3-Line Briefing
- ARM — which Nvidia failed to acquire even after mobilizing roughly 61 trillion won, blocked by regulatory hurdles — is once again in the spotlight as the de facto standard supplier in the AI chip design IP market.
- ARM does not manufacture chips itself but licenses out its designs under a license-and-royalty model, making it a tollgate player that collects fees from nearly every chipmaker as AI chip shipments grow.
- For Korean investors, the key point is that the profit and loss across the entire supply chain moves together — from Samsung Electronics (005930) and SK hynix, which use ARM cores, to Nvidia, which designs accelerators on top of ARM.
What's Changing
The essence of this issue is not simply a failed-acquisition story, but the structural question of who controls the chokepoint of semiconductor design in the AI era. The fact that Nvidia, despite committing enormous capital, failed to get its hands on ARM means that competition authorities viewed the ripple effects of a single company monopolizing the design standard as that significant. In other words, it is a signal that ARM's very neutrality is being valued as shared infrastructure for the AI chip ecosystem.
What has changed is that ARM's center of revenue gravity is shifting rapidly from a smartphone focus toward data centers, AI accelerators, and automotive. In the past, mobile application processors (APs) drove earnings, but now the licensing of high-value cores going into AI servers and the per-chip royalties have emerged as the growth engine. Because the royalty rate ARM earns rises with the design complexity of the latest cores, the same shipment volume contributes more to revenue.
The crucial part from Korea's standpoint is the composition of downstream customers. A significant share of the mobile and server chips made by Samsung Electronics and SK hynix, Nvidia's Grace-series CPUs, and even Apple silicon are all built on the ARM instruction set architecture. Shifts in ARM's bargaining power directly affect these customers' costs and design flexibility.
Reading the Numbers and Context
The roughly 61 trillion won cited in the headline is the deal value at which Nvidia sought to buy ARM from SoftBank. Announced in 2020 at around $40 billion, the valuation swelled as Nvidia's share price rose, and the deal ultimately collapsed in 2022 amid opposition from multiple regulators including the UK, the US, and the EU. ARM then chose an independent path, listing directly on the Nasdaq in 2023.
The core mechanism is a dual structure of upfront license fees and per-chip royalties. Customers pay a license fee when they adopt a design, and subsequently pay additional royalties in proportion to the volume of chips produced. Because trailing royalties accumulate as AI accelerators and data center CPUs proliferate, investors should look not only at short-term earnings but also at the pipeline of adopted designs that lead to mass production over several years.
Beneficiary and Affected Stocks
- ARM: As the standard supplier of AI chip design, it sits at the peak of structural benefit, where rising shipments and adoption of high-value cores push up royalty rates.
- Nvidia: A key customer that designs its own data center CPUs on the ARM architecture to expand its AI acceleration platform — a direct beneficiary of ARM's ecosystem expansion while also bearing the cost burden, both sides of the coin.
- Samsung Electronics (005930): Expected to see improved utilization in ARM-core-based mobile and server chip design and foundry orders if AI chip demand expands.
- SK hynix: A memory-side beneficiary, as demand for high-bandwidth memory (HBM) for AI servers moves in step with the spread of ARM-based accelerators.
- SoftBank Group: As ARM's largest shareholder, changes in its stake value are directly reflected in the group's asset value.
Risk Check
- Valuation burden: Since its listing, ARM has had future growth expectations priced into its shares, leaving persistent debate over overvaluation relative to earnings.
- Rise of alternative camps: If large customers partially adopt the royalty-free open instruction set architecture (RISC-V), ARM's long-term bargaining power could weaken.
- Customer concentration and conflict: Some key customers are also competing design houses, so if in-house chip development accelerates, the royalty base risks being shaken.
- Macro variables: If the AI investment cycle slows or data center capital expenditure turns down, the pace of royalty accumulation will slow alongside it.
One-Line Conclusion
ARM holds a structural growth narrative thanks to its scarcity as the shared design chokepoint of the AI chip ecosystem, but with a high valuation and the counterweights of RISC-V and customer in-housing operating at the same time, this is a stretch to weigh expectations against risks by checking quarterly royalty revenue trends, shifts in the data center mix, and large customers' design-adoption disclosures at every quarterly earnings release.
This article is content automatically summarized and analyzed based on the original news report. View original (Maeil Business Newspaper, Securities)
