic leading-edge logic manufacturing that the CHIPS Act was designed to create. Whether Washington treats a SK Hynix acquisition as a win — a foreign company investing in American production — or as a failure of its flagship industrial policy will shape how the government responds.
The political layer complicates the arithmetic. American officials have made semiconductor self-sufficiency a bipartisan priority, and the prospect of a Korean company taking over a showcase US project would draw scrutiny from both parties, even if the jobs stay. SK Hynix has been careful in its public statements to emphasize its commitment to US investment without confirming any specific project — a posture that lets it negotiate while keeping Washington informed.
The next few weeks should resolve the ambiguity. SK Hynix’s earnings call will give executives a chance to address the reports; Intel’s next public statements on its Ohio plans will show whether the campus is being prepared for a handover or for continued construction. Until then, the two tracks — the categorical denial and the persistent sourcing — will keep running side by side, and the market will keep weighing which one to believe.
