U.S.-China Relations
NextFin News - South Korea is rewriting a 73-year-old espionage statute to treat stolen chip designs as a national-security crime, exposing leakers to up to 30 years in prison as Seoul races to keep China from catching its semiconductor industry.
The revised law takes effect on September 13, 2026, and for the first time since the provision was written in 1953, expands the legal definition of espionage beyond North Korea to any "foreign country or an organisation equivalent thereto." The change elevates industrial secrets from a corporate offense to a threat against the state, reflecting how deeply semiconductors are now woven into South Korea's economy and its alliance with Washington. A companion amendment filed in late August would push the minimum sentence for stealing designated national core technologies from three years to ten and name foreign corporations as covered beneficiaries for the first time.
The Law, the Leak, and the Rival Closing the Gap
The timing is not accidental. Half of all detected cases involving leaks of sensitive technology to foreign entities in 2025 involved China, according to South Korean police data. The most damaging single case on prosecutors' files reads like a blueprint for the new legislation: ten former Samsung Electronics employees stand indicted for allegedly transferring 10-nanometer DRAM process technology to ChangXin Memory Technologies (CXMT), a move the Seoul Central District Prosecutors' Office says cost Samsung about 5 trillion won (roughly $3.6 billion) in lost sales in 2024 alone.
Prosecutors allege the former employees used office relocations and airport layovers to move data undetected, and that a comparison of CXMT materials with Samsung's showed a 98.2% match. Samsung had spent 1.6 trillion won over five years developing that technology. The leak allowed CXMT to mass-produce China's first 10nm-class DRAM in 2023 and laid the groundwork for its push into high-bandwidth memory, the chips that power AI data centers.
CXMT's rise illustrates the stakes. Its share of the global DRAM market was below 1% in early 2024; by the second quarter of 2026 it had reached 10%, according to Counterpoint Research. In July, following its initial public offering, CXMT became China's largest onshore-listed company. Samsung and SK Hynix still command about 38% and 25% of the DRAM market respectively, but the gap that once looked unbridgeable is narrowing — and the new frontier is high-bandwidth memory, where SK Hynix has said its 2026 production is already sold out.
"China's speed is beyond imagination. It's why I feel a sense of intense urgency," South Korea's Industry Minister Kim Jung-kwan said in August.
China's Ministry of Foreign Affairs did not respond to a request for comment.
How the Leaks Actually Happen
The revised law targets the endpoint of a leakage chain that begins with recruitment. Chinese companies, according to Jon Woo-jung, an intellectual-property professor at the Korea Advanced Institute of Science and Technology, keep lists of South Korean specialists to contact with offers of high salaries and benefits when they need specific technology. Departing employees tell their former employers they are moving to a different industry; private investigators hired to check on them report finding the same people hunched over toolmaking blueprints weeks later.
That pattern exposes a structural weakness in South Korea's chip ecosystem. Over 85% of tech leaks between 2020 and 2024 afflicted small or medium-sized businesses, according to data from lawmaker Choi Soo-jin's office. Large chipmakers typically maintain in-house risk-management teams that can detect anomalies and preserve evidence for a criminal case. Their smaller downstream suppliers — the firms that make the materials, components, and tooling that feed the fabs — often lack the resources for extensive security controls and are therefore more vulnerable to poaching.
From 2020 through June 2026, 33 national core technology cases and 110 industrial technology cases were transferred overseas, according to figures cited in the amendment filed by Representative Go Dong-jin of the ruling People Power Party. Those numbers frame the problem as systemic rather than episodic, which is why Seoul is building enforcement muscle alongside the statutory change: the government plans to expand the technology police force from 27 to 61 officers and create a new Technology Divulgence Police Division with 21 investigators, including specialists in electrical, chemical, and mechanical engineering alongside patent examiners and attorneys.
Why the Old Law No Longer Fit
For decades, South Korea's courts were hemmed in by a narrow espionage definition that applied only to those proven to be spying for North Korea. That created a perverse asymmetry: a worker passing DRAM blueprints to Pyongyang could be charged with espionage, while one selling the same secrets to a foreign competitor faced only commercial statutes with lighter penalties and weaker investigative tools. The new Article 98-2 closes that gap, covering anyone who detects, collects, leaks, delivers, or brokers national secrets for a foreign country or equivalent organization.
But the statute still requires prosecutors to prove that the transfer was made "under the direction or instigation" of the foreign entity — a nation-state link that legal analysts say will be difficult to establish in court. Chae Seong-joon, a military-science professor at Seokyung University and a former official with South Korea's spy agency, noted that even with the expanded framework, "it will still be a hard crime to prosecute as you need to prove nation-state involvement."
That enforcement gap is one reason the ruling party is pushing further. Go's amendment, filed on August 31, arrives less than two weeks before the September 13 effective date. It adds "foreign corporation" as an explicit criminal purpose and designates national core technologies — a list maintained by the Ministry of Trade, Industry and Energy across 12 sectors, including advanced semiconductor manufacturing processes — as protected categories. The practical boundary between espionage and legitimate export turns on the "direction or instigation" test, which courts have not yet applied to the new statute.
Seoul is not alone in treating chip know-how as strategic infrastructure. Taiwan, home to Taiwan Semiconductor Manufacturing Co., made economic espionage a crime in 2022 with legislation specifically targeting technology transfer to China. The difference is one of scope: South Korea's revision is its first major overhaul of the espionage statute in 73 years, and it applies to all foreign states rather than a single named adversary.
The American Shadow Over Seoul's Statute
The law has a global dimension that extends well beyond the Korean peninsula. Since 2024, the United States has imposed sweeping export controls on China and pressed allies to restrict sales of advanced chips and chipmaking equipment. In February, the Pentagon blacklisted CXMT, citing security concerns. The revised espionage law effectively aligns South Korea's legal architecture with Washington's containment strategy — and it does so in a way that reaches beyond China.
"The Americans are telling Korea: 'Don't be a weak link, stop leaking all your technology to China, stop exposing your supply chains, stop having Chinese investment in sensitive industries, because ultimately that also affects America,'" said Ben Forney of Lazlo Intelligence, a corporate-intelligence firm.
The statute does not distinguish between adversaries and allies. A Korean engineer who transfers proprietary process data to a U.S., Japanese, or European employer falls inside the same legal framework as one who transfers to a Chinese state-backed chipmaker. Legal experts and defense exporters have flagged this as a concern: routine technology transfers to allied joint-venture partners and defense customers could require careful legal management under the new rules. For South Korea's defense industry, which has spent recent years chasing cross-border export business, the question of where legitimate exporting ends and a criminal act begins has moved from theoretical to urgent.
What the Market Has Priced In — and What It Hasn't
For Samsung and SK Hynix, the immediate backdrop is unusually strong. Semiconductors accounted for 47.1% of South Korea's exports in the first ten days of September 2026, the largest share any single product category has commanded in an early-month flash release, according to Korea Customs Service data. Monthly exports topped $100 billion for the first time in June, driven by record chip shipments. The government recently unveiled an $880 billion plan to develop a massive semiconductor hub in the country's south, a bet that the memory leaders can hold their technological edge.
But the law itself does not create that edge; it only raises the cost of losing it. The second-order question investors should ask is whether the statute changes behavior at the margin where leaks actually occur. If over 85% of leaks originate in small suppliers that lack security budgets, then a law aimed at 30-year sentences for nation-state espionage may deter the headline cases while leaving the bulk of the vulnerability untouched. The enforcement build-out — 61 technology officers and a specialized divulgence division — is a more meaningful signal than the maximum penalty, because it targets the supply-chain tier where the losses actually accumulate.
There is also a compliance cost that the market has not fully priced. Because the law reaches transfers to allied firms, Korean chipmakers and their suppliers will need to vet a wider set of cross-border collaborations, joint ventures, and hiring decisions. That friction is a feature for Washington and a tax on Seoul's global integration. For foreign partners, the near-term risk is not prosecution — the nation-state test is a high bar — but delay and legal uncertainty around talent movement and technology sharing. In a sector where a single generation of process technology can be worth billions, even modest delays in collaborative development carry real option value.
The Counter-Thesis: Deterrence Is Already Working
The strongest case against a skeptical reading is that South Korea's penalties have been climbing for years and leaks have continued anyway — which suggests the binding constraint is enforcement, not sentencing. Courts have handed down stiff terms under existing commercial statutes; an ex-Samsung manager received a six-year sentence in a China leak case. From this view, the new law's real value is not the 30-year maximum but the investigative authority it unlocks: the ability to pursue cases against foreign entities and to embed technical specialists in the investigative process earlier.
That argument has force, but it concedes the central point. If the binding constraint is evidence-gathering rather than penalty severity, then the law's effectiveness will be determined by whether the expanded technology police force can actually prove nation-state direction — the very element Chae and other analysts flag as the hardest to establish. The falsifying signal is straightforward: if prosecutors secure convictions under Article 98-2 within 12 to 18 months of the September 13 effective date, particularly in cases involving foreign corporations rather than state organs, then the deterrence thesis holds. If the docket stays empty or produces only plea deals under lesser commercial charges, the revision will have been more signal than shield.
What to Watch Next
In the short term, watch the legislative track on Rep. Go's amendment and whether it clears the National Assembly before or after the September 13 effective date — a ten-year minimum sentence would reset the deterrent baseline for every pending case. Also watch CXMT: its post-IPO trajectory and whether HBM2 production scales through 2026 will test how much ground Chinese memory makers can still gain without fresh leaks.
Over the medium term, the key data point is the composition of prosecuted cases. A shift from SME-focused commercial charges to Article 98-2 national-security indictments would indicate the new framework is functioning as designed. Watch, too, for the first test of the allied-transfer boundary: any case involving a U.S., Japanese, or European recipient would clarify where Seoul draws the line between espionage and legitimate export.
Over the longer term, the question is structural: semiconductors have been reclassified from commercial assets into national-security infrastructure across the U.S. alliance network, and that regime change is not reversible on its own. Taiwan's 2022 law, Washington's export controls, and now Seoul's espionage revision are three points on the same line. The base case is that the law deters the most brazen transfers while doing little for the small-supplier leakage that accounts for most losses. The upside case is that the investigative build-out catches cases the old statute could not touch, slowing China's catch-up in memory. The downside case is that compliance friction spreads to allied partnerships, raising costs for Korean firms without materially reducing leaks.
South Korea has just declared that a stolen DRAM blueprint is no longer a trade secret but a state secret — and the next 18 months will show whether that declaration changes anything on the factory floor.
Explore more exclusive insights at nextfin.ai.
