A
Arm Holdings plc American Depositary Shares
ARM · Technology
Q3 FY2026
Earning ended Dec 31, 2025Report generated Mar 15, 2026
The World's Greatest Toll Road, Priced for Eternal Perfection
The market has already awarded ARM the full value of the CSS royalty ramp and the AI data center pivot — but those tailwinds haven't yet materialized in actual cash flows, meaning investors are paying peak-cycle multiples for a story that is still being written, not one that has already delivered.
Flowly Score
Mar 27 Close
-85%$144.13
Fair Value
$22.00
46
Below Averageat $144.13
Business Strength8/10
The ISA moat is genuinely rare — three decades of accumulated ecosystem lock-in means even trillion-dollar hyperscalers can't walk away — but the SoftBank governance overhang and the Arm China joint venture's documented control failures prevent a top score; minority shareholders are passengers, not co-pilots.
Financial Resilience6/10
The balance sheet is clean and the IP model requires almost no physical capital in steady state, but the recent quarter's pattern — revenue surging while free cash flow collapsed nearly in half — is precisely the wrong direction for a business that should consistently over-convert earnings to cash; until that reversal explains itself, resilience gets a haircut.
Growth & Trajectory8/10
The CSS licensing strategy is a genuine value-chain ascent — not incremental improvement but a structural repricing of what ARM charges per chip — and the data center royalty doubling year-over-year with hyperscalers packing triple-digit core counts into custom silicon is the clearest multi-year compounding story in semiconductors right now.
Valuation2/10at $115.75
At over 600x trailing free cash flow, the current price isn't pricing in a bull case — it's pricing in a bull case that has already fully executed flawlessly, leaving zero margin for execution risk, royalty ramp timing delays, or a single bad macro quarter; the gap between intrinsic value and market price is not a rounding error.
Risk3/10
The convergence of Chinese state-backed RISC-V substitution, an Arm China JV that ARM doesn't fully govern, a 90% controlling shareholder with its own financial agenda, and a valuation that requires perfect execution creates a rare layering of moat risk, governance risk, and price risk simultaneously — any one of these alone would warrant caution; together they demand it.
ARM is a genuinely exceptional business — perhaps the purest architectural moat in technology — but exceptional businesses and exceptional investments are not the same thing, and right now the price assumes a flawless ten-year execution of a thesis that hasn't yet shown up in the cash register. The CSS strategy is real, the AI inference tailwind at the edge is real, the data center royalty ramp is real — but at these multiples, every one of those things needs to arrive on schedule, at the expected royalty rates, without China disruption, without RISC-V inroads, and without SoftBank making a move that disadvantages the public float. That's not investing — it's a parlay. Where this business is going is actually compelling: the shift from licensing discrete CPU cores to bundling entire compute subsystems is a quiet repricing of ARM's value proposition, transforming it from a component supplier to a platform — and platforms command structurally higher economics. As hyperscalers race to differentiate their AI infrastructure through custom silicon, they are pulling ARM's highest-royalty-rate products into the fastest-growing segment of the chip market. The v9 architecture upgrade cycle creates a second compounding lever — an installed base of licensees being nudged toward a higher royalty tier, independently of unit volume growth. These are real inflection points, not hype. The single biggest specific risk is the Arm China structure colliding with accelerating geopolitical decoupling. China is a meaningful revenue source flowing through a joint venture that ARM demonstrably cannot fully govern — proven not in theory but in practice during the years-long CEO standoff. Beijing has both the policy intent and the capital to subsidize RISC-V adoption as a strategic alternative to Western ISA dependence. If Chinese chip volume migrates to RISC-V at scale, ARM loses the royalties, loses the network reinforcement from high-volume designs, and loses the credibility of claiming universal architecture dominance — three simultaneous blows to the thesis that is already fully priced in.