💡 TL;DR - The 30 Seconds Version
👉 Arm Holdings shares dropped 8% to $149.50 after announcing plans to build its own chips instead of just licensing designs to others.
📊 The company reported Q1 revenue of $1.05 billion (up 12%) but forecast earnings of 29-37 cents per share, missing the 36-cent analyst consensus.
🏭 Arm will develop chiplets and complete solutions for AI infrastructure, including the $500 billion Stargate project with OpenAI and SoftBank.
⚔️ The strategy puts Arm in direct competition with major customers like Nvidia and Amazon, who currently license its designs to build their own chips.
📈 Arm trades at over 80 times expected earnings compared to Nvidia's 35 times, raising questions about whether the premium valuation makes sense.
🚀 The pivot represents a bet that Arm can capture more value from the AI boom by moving beyond licensing into the expensive, crowded chip market.
Arm Holdings shares fell 8% Wednesday after the British chip designer announced something that changes everything. The Cambridge company doesn't want to just license technology anymore. It wants to build the chips itself.
CEO Rene Haas told investors Arm is "consciously deciding to invest more heavily" in making its own chipsets, chiplets, and complete solutions. That's a big departure from how Arm has operated for decades — selling intellectual property to companies like Nvidia, Amazon, and Apple.
Investors weren't having it. Arm's stock dropped to $149.50 after hours, losing $13.82 per share. The disappointing forecast didn't help either. Arm expects to earn 29 to 37 cents per share this quarter, missing the 36-cent consensus.
Strategic Gamble With High Stakes
Arm's move makes sense if you think about it. The company has watched customers like Nvidia build massive businesses using Arm's designs while paying relatively modest licensing fees. Now Arm wants more of that money.
Haas said the company would explore "chiplets or even possible solutions" — basically complete hardware and software packages that customers can use right away. No assembly required.
The timing matters here. Arm sits at the center of SoftBank's $500 billion Stargate project with OpenAI and Oracle to build AI data centers across America. Haas confirmed that chip design work for those facilities has started, though he wouldn't share details about what they're building or when we'll see it.
Numbers Tell a Mixed Story
Arm's first-quarter results show why this pivot feels necessary. Revenue climbed 12% to $1.05 billion, just missing the $1.06 billion estimate. The growth came with some asterisks though.
Royalty revenue jumped 25% to $585 million as more AI chips shipped. That's the bright spot. But licensing revenue dropped 1% to $468 million, suggesting fewer companies are signing up for new Arm designs.
Net income fell 42% to $130 million. Some of that reflects the heavier R&D spending Haas keeps talking about. He calls it necessary investment. Investors see profit getting squeezed.
Market Headwinds Don't Help
Arm faces problems beyond its strategic shift. Global smartphone shipments grew just 1% last quarter, according to International Data Corporation. That's bad news for a company still dependent on mobile chip royalties.
Trade tensions make forecasting harder too. Haas said Arm hasn't seen customers rushing orders to beat potential tariffs, but the uncertainty clouds everything. The company's revenue guidance of $1.01 billion to $1.11 billion next quarter reflects this caution.
China creates particular headaches. Analysts noted weakness in Android smartphone royalties from Chinese manufacturers — historically a growth engine for Arm.
Customer Relationships Get Complicated
Here's where things get messy. Arm's plan to build complete chips puts it head-to-head with some major customers. Companies like Nvidia and Amazon currently license Arm's designs to build their own processors. If Arm starts selling finished chips, those partnerships could turn awkward fast.
J.P. Morgan analysts led by Harlan Sur called this out directly. "We are increasingly concerned with its strategy to develop full chip solutions," they wrote, pointing to obvious conflicts of interest.
The semiconductor world has seen this before. When suppliers become competitors, things can get ugly. Arm needs to handle these relationships carefully or risk losing the customers that still pay most of its bills.
Breaking Into an Expensive, Crowded Market
Making advanced AI chips costs serious money. Developing cutting-edge processors can run $500 million just for the silicon, before you add server hardware and software. That's a lot of cash for uncertain returns.
The competition is brutal too. Nvidia owns AI data center chips. Amazon, Microsoft, and others built their own custom processors. AMD is fighting hard for market share. Well-funded startups launch regularly.
Haas sounds confident about Arm's chances. He pointed to the company's reach from milliwatt smartphone chips to megawatt data center processors. "This wide spectrum equips Arm to deliver scalable and efficient solutions unmatched by competitors," he said during the call.
Maybe. But spanning markets and winning them are different things.
Stock Price Raises Questions
Arm has jumped 150% since going public in 2023. The stock trades at over 80 times expected earnings — more than double Nvidia's 35 times multiple and way above traditional chip companies.
That premium assumes Arm will grab more AI infrastructure spending. But Wednesday's disappointing forecast and strategic uncertainty make the valuation look stretched.
Still, at least two brokerages raised price targets after earnings. The median target now sits at $155, suggesting modest upside from current levels. Some analysts clearly still believe in the story.
Why this matters:
• Arm is risking its profitable licensing business on an expensive, crowded chip market — a bet that could reshape semiconductor power or blow up spectacularly.
• The company risks losing major customers while chasing AI money, creating the classic innovator's dilemma where success in one area kills the core business.
❓ Frequently Asked Questions
Q: What exactly are chiplets and why does Arm want to make them?
A: Chiplets are smaller, specialized chips that work together like Lego blocks to form complete processors. Instead of one massive chip, companies combine chiplets for different functions. This approach costs less and lets companies customize processors more easily.
Q: How much money does Arm actually make from licensing versus royalties?
A: In Q1, Arm earned $468 million from licensing (down 1%) and $585 million from royalties (up 25%). Licensing provides upfront payments when companies use Arm's designs. Royalties generate ongoing revenue from every chip sold using those designs.
Q: Why is Arm's stock price so much higher than other chip companies?
A: Arm trades at 80+ times expected earnings versus Nvidia's 35 times and AMD's 35 times. Investors pay this premium because Arm's designs power 99% of smartphones and the company could capture more AI infrastructure spending as demand grows.
Q: What is the Stargate project that Arm mentioned?
A: Stargate is a $500 billion plan by SoftBank, OpenAI, and Oracle to build AI data centers across America over four years. Arm will design custom chips for these facilities, potentially generating significant new revenue beyond traditional licensing.
Q: How dominant is Arm in the smartphone market?
A: Arm holds a 99% market share in smartphone processors. Apple, Samsung, Qualcomm, and virtually every other phone chip maker uses Arm's designs. This dominance generates steady royalty payments but limits growth as smartphone sales have plateaued globally.
Q: Does SoftBank's ownership influence Arm's strategy to build chips?
A: Yes. SoftBank CEO Masayoshi Son wants Arm central to his AI investments. SoftBank owns Arm and leads the Stargate project, creating natural synergies. Son sees chip development as key to capturing more AI infrastructure value.
Q: Who are Arm's biggest customers that might become competitors?
A: Nvidia uses Arm designs in AI data center chips, Apple builds Arm-based processors for iPhones and Macs, Amazon creates custom Arm chips for AWS servers, and Qualcomm makes Arm-based smartphone processors. All could compete with Arm's new chip business.
Q: When might we see Arm's own chips hit the market?
A: Arm hasn't provided a timeline. CEO Rene Haas said chip design work for Stargate facilities has started but declined to share launch dates. Given typical chip development cycles of 2-3 years, expect announcements in 2026-2027 at earliest.