Summary

A May 15, 2026 exclusive report by Bloomberg revealed that the U.S. Federal Trade Commission (FTC) has been pursuing an antitrust investigation into UK-based Arm Holdings, a subsidiary of SoftBank Group. The FTC is reportedly examining whether Arm has been attempting to illegally monopolize a portion of the semiconductor market by expanding its own chip business while refusing to license its CPU (central processing unit) designs or degrading their quality. Following the report, Arm shares fell roughly 8% the same day, wiping out approximately 3 trillion yen in market capitalization in a single day. This article first lays out the fundamentals of the FTC and antitrust law, then examines the differences among various news outlets' reporting and the history of the dispute with Qualcomm, before offering a multifaceted reading—from the perspective of Silicon Valley venture capital (VC)—of what this investigation reveals about the emergence of "platform risk" and the shift of capital toward RISC-V.

The Day the "Switzerland of Semiconductors" Stands in the Defendant's Dock for Monopoly

Its technology is used in more than 99% of the world's smartphones, and roughly 70% of the global population interacts with its designs every day — so the UK's Arm Holdings has long described itself. On May 15, 2026, the company found itself in an unexpected position. That day, a Bloomberg exclusive revealed that the U.S. competition authority, the Federal Trade Commission (FTC), had been pursuing an antitrust investigation into Arm's semiconductor technology licensing business. The article was based on accounts from multiple people familiar with the matter, and as the cautious phrasing "Said to face" in the headline suggests, this was not an official announcement by the authority but a scoop based on insider testimony.

According to the report, the FTC is investigating whether Arm "is illegally attempting to monopolize a portion of the semiconductor market." The focus is clear: Is Arm refusing to license its blueprints for designing CPUs (central processing units) to third parties, or degrading the quality and terms of such licenses, while at the same time accelerating the development of its own in-house chips? The FTC is said to have notified Arm of the investigation earlier in 2026 and demanded the preservation of related documents. A document preservation request is a typical opening move signaling that a regulator has launched a full-fledged investigation.

For 35 years, Arm has not manufactured specific chips itself, but rather licensed its blueprints and instruction set (the common language by which software communicates with the processor) to any company. Nvidia, Apple, and Qualcomm — fiercely competing with one another — all use Arm's technology as a common foundation. Precisely because Arm competes with none of them, everyone can safely climb aboard. It is due to this neutrality that Arm has often been called the "Switzerland of the semiconductor industry." Now, this guardian of neutrality is itself under suspicion of monopolization. In this article, we will start with the basics of what the FTC and antitrust law are, examine the discrepancies in how various newspapers reported the story, trace the long history of conflict with Qualcomm, and then take a step further to read how Silicon Valley VCs are receiving this episode, where new moves may emerge going forward, and the implications for Arm shares and SoftBank Group shares. Note that dollar-denominated figures in this article, unless otherwise specified, are accompanied by reference values converted at approximately ¥159 to the dollar (the exchange rate level as of May 15, 2026).

What Is the FTC? The Independent Regulatory Agency That Oversees Competition in the U.S. Market

The FTC (Federal Trade Commission) is an independent U.S. agency established under the FTC Act (Federal Trade Commission Act) enacted in 1914. Headquartered in Washington, D.C., its mission is broadly divided into two areas. One is competition policy (antitrust) to protect competition in the marketplace, and the other is consumer protection to shield consumers from unfair business practices. What concerns Arm falls under the former, an area handled by the Bureau of Competition.

The FTC is run by five commissioners. Commissioners are nominated by the president and appointed following Senate confirmation, with a seven-year term. No more than three may come from the same political party, by design preventing political tilt to one side. The chair is nominated by the president, and as of May 2026, the chair is Andrew Ferguson, appointed by the Trump administration in January 2025. On the operational side, the FTC requires companies to preserve documents, and gathers materials and testimony through CIDs (Civil Investigative Demands) and subpoenas. Once suspicions of a violation solidify, the FTC chooses—based on grounds such as Section 5 of the FTC Act—whether to bring the case into its internal administrative adjudication (going through a hearing before an administrative law judge, then to the Commission, and further to a federal court of appeals), or to file suit in federal district court. The "document preservation demand" sent to Arm this time marks the entry point of this investigative process.

What is important is that, for the FTC, Arm and SoftBank are by no means unfamiliar counterparts. The FTC has a clear prior history. In December 2021, the FTC moved to file suit to block Nvidia's plan to acquire Arm for roughly $40 billion (the figure based on the FTC's official designation; converted at the exchange rate, this comes to roughly over 6 trillion yen). Under then-chair Lina Khan, the FTC argued that Arm's technology is an "essential input" supporting competition between Nvidia and its rivals across multiple markets, and that if Nvidia gained control of it, Nvidia would gain access to competitively sensitive information of Arm's licensees (some of whom are Nvidia's rivals), giving it both the ability and the incentive to weaken its competitors. Compounded by pressure from regulators in various countries, Nvidia abandoned the acquisition in February 2022. The FTC withdrew its suit. Furthermore, in 2025, the FTC issued a "Second Request (a full-scale demand for additional information)" with respect to SoftBank's acquisition of U.S.-based Ampere Computing. The competitive issues surrounding Arm are terrain that U.S. authorities have already examined many times over.

There is an ironic structure here. The FTC's 2021 logic, stated in one sentence, was: "To preserve Arm's neutrality, we cannot allow it to be placed under the ownership of a particular company, namely Nvidia." Arm's continuing to be a neutral licensor not under anyone's umbrella—this, the FTC believed, was what supports competition across the entire semiconductor industry. Yet in 2026, what the FTC is concerned about is not the question of who the owner is, but the very act of Arm itself ceasing to be a neutral foundation and descending downstream with its own chips. The very neutrality it sought to protect is being relinquished by none other than Arm itself. The tone of antitrust enforcement has changed with the change in administration, but the basic posture of scrutinizing big tech has been maintained, and herein lies the basis for viewing the investigation of Arm as proceeding with continuity regardless of the administration's political color.

Understanding Antitrust and Anti-Monopoly Law: From the Sherman Act to EU and Japanese Competition Law

"Antitrust Law" is the general term for anti-monopoly law in the United States. The etymological root, "trust," refers to the holding and trust arrangements that giant corporations used in the late 19th century to dominate markets in sectors such as oil and railroads. Against the backdrop of public backlash against such monopolies, exemplified by Standard Oil, the United States enacted the Sherman Act in 1890. This is the origin of antitrust law, and it shares the same underlying philosophy as Japan's "Anti-Monopoly Act" and the EU's "Competition Law." Despite the differing names, all of these legal frameworks share the common purpose of safeguarding fair competition in the market.

U.S. antitrust law is broadly structured around three pillars. The first is the Sherman Act: Section 1 prohibits "restraints of trade" that unreasonably restrict competition, such as cartels and bid-rigging, while Section 2 prohibits "monopolization," "attempts to monopolize," and "conspiracies to monopolize." The second is the Clayton Act, enacted in 1914, whose Section 7 prohibits mergers and acquisitions that may substantially lessen competition. When the FTC blocked Nvidia's acquisition of Arm in 2021, it was this Clayton Act lineage that was invoked. The third is the FTC Act, also enacted in 1914, whose Section 5 broadly prohibits "unfair methods of competition." The FTC has interpreted this Section 5 as a distinct authority that encompasses Sherman Act violations while extending to a broader range of conduct. In the United States, the FTC and the Antitrust Division of the Department of Justice (DOJ) jointly enforce competition law under a dual structure.

The core of this investigation lies in the doctrine of Section 2 of the Sherman Act. What is easily misunderstood here is that "monopoly itself is not illegal." Achieving a high market share through a superior product, exceptional managerial judgment, or even historical happenstance is, on the contrary, permitted as the fruit of competition. What becomes illegal is the acquisition, maintenance, or expansion of monopoly position through "exclusionary conduct" that unjustly forecloses competitors. The accusations currently directed at Arm fall within the domain of "monopolization" as contemplated by Section 2 of the Sherman Act—or more precisely, "attempted monopolization," a category whose elements require exclusionary conduct, specific intent, and a dangerous probability of success. Indeed, multiple reports indicate that the FTC is examining "whether Arm has the ability or intent to refuse or degrade licensing," which aligns with the attempted monopolization framework that turns on "capability and intent."

A key conduct category here is the concept of "refusal to deal." As a general rule, even a dominant firm bears no general obligation to assist its competitors. However, U.S. Supreme Court precedent (such as the Aspen Skiing case) has held that suddenly terminating a long-standing, profitable cooperative relationship—sacrificing short-term profits in order to crush a competitor—can constitute a Section 2 violation. Qualcomm's argument, discussed later, that "Arm operated an open network for more than 20 years and has only now begun to close it," is structured precisely to invoke this judicial doctrine.

Competition law is not unique to the United States. In the EU, Article 101 of the Treaty on the Functioning of the European Union (TFEU) prohibits anti-competitive agreements, while Article 102 prohibits "abuse of a dominant market position," with enforcement carried out by the European Commission's Directorate-General for Competition (DG COMP). Refusal to supply and discriminatory licensing terms fall squarely within the abuse categories that Article 102 captures. In Japan, the Anti-Monopoly Act enacted in 1947 (formally, the "Act on Prohibition of Private Monopolization and Maintenance of Fair Trade") prohibits private monopolization, unreasonable restraints of trade, and unfair trade practices as its three pillars, and is administered by the Japan Fair Trade Commission (JFTC). What is closest to the current Arm matter would be "refusal to deal" and "exclusive dealing" within unfair trade practices, as well as the type of "private monopolization" that excludes the business activities of other companies. South Korea also has a Korea Fair Trade Commission (KFTC), which polices abuse of a dominant market position. In short, what the United States calls "antitrust law" is simply called "Anti-Monopoly Act" in Japan and "competition law" in the EU; and a company operating globally, such as Arm, will find itself simultaneously under the scrutiny of multiple such jurisdictions.

How Each Newspaper Reported It — Bloomberg's Exclusive Scoop and the "Unable to Confirm" Caveat

This investigation came to light not through an official announcement by the authorities, but through media reports. The first to report it was Bloomberg, in its May 15 article "Arm Holdings Said to Face US Antitrust Probe Over Chip Tech." Citing "people familiar with the matter" as its source, the article reported that the FTC was investigating Arm's semiconductor technology licensing, and that the notice of investigation and document preservation requests had been issued in early 2026. In Japanese, Bloomberg Japan distributed it the following day, on the 16th, under the title "US FTC Conducts Antitrust Investigation into Arm – Sources," and it was reposted on outlets such as Yahoo! News.

Reuters then followed up by citing Bloomberg's reporting. Here lies a reservation that cannot be overlooked when reading each outlet's coverage. Reuters explicitly stated that, regarding Bloomberg's description that the South Korean authorities' investigation originated from a complaint filed by Qualcomm, "Reuters was unable to independently confirm this." Reuters further wrote that "Arm declined to comment, and the FTC did not immediately respond to a request for comment." Clearly identifying primary sources, and explicitly noting where its own verification has not been possible—this restrained posture is something that cannot be ignored in assessing the reliability of the reporting. Many outlets, including Investing.com and Malaysian media, reported the story by reposting this Reuters dispatch, sharing the same skeleton: "the FTC is investigating whether there has been illegal monopolization," "Arm was notified in early 2026 with a request for document preservation," "the dispute with Qualcomm forms the background," and "South Korean authorities conducted an on-site inspection of the Seoul office in November 2025."

On the other hand, some outlets put their emphasis on stock prices and market reactions. In addition to the content of the FTC investigation, Investing.com also reported that Arm's stock fell 8.46% to close at $209.16, and that Arm's in-house chip plans are projected to generate $15 billion (approximately ¥2.39 trillion) annually within five years. Outlets such as CoinCentral and AInvest led with the headline of "investors' hypersensitive reaction to regulatory risk," reflected in the stock's sharp drop and surging trading volume. In the Japanese-language sphere, the Nikkei has continuously covered the FTC, SoftBank, and Arm, and NOVAIST published an explanatory article that traced the timeline of the Nuvia-related litigation in detail.

There are two points that should be explicitly noted here as "discrepancies in reporting." First, whether the investigation originated from a Qualcomm complaint was reported by Bloomberg, but Reuters appended the caveat that it "could not independently confirm" this—meaning the handling of certainty varies by outlet. Second, there is variation in the magnitude of the stock price decline as well. Investing.com gave the regular-session closing price as "down 8.46% at $209.16," Reuters-affiliated articles reported a very modest initial reaction of "less than a 1% decline in Friday after-hours trading, $207.96," and TradingKey, which deals with market data, recorded the May 15 decline at 7.47%. These are not contradictions, but rather can be read as reflecting differences in time periods: against a regular-session close (around $209) that had already fallen significantly, after-hours trading was nearly flat ($207.96). Presenting such differences as they are, rather than rounding them down to a single figure, is the starting point for reading this matter calmly. And what is decisively important is the fact that, at this point, this entire matter is a "report" relying on Bloomberg's anonymous sources, and that neither the FTC nor Arm has officially acknowledged it. A highly reliable scoop and an official announcement by the authorities are two different things.

What Is Being Deemed "Suspected of Being Illegal" — The Withholding of CPU Blueprint Licenses

First, it must be emphasized that the FTC has not yet filed a complaint, and there is no fact that it has determined any specific conduct by Arm to be "illegal." The current stage is strictly an investigation, and what appears in the reporting are "suspicions" and "points of contention." With that in mind, let us look concretely at what is being treated as problematic.

Arm's revenue has a two-tier structure. One is the up-front fee (license fee) that companies pay when they enter into a contract to use Arm's designs or instruction set architecture (ISA). The other is the royalty that accrues each time a chip using that design is shipped. Of the $4.92 billion (approximately ¥780 billion) in total revenue for the fiscal year ending March 2026 (FY2026), royalty income reached $2.61 billion (approximately ¥415 billion). High-margin, long-term revenue that keeps flowing in as long as chips continue to go out into the world—this is the core of Arm's business, and its source lies in the fact that Arm holds the ISA, a "common language that everyone needs."

There are broadly two forms of license. One is the "TLA (Technology License Agreement)," which grants the right to use Arm's already-designed cores (such as Cortex and Neoverse) as-is. The other is the "ALA (Architecture License Agreement)," a more advanced and rare contract that grants the right to independently design proprietary CPUs that comply with Arm's ISA. Apple, Nvidia, and Qualcomm have all held this ALA. What the FTC is concerned about is the structure in which Arm, holding the essential foundation that is the ISA, simultaneously steps into the position of selling its own chips, and may then have the "ability and motive" to withhold licenses from rival third parties, delay providing its best technology, or impose unfavorable terms. If the referee also plays as a player, strategic information about other players ends up in the hands of a competitor, and the way the whistle is blown could be skewed in their own favor—this self-preferencing, discriminatory licensing, and vertical foreclosure are the very heart of the investigation. As specific conduct, what is envisioned includes refusing licenses themselves, intentionally degrading terms or quality, delaying access to new versions, and selectively raising royalties.

The specific conduct frequently cited is the allegations Qualcomm has made to authorities in various countries. According to Qualcomm, Arm operated an "open network" available to anyone for more than 20 years, yet in recent years it began restricting access to its technology. In particular, Qualcomm complains that it was not allowed to use, without Arm's consent, the Arm license held by Nuvia, the design company Qualcomm acquired, and furthermore that Arm dangled the possibility of revoking its architecture license. Arm flatly denies this, retorting in strong language that Qualcomm's claims of anticompetitive conduct are "baseless and nothing more than a desperate and underhanded attempt to gain leverage for its own competitive interests in an ongoing commercial dispute." However, this comment is purely a rebuttal to Qualcomm's claims, not Arm's official explanation of the FTC investigation itself. Note that Arm's own statement that it has come to account for roughly 50% of CPU compute share among hyperscalers (major cloud operators) in the data center market is a figure that regulators are likely to look at when measuring "market power."

All-Out War with Qualcomm — From the Nuvia Lawsuit to Simultaneous Global Litigation

The current FTC investigation did not come out of the blue. It represents the "regulatory front" of an all-out, multi-year war between two semiconductor giants: Arm and Qualcomm. The trigger came in 2021, when Qualcomm acquired Nuvia, a startup developing high-performance CPUs for servers, for approximately $1.4 billion (about ¥223 billion). Arm argued that the ALA rights Nuvia held with Arm did not automatically transfer to Qualcomm, and that the royalty rates differed as well; in 2022, it sued Qualcomm in the federal district court in Delaware, even demanding the destruction of designs derived from Nuvia.

The outcome of the courtroom battle leaned toward a victory for Qualcomm. In the December 2024 jury trial, it was recognized that Qualcomm had not violated its own ALA, and that the Snapdragon X—equipped with Qualcomm's proprietary CPU core "Oryon," which incorporated Nuvia-derived technology—was properly licensed (the jury was split on whether Nuvia had violated its original license). From September to October 2025, the judge at the Delaware federal district court finalized Qualcomm's "complete victory" and denied Arm's motion for a retrial. Arm has indicated its intent to appeal the ruling. Furthermore, the countersuit Qualcomm filed against Arm (alleging breach of contract and tortious interference) remains pending, and the legal battle between the two companies continues.

Having stumbled in court, Qualcomm expanded the front to regulatory authorities. In March 2025, Qualcomm was reported to have filed antitrust complaints with competition authorities in the U.S., EU, South Korea, and elsewhere, alleging abuse of Arm's dominant market position. Qualcomm claims that Arm has restricted access to licenses and is withholding critical technology. In November of the same year, the Korea Fair Trade Commission (KFTC) conducted a surprise on-site inspection at Arm's Seoul office. According to the Korea Times, the KFTC's investigation is directed at suspicions that Arm has reversed the open posture it maintained for many years and is undermining fair competition by restricting access to its technology. As for the EU, while multiple outlets have reported that Qualcomm filed a competition-law complaint with the European Commission, no official announcement that the Commission has launched a formal investigation has been confirmed, and at present it is more accurate to view the matter as being at the "complaint pending" stage.

Lined up this way, the U.S. FTC investigation that surfaced in 2026 can be positioned as the latest and heaviest move in the encirclement Qualcomm has mounted on a global scale. That said, from a VC's perspective, it is worth adding that even if a single company's complaint is the entry point, once a competition authority opens an investigation, its scope can extend beyond the complainant's interests to Arm's licensing practices broadly, and even to the terms of its data center deals. "Narrow at the entrance, wide at the exit"—this characteristic is precisely what matters when evaluating regulatory risk.

Arm's Tectonic Shift—Abandoning "Neutrality" to Sell Its Own Chips

Why is Arm's neutrality being called into question now? The roots lie in a tectonic shift in Arm's own business model. Founded in 1990, the company was acquired and taken private by SoftBank Group in 2016 for approximately £24 billion (roughly ¥3.3 trillion at the exchange rate of the time), and relisted on the Nasdaq on September 14, 2023. The offering price was $51 per share, and the amount raised was approximately $4.87 billion (about ¥770 billion), making it the largest IPO of that year. SoftBank still holds roughly 90% of the outstanding shares.

What sustained Arm throughout its 35-year history was the principle that "we sell designs but don't make chips, so we don't compete with anyone." On March 24, 2026, however, Arm broke this principle itself. It announced its first in-house chip, the "Arm AGI CPU." This is a full-scale mass-production CPU for AI data centers, integrating 136 of Arm's Neoverse V3 cores, manufactured on TSMC's 3-nanometer process, and is claimed to deliver up to twice the performance of x86. The first customer is Meta Platforms, which is said to be deploying this chip in combination with its own AI accelerator "MTIA" and to have signed a multi-generational collaboration roadmap. OpenAI, Cloudflare, and Cerebras have also been named as deployment partners. Arm has indicated that it expects this chip business to reach annual revenue on the scale of $15 billion (about ¥2.39 trillion) within five years. Arm, which had been the side that "lent out" designs, has shifted to becoming the side that "sells" chips it designed itself, beginning to compete on the same playing field as its former customers. Furthermore, SoftBank acquired the U.S. firm Ampere Computing, which makes Arm-based server CPUs, for $6.5 billion (about ¥1.03 trillion) (announced in March 2025, completed in November of the same year), strengthening the group's capabilities in data-center chips. Prior to this, it had also been reported that Arm was considering plans to raise royalty rates for the Armv9 generation by up to 300%.

Arm's CEO Rene Haas explains this shift as "customer-driven." In an interview on Stratechery, run by prominent tech analyst Ben Thompson, Haas said, "If there's a benefit, customers will buy it; if not, they won't," and "Since that day, we have continued to offer all the products we had been offering. A chip has simply been added to that lineup," emphasizing that Meta itself had asked Arm to complete the design. Just as Google continues to buy Nvidia's GPUs even while having its own in-house TPUs, the market adjusts autonomously—such is his line of reasoning. But the structural issue that regulators see lies precisely in the one point that interview did not delve into. Once a referee also becomes a player, the competitively sensitive roadmaps of each licensee end up "in the hands of a competitor." The company's framing of "the market will decide" versus the authorities' framing that "the owner of the foundation has incentives to shut competitors out"—how one reads the gap between these two will be the dividing line in how this case is evaluated.

Silicon Valley VCs' Take—"Platform Risk" Has Become Reality

From here, I will offer an integrated reading of this investigation from the perspective of Silicon Valley venture capital (VC), which is the central focus of this article. A VC, distilled to its essence, is an entity that underwrites and prices "platform risk." Which foundation does a startup build its business upon, and when and how might the owner of that foundation change its behavior?—discerning this is the core work of a VC. There is a lesson repeatedly told in the VC world. The case of apps that depended on Twitter's (now X's) API and were suddenly locked out; "Sherlocking," in which Apple itself absorbs a feature that grew on the App Store by bundling it as standard—these illustrate the danger of riding on a third-party platform you cannot control.

In that context, Arm was an exceptionally unusual presence. Precisely because it remained a "neutral foundation that competed with no one" for 35 years, semiconductor and AI hardware startups could confidently build their businesses on top of Arm IP. Arm was one of the few "safe dependencies." However, with the launch of the AGI CPU and SoftBank's acquisition of Ampere, Arm became both a supplier and a potential competitor. The FTC investigation means that the authorities have officially recognized as a "point of contention" a risk that VCs had already begun to be aware of. The practical question for VCs is this: when Arm (and Ampere under SoftBank) can compete downstream and also controls licensing terms, can startups confidently build chip businesses on top of Arm IP? This structure—"the platform competes with its own ecosystem"—follows the same pattern as the disputes over Amazon's marketplace and its private-label brands, and the operation of the App Store, and it is a pattern with a sense of déjà vu for VCs. Arm is none other than the semiconductor version of this.

Another reason VCs and crossover investors take this investigation seriously lies in the "mechanism of Arm's valuation." The bullish scenario for Arm, and by extension the basis for SoftBank Group's net asset value (NAV), rests on Arm's pricing power—its ability to raise royalty rates and capture more value per chip. The reported plan to raise rates by up to 300% is emblematic of this. Yet the antitrust investigation targets the very "mechanism itself" of that pricing power. If regulators impose constraints on license differentiation or self-preferencing, the premise of "valuing Arm like a growth stock even though it's a royalty business" will be shaken. From a VC perspective, this investigation is not a mere transient headline, but a specific risk factor that affects the valuation multiple.

And the impact on the arithmetic of exits cannot be ignored either. Regulators' gaze toward acquisitions involving chips and AI has clearly sharpened. On March 20, 2026, Democratic Senators Elizabeth Warren and Richard Blumenthal sent a letter regarding Nvidia's roughly $20 billion (about ¥3.2 trillion) deal involving Groq. This is a technique known as a "reverse acquihire," which combines technology licensing with the poaching of core talent to effectively take control without buying the company itself, and the two senators asked whether this might constitute an evasion of antitrust law that bypasses merger review. The FTC blocked the Nvidia-Arm acquisition in 2021, issued a second request on the SoftBank-Ampere acquisition in 2025, and is now examining Arm's licensing practices themselves—this sequence shows that regulatory scrutiny of the semiconductor IP layer is structurally intensifying. For deep-tech and semiconductor startups, M&A is the dominant exit, and as regulatory friction grows, that pathway becomes slower, narrower, and more conditional. On the other hand, conversely, it also means that it becomes harder for dominant firms to enclose entire ecosystems, leaving room for independent startups to scale on their own and make it all the way to a public listing.

RISC-V as Insurance for VCs—The Money Is Already Moving

The VC community's reception goes beyond abstract debate. Capital is already moving in concrete ways. Its destination is "RISC-V." RISC-V is an instruction set architecture whose ISA is openly published, has no specific owner, and incurs neither licensing fees nor royalties. As "insurance" against Arm's licensing dominance, VCs have been investing heavily here.

A symbolic example is U.S.-based SiFive, which develops RISC-V processors. In April 2026, the company completed a $400 million (approximately ¥64 billion) funding round, bringing its valuation to $3.65 billion (approximately ¥580 billion). The round was led by Atreides Management, with participants including Nvidia, hedge fund Point72, and Apollo Global Management. Reports position this as the final private round before going public. Funding for RISC-V startups is spreading elsewhere as well. Akeana, an emerging company that recently exited stealth, raised over $100 million (approximately ¥16 billion), with backers including Kleiner Perkins, Mayfield, and Fidelity. AheadComputing, founded by former Intel engineers, raised a $21.5 million (approximately ¥3.4 billion) seed round led by Eclipse. Corporate movements are pointing in the same direction. Qualcomm acquired Ventana Micro Systems, which develops RISC-V cores, seeking to secure "roadmap sovereignty" that avoids dependence on Arm and licensing disputes.

The VC logic is simple. Every time Arm tightens its licensing, every time the dark clouds of regulation grow denser, the relative value of an alternative architecture that requires no royalties and has no single owner rises. Frankly, the FTC investigation is a tailwind for portfolios invested in RISC-V. The trend demanding "silicon sovereignty" in semiconductors—the desire of companies to build chips uniquely optimized for the AI era without being bound by anyone's permission—is further propelled by the investigation into Arm's conduct. SiFive's anticipated IPO from late 2026 onward, along with announcements of expanding RISC-V adoption, are worth reading as the market's "revealed preference." If the market seriously believes Arm's regulatory risk is high, then capital and talent should flow more quickly toward open architectures.

What must not be forgotten is the fact that SoftBank Group itself is one of the world's leading tech investors—that is, a massive VC. Arm is at the core of SoftBank's strategy to transform into an "AI holding company," serving as collateral underpinning its enormous investment in OpenAI and its AI infrastructure vision, and also as its funding engine. Dark regulatory clouds over Arm are equivalent to dark clouds over SoftBank's entire AI strategy. The VC community is closely watching this investigation not as an issue concerning Arm alone, but as a matter touching the preconditions for the entire AI infrastructure investment boom.

Impact on Arm and SoftBank Group shares—the "perfect price" put to the test

Let us accurately organize the market's reaction, including the variations in reporting. On Friday, May 15, the day the news broke, Arm shares plunged. Investing.com reported the regular-session closing price down 8.46% at $209.16 (approximately ¥33,000), while TradingKey, which handles market data, recorded the day's decline at 7.47%. Meanwhile, an article distributed by Reuters immediately after the initial report stated that in after-hours trading the share price fell "less than 1%, to $207.96." This is not a contradiction; it can be read as reflecting a difference in time frames: against the regular-session close (about $209), which had already ended sharply lower, after-hours trading was nearly flat ($207.96). Cross-referencing the various reports, the most reasonable interpretation is that Arm shares closed down approximately 8% in regular trading on May 15.

The impact of this decline is not small. Arm's market capitalization stood at around $220–230 billion (approximately ¥35–37 trillion) as of mid-May, meaning the roughly 8% plunge wiped out about $19 billion (approximately ¥3 trillion) in market value in a single day. The backdrop is that Arm shares had been trading at a "priced-for-perfection" level. Arm stock had roughly doubled year-to-date going into 2026, and on May 6 had just surged 13.6% in response to a quarterly earnings report that was the best ever as a quarter (FY2026 Q4 revenue of $1.49 billion = approximately ¥237 billion, up 20% year-on-year, with data-center royalties doubling year-on-year). While several brokerages raised their target prices, the average analyst target was below the current price, and there were voices warning that the stock had "risen too far." When a new factor—regulatory risk—collides with a share price that has already priced in perfection, a sharp correction is half-inevitable. It is reasonable to read this as a phase in which the market is not viewing the foundation of the business as broken, but is beginning to layer on a commensurate "regulatory risk premium."

What could ripple out more seriously is the parent company SoftBank Group (TSE: 9984). SoftBank Group holds roughly 90% of Arm shares, with floating shares circulating in the market amounting to only around 10%. This thin float itself is significant: the scarcity of liquidity makes the share price prone to swings, and that market valuation in turn governs the bulk of SoftBank Group's NAV. Arm is, on SoftBank Group's books, the single largest asset on its own, and the value of that stake corresponds to about 90% of Arm's market capitalization (on the order of $220–230 billion in mid-May)—that is, roughly $200 billion (on the order of ¥32 trillion). SoftBank Group also carries borrowings collateralized by assets including Arm shares, and if Arm shares are persistently re-rated downward (de-rated) on regulatory grounds, pressure will mount on both NAV and financial strategy. The Nikkei has repeatedly pointed out the structure in which SoftBank Group shares move almost in lockstep with Arm shares, and a decline in Arm shares dampens expectations for an improvement in SoftBank Group's investment returns.

That said, SoftBank Group itself is currently on strong footing. The company announced its FY March 2026 results on May 13, with the Vision Fund business posting profits of approximately $46 billion (approximately ¥7.3 trillion) for the year, centered on valuation gains on its OpenAI investment, and full-year net income reaching the ¥5 trillion range. In its earnings, SoftBank Group established a new "AI Computing Segment" bundling Arm, Graphcore, and Ampere, effectively demonstrating itself just how central Arm is to the group's strategy. What matters is that this news broke on Friday night U.S. time. The Tokyo market, where SoftBank Group shares trade, is closed over the weekend, so the first real trading opportunity to price this in will be the sessions from Monday, May 18—the date of this article—onward. Note that as of the time of writing, we have not been able to confirm via primary sources "the specific percentage decline in SoftBank Group shares following the FTC report," and here we limit ourselves to describing the linkage mechanism and historical tendencies. From a VC perspective, the transmission pathway to SoftBank Group is clear—Arm's concentration within NAV, the thin float, and leverage. Through these three points, Arm's regulatory risk can be transmitted, amplified, to SoftBank Group's valuation.

Future Timeline—When and What Will Be Observed

Finally, we look ahead to what new developments may be observed, and when, in line with the FTC's investigative procedures. As a premise, antitrust investigations that question conduct (monopolization) take time. Because there is no statutory review deadline as there is with merger review, it is not unusual for the FTC to take 12 to 24 months or more to reach any conclusion before deciding "whether or not to file suit." The process generally proceeds as follows: the non-public commencement of an investigation and document preservation requests (where Arm currently stands), the collection of materials and testimony via CIDs or subpoenas, a recommendation from the FTC staff to the Commission, and then the Commission's decision. There are multiple exits, any of which is possible: closing the investigation without action, a consent order in which the company promises to remediate its conduct, a filing for an internal FTC administrative adjudication, or a complaint filed in federal district court.

The near-term milestones that investors and VCs should mark on their calendars are clear. First, whether the FTC goes beyond document preservation requests and issues formal CIDs or subpoenas to Arm and to third parties (other licensees including Qualcomm). If this happens, the seriousness of the investigation will move up a notch. Second, Arm's Q1 FY March 2027 earnings (the April–June 2026 quarter) are scheduled roughly from late July to early August. How this investigation is reflected in the earnings materials and in the risk factor disclosures in statutory filings, and how management comments on it, will be the first litmus test. SoftBank Group's Q1 earnings will follow around August. Third, there are developments outside the United States, namely whether the European Commission, having received Qualcomm's complaint, will open a formal investigation under Article 102 of the TFEU in Europe, and how the Korean KFTC will move toward its review findings following its on-site inspection. Fourth, the litigation between Arm and Qualcomm will proceed in parallel through the appeals proceedings and Qualcomm's counterclaim. In addition, the volume production ramp of the AGI CPU (scheduled for the second half of 2026) and the acquisition of new customers, as well as whether Arm will revise the terms of its standard license agreements, will also serve as material for divining the direction of the investigation.

From the VC perspective, we would like to define "the moment the investment thesis changes." If the FTC moves to file a formal complaint, or if the matter is settled with a consent order accompanied by behavioral remedies that mandate non-discriminatory licensing, that will be a genuine turning point for Arm's business model. Conversely, if the investigation quietly concludes without any noteworthy measures, Arm's stock will likely be re-rated as the regulatory risk premium is peeled away. In the end, what this investigation poses to Silicon Valley VCs is a structural question that goes beyond the legal troubles of a single semiconductor company: should a company be allowed to own the "rules" of an ecosystem while simultaneously competing "inside" that ecosystem? The answer to this question will, beyond the stock prices of Arm and SoftBank Group, define how the IP that forms the foundation of computing over the next decade will be designed, funded, and where it will compete. The FTC's single document preservation request is merely the entryway to that larger question.

Sources