South Korea is projecting calm in the immediate aftermath of the United States’ decision to impose a 25% tariff on certain advanced artificial intelligence chips, but beneath the reassurance lies a deeper unease about where Washington’s semiconductor policy is heading and what it could ultimately mean for Asia’s chip powerhouses.

Trade Minister Yeo Han-koo said on Saturday that the first phase of the U.S. measures would have only a limited impact on South Korean companies, largely because the tariffs target high-end AI processors produced by firms such as Nvidia and AMD, rather than the memory chips that dominate South Korea’s export profile. Samsung Electronics and SK Hynix, the world’s two largest memory chipmakers, derive the bulk of their semiconductor revenue from DRAM and NAND products, which remain outside the scope of the proclamation for now.

Yet Yeo’s remarks were careful not to sound complacent. He warned that it was “not yet time to be reassured,” underlining uncertainty over how quickly and how broadly the United States might expand the measures. The government, he said, would continue close consultations with industry to prepare for potential escalation and to safeguard South Korean interests.

The tariffs stem from a proclamation signed by President Donald Trump on Wednesday following a nine-month investigation under Section 232 of the Trade Expansion Act of 1962, which allows the U.S. to restrict imports deemed a threat to national security. The initial action applies a 25% duty to selected advanced AI chips that meet specific performance benchmarks, including Nvidia’s H200 processor and AMD’s MI325X, both critical components in cutting-edge AI training systems.

The White House has sought to reassure markets by stressing that the tariffs are narrowly tailored. According to an accompanying fact sheet, the duties will not apply to chips and derivative devices imported for U.S. data centers, startups, non-data-center consumer electronics, civil industrial uses, or public sector applications. That carve-out is significant, given that hyperscale data centers account for a substantial share of global demand for AI hardware and remain a key end market for memory chips supplied by South Korean firms.

However, the administration has also been explicit that broader action could follow. The fact sheet said the United States may, in the near future, impose wider tariffs on semiconductors and related products to incentivize domestic manufacturing. That warning was sharpened on Friday by Commerce Secretary Howard Lutnick, who said South Korean and Taiwanese chipmakers that do not invest more heavily in U.S. production could face tariffs of up to 100%.

Lutnick’s comments, delivered at a groundbreaking ceremony for Micron’s new semiconductor plant in upstate New York, underscored the strategic thrust of U.S. policy. Washington is using trade pressure alongside subsidies to accelerate the reshoring of semiconductor manufacturing and reduce dependence on overseas suppliers, particularly in Asia.

This creates a delicate balancing act for South Korea. Its chipmakers are already among the largest foreign investors in U.S. semiconductor manufacturing, drawn by incentives under the CHIPS Act and by the need to stay close to key customers. Samsung is building a multibillion-dollar fabrication complex in Texas, while SK Hynix has announced plans linked to advanced packaging and memory production in the United States.

Even so, the prospect of expanding tariffs raises concerns that trade measures could eventually spill over from logic and AI processors into memory chips or products that incorporate them. That would strike at the core of South Korea’s export economy. Semiconductors account for a significant share of the country’s overseas sales, and any disruption to global chip trade risks knock-on effects for growth, investment, and employment.

There is also anxiety about the precedent being set. Section 232 investigations were once used sparingly, but their application to semiconductors signals a more assertive use of national security arguments in trade policy. For allies like South Korea, which are deeply integrated into U.S.-centric supply chains, this blurs the line between strategic cooperation and economic pressure.

In the short term, analysts say South Korean firms are likely to benefit indirectly from the tariffs, as U.S. restrictions on advanced Chinese chips tighten and demand for memory tied to AI workloads continues to grow. Over the longer term, however, the risk is that an expanding web of tariffs and localization requirements fragments the global semiconductor market, raising costs and complicating investment decisions.

Yeo’s message reflects that dual reality. The first-phase impact may be limited, but the trajectory of U.S. policy points to a more uncertain and politicized trade environment. The challenge for South Korea will be to leverage its strategic importance to the U.S. semiconductor ecosystem while guarding against measures that could, over time, erode the foundations of its chip export dominance.

Elon Musk is seeking as much as $134 billion in damages from OpenAI and Microsoft, arguing that both companies benefited enormously from his early backing of the artificial intelligence startup and that those gains were improperly obtained after OpenAI abandoned its original mission.

The bid is fast becoming one of the most consequential legal battles in the global AI race, not only because of the staggering sums involved, but because of what it could mean for how artificial intelligence is funded, governed, and commercialized.

At its core, the lawsuit is an attempt by Musk to retroactively redefine his role in OpenAI’s origin story — from philanthropist and early supporter to aggrieved architect whose contributions, he argues, laid the groundwork for extraordinary private gains that he was unfairly cut out of.

In the filing submitted on Friday, Musk contends that OpenAI’s transformation from a nonprofit research lab into a for-profit enterprise tightly aligned with Microsoft fundamentally altered the bargain under which he provided early funding, credibility, and strategic support. He argues that those changes enabled OpenAI to generate between $65.5 billion and $109.4 billion in value, while Microsoft extracted a further $13.3 billion to $25.1 billion through its exclusive cloud, licensing, and distribution arrangements.

The numbers themselves are striking, but so is the legal theory underpinning them. Musk is not claiming conventional investor damages. Instead, he is seeking “disgorgement” — the return of what he describes as “wrongful gains” — a remedy more commonly associated with fraud or breach of fiduciary duty than with disputes between founders and donors. His filing explicitly likens his role to that of an early startup investor who took existential risk and therefore deserves outsized returns once the venture succeeds.

OpenAI and Microsoft reject that framing outright. They argue that Musk was never promised equity-like returns, that OpenAI’s restructuring was necessary to raise capital at the scale required to compete in frontier AI, and that Musk’s claims amount to an unprecedented attempt to siphon value out of an organization he voluntarily left seven years ago.

The companies’ counter-filing attacks the credibility of Musk’s damages model, warning that it invites jurors to speculate about hypothetical alternate histories of OpenAI — scenarios in which Musk remained involved, the nonprofit model remained intact, and Microsoft’s role was either reduced or nonexistent. They argue that such speculation is not only unverifiable but fundamentally misleading.

Beyond the courtroom, the case exposes deeper fault lines in the AI ecosystem.

First is the unresolved tension between public-interest AI and commercial reality. OpenAI was founded on the idea that artificial general intelligence should benefit humanity broadly, not concentrate power or wealth. Musk’s lawsuit leans heavily on that original mission, portraying the Microsoft partnership and profit-seeking structure as a departure so severe that it invalidates earlier commitments. OpenAI counters that without commercialization, the organization would have been unable to fund the computing power, talent, and infrastructure needed to remain competitive.

Second is the question of control. Musk left OpenAI in 2018 after disagreements over leadership and direction. Since then, the company has become one of the most influential forces in technology, while Musk has launched xAI to compete directly with it. That competitive overlap looms over the case. OpenAI and Microsoft have already characterized Musk as a “donor-turned-competitor,” suggesting his legal action is as much about weakening a rival as it is about righting alleged wrongs.

Third is the precedent the case could set. If a jury accepts Musk’s argument, it could unsettle how nonprofits, research labs, and mission-driven organizations attract early funding. Future donors and founders may demand clearer exit rights, financial upside, or contractual safeguards, potentially accelerating the shift away from nonprofit structures in cutting-edge technology fields.

The stakes are heightened by the remedies Musk is seeking. In addition to massive financial damages, his filing leaves open the possibility of punitive penalties and injunctive relief. Even a narrowly tailored injunction could disrupt OpenAI’s governance or constrain aspects of its partnership with Microsoft — a relationship that underpins products across cloud computing, enterprise software, and consumer AI services.

The case carries reputational as well as financial risk for Microsoft. While the company insists it did not aid any breach of OpenAI’s mission, a prolonged trial will inevitably scrutinize how much influence it exerts over OpenAI’s strategy and whether its commercial interests shaped decisions that were once framed as altruistic.

The dispute is gradually evolving into a referendum on the economics of AI itself: who bears the early risks, who captures the rewards, and whether lofty founding ideals can survive once AI systems become engines of immense profit and power. Whatever the jury decides, the outcome is likely to ripple far beyond Musk, OpenAI, and Microsoft, shaping how the next generation of AI ventures are built, funded, and controlled.