⚡ Earnings Preview Pre-Print ARM · NASDAQ Semiconductors / IP Licensing Published Apr 28, 2026 · Q4 FY26 Print: May 6

Arm Holdings
Q4 FY26 Earnings Preview

Arm is the invisible toll booth on every computing device on Earth — 30 billion+ chips shipped per year, royalties on every one. The Q4 FY26 print on May 6 is the first chance to hear FY27 guidance, CSS royalty rate progression, and whether the data center royalty displacement of smartphone is ahead of schedule. That's the entire call in three sentences.

Next Earnings
May 6, 2026
Current Price
~$125
Market Cap
~$130B
Non-GAAP Gross Margin
~97%
🏛️
01 / 07
Company Snapshot

Arm Holdings (NASDAQ: ARM) designs the processor architectures that run virtually every smartphone, tablet, and an accelerating share of AI data center infrastructure on Earth. The company does not manufacture chips — it licenses instruction set architectures (ISAs), CPU/GPU/NPU cores, and increasingly complete Compute Subsystems to semiconductor companies, who pay an upfront license fee and then a royalty on every chip they ship. That asset-light model produces a ~97% gross margin and positions Arm at the center of every major computing trend without taking manufacturing risk.

Founded in 1990 as a joint venture between Acorn Computers, Apple, and VLSI Technology in Cambridge, UK, Arm was acquired by SoftBank in 2016 for $32 billion. It relisted on NASDAQ on September 14, 2023 at $51/share — the largest IPO of that year — raising ~$4.87 billion, with SoftBank retaining approximately 90% ownership. CEO Rene Haas, who took over in 2022 from Simon Segars, has repositioned Arm's narrative from "smartphone royalty engine" to "compute platform of choice from milliwatts to megawatts."

Arm's revenue model has two components: licensing revenue (upfront fees from chip designers for access to IP), which is lumpy quarter-to-quarter but signals future royalty streams; and royalty revenue (per-unit fees on every chip shipped using Arm IP), which is recurring and scales with chip volumes and architecture mix. The key valuation driver is the architecture transition from Armv8 to Armv9 — v9 pays approximately 2x the royalty rate of v8, and as the v9 mix in shipping chips increases, royalty revenue grows without requiring more chip volumes.

Founded
1990
HQ
Cambridge, UK
CEO
Rene Haas
Employees
~7,700
Exchange
NASDAQ: ARM
FY26 Revenue (est.)
~$4.9B
Non-GAAP Gross Margin
~97%
SoftBank Ownership
~90%
Armv9 Architecture Armv8 (Legacy) Neoverse (Data Center) Cortex-A/X (Mobile) Compute Subsystems (CSS) Mali GPU Ethos NPU Total Access (ATA) Licensing
♟️
02 / 07
Strategic Moves

Arm's strategic agenda is concentrated on two structural transitions: the CSS royalty rate step-change and the Armv9 mix inflection. Both are happening simultaneously and both are systematically under-modeled by sell-side consensus.

Compute Subsystems (CSS) — the royalty rate step-change
2024–2026
CSS: pre-integrated chip blueprints that bundle CPU clusters, interconnects, memory interfaces, and system IP. CSS v2 royalties are over 10% of chip ASP, versus ~1-2% for a typical Armv8 CPU license. As of Q3 FY26, Arm had signed 21 CSS licenses across 12 companies, with 5 customers already shipping CSS-based chips. This is ahead of Arm's internal plan. CSS licensees span smartphones, PCs, data centers, and automotive.
Armv9 royalty rate uplift — the mix inflection math FinTwit ignores
2024–2026
Armv9 chips pay roughly 2x the royalty rate of Armv8. As of Q3 FY26, v9 represented approximately 30% of royalty revenue. If v9 goes from 30% to 50% of the royalty base (all else equal), royalty revenue grows ~13% from architecture mix alone — with zero volume growth. Qualcomm's Snapdragon 8 Elite, Apple's A18, AWS Graviton4, Google Axion, and Microsoft Azure Cobalt are all v9. The royalty rate uplift from v9 penetration is a structural tailwind systematically under-modeled by sell-side consensus.
Data center — 100%+ YoY royalty growth, smartphone displacement ahead
Q3 FY26
In Q3 FY26, Arm's data center royalty revenue grew more than 100% year-over-year. Arm's CFO Jason Child explicitly projected data center royalty will surpass smartphone royalty within 2–3 years. AWS Graviton4, Google Axion, Microsoft Cobalt 100, NVIDIA Grace, and Meta custom Arm-based CPUs all pay Arm royalties. Arm-based cloud CPU market share approaching ~50% of new hyperscaler server deployments.
No FY26 full-year guidance — a deliberate macro hedge
February 4, 2026
Arm was the only major tech company to explicitly decline to give FY26 full-year guidance, citing macroeconomic uncertainty (tariffs). This kept expectations anchored. The Q4 print will be the first opportunity to hear management frame FY27 expectations. The market will treat any quantified FY27 commentary as the catalyst — this single data point will drive the stock reaction more than the Q4 actuals themselves.
📊
03 / 07
Financial Trajectory

Arm's revenue trajectory reflects two dynamics: licensing revenue is episodic and lumpy (deal timing drives quarter-to-quarter swings), while royalty revenue is structurally growing as v9 mix increases and data center ramps. The last four quarters:

QuarterRevenueRoyaltyLicenseNon-GAAP EPSYoY Growth
Q1 FY26 (Jun 2025)$1.05B~$590M~$472M$0.35+12%
Q2 FY26 (Sep 2025)$1.14B$620M$515M$0.39+34%
Q3 FY26 (Dec 2025)$1.24B$737M$505M$0.43+26%
Q4 FY26 Guide (Mar 2026)$1.47B ±$50Mup low-teens YoYup high-teens YoY$0.58 ±$0.04~+18%

Royalty Revenue as the quality signal: Royalty revenue is the recurring engine — $737M in Q3, up 27% YoY with data center growing >100%. The Q4 guide calls for royalties up "low teens" YoY — approximately $650–670M. Sequential decline Q3→Q4 is normal seasonality. Watch whether Q4 royalty beats or misses the implied ~$655M consensus.

License Revenue lumpiness: Q3 license came in at $505M vs. $519M consensus — the stock dropped 7%+ despite beating on total revenue. ACV (Annualized Contract Value) grew 28% YoY in Q3. Q4 license guidance implies ~$815-830M (high teens YoY on the Q4 FY25 base of ~$700M). Any miss on license will likely trigger a similar stock reaction even with a headline revenue beat.

FY26 full year estimate: Q1+Q2+Q3 actual ($3.43B) + Q4 guided midpoint $1.47B = ~$4.90B FY26 revenue (vs. $4.01B FY25 = +22% YoY). Non-GAAP OpEx guided to ~$745M in Q4 — $2.9B+ annualized. TTM Free Cash Flow: $893M as of Q3 FY26.

⚔️
04 / 07
Competitive Position

Arm competes in the invisible layer: it doesn't build chips, it provides the blueprints. This makes its competitive position simultaneously more durable (no fab risk, no design execution risk) and more vulnerable (a superior ISA alternative would threaten the entire royalty stream, not just one product line).

RISC-V 🔴
The structural threat FinTwit dismisses too easily. RISC-V is an open-source ISA that charges no licensing fee. Already taken meaningful share in embedded microcontrollers and IoT. As of 2026: RISC-V is not there yet for hyperscale data center CPU workloads. But ahead of where most investors expect in AI accelerators — Tenstorrent and other AI chip startups are building RISC-V-based AI chips that don't pay Arm royalties. The 5-year risk is real; the 2-year risk is manageable.
x86 (Intel/AMD) 🟢
Arm winning, not threatened. Intel and AMD are losing cloud server CPU share to Arm-based alternatives. AWS Graviton, Google Axion, Microsoft Cobalt are all in production. This is not a threat to Arm — it's a tailwind. Royalty rates on Neoverse-based cloud chips are higher than on mobile Cortex chips, volumes growing, revenue mix shift accelerates. Every new hyperscaler Arm-based CPU deployment is a royalty win at higher per-chip rates than smartphones.
NVIDIA Grace 🟡
Partner AND competitor. NVIDIA's Grace CPU (Hopper and Blackwell superchips) is built on Arm's Neoverse N2 — NVIDIA pays Arm royalties. This makes NVIDIA simultaneously Arm's largest customer and a company with long-term incentives to reduce Arm dependency. In the near term (2-3 years), Grace is an Arm royalty win. In the 5+ year horizon, NVIDIA is a potential defection risk.
Qualcomm (litigation) 🔴
The existential binary risk. In October 2022, Arm gave Qualcomm a 60-day notice to cancel its license regarding Nuvia IP. Qualcomm argues its master license covers Nuvia's designs; Arm argues the Nuvia CPU requires a new license with higher royalties. This dispute has gone to trial. Qualcomm represents ~$500M+ of Arm's annual royalty revenue. A ruling in Arm's favor would significantly increase Qualcomm's royalty obligations. A ruling in Qualcomm's favor could depress royalty rate expectations systemically.
Bottom line: Arm's competitive position is unique — it's not fighting for design wins against other chip companies. It's fighting to remain the default ISA as computing architectures evolve. The near-term view (2-3 years) is favorable: data center royalties accelerating, v9 mix inflecting, CSS commanding 10%+ ASP royalty rates. The long-term view (5+ years) requires watching RISC-V data center adoption and whether hyperscalers with custom silicon ambitions reduce their Arm-based CPU deployments.
⚠️
05 / 07
Key Risks
📬 Post-Print Update — ARM prints May 6 AMC

Get the post-print update. We'll email you within 24h of the print with actuals, thesis verdict, and an updated bull/bear range.

Arm's risk profile is unusual: not execution risk (it doesn't build anything) and not market risk (compute demand only grows). It's structural risk — the risk that the computing industry migrates to ISAs that don't pay Arm royalties. Near-term, the risks are valuation and guidance disappointment.

Bull Case
$175–200 price target conditions
FY27 revenue guide ≥$6.5B (>30% YoY) on Q4 call · CSS licenses reach 30+ and first CSS chip revenue >$200M in FY27 guide · v9 royalty mix reaches 45%+ by mid-FY27 · Qualcomm dispute settled favorably · Data center royalty surpasses smartphone in FY27 · Stock re-rates toward 40x revenue → $200+
Bear Case
$80–95 price target conditions
FY27 guidance ≤$5.5B → implies deceleration to ~12% growth, multiple compression toward 20x revenue · Qualcomm dispute resolved in Qualcomm's favor · RISC-V data center adoption accelerates with hyperscaler announcement · License revenue continues missing consensus · China revenue declines >30%
High
SoftBank concentration overhang
SoftBank owns ~90% of ARM — ~950M shares worth ~$119B. Every 1% SoftBank sale = ~10M shares (~$1.25B) potential supply. SoftBank's Vision Fund losses and leverage create secondary offering risk. This overhang suppresses valuation multiples and will continue until SoftBank's stake normalizes below 50%.
High
Valuation math at 160x+ GAAP P/E
ARM trades at one of the highest multiples in the semiconductor sector. Any guidance miss or growth deceleration triggers disproportionate multiple compression. The stock has dropped 7-14% after every recent earnings print despite beating on revenue. At ~$130B market cap on ~$4.9B revenue, the stock prices in sustained 25%+ royalty growth for a decade.
Medium
License revenue timing risk
A single deal pushed from one quarter to the next can cause a license revenue miss even when underlying demand is strong. Q3 FY26 demonstrated this: $505M vs. $519M consensus caused ~7% stock decline. Q4's "high teens YoY" license guide creates a ~$820-830M expectation. If a large ATA agreement slips, a miss is likely even if royalty revenue is healthy.
Medium
China revenue exposure (~22% of revenue)
US export control expansion targeting Arm IP would directly impact royalties from Chinese chip shipments. Any commentary on China revenue trajectory on the Q4 call will be closely watched. The geopolitical trajectory — US-China chip restrictions tightening — is a structural headwind that will not reverse.
🚀
06 / 07
What FinTwit Is Missing

The sell-side consensus on ARM focuses on unit volumes and total revenue. It systematically misses the rate story — how CSS and v9 mix change the per-chip royalty Arm earns even when volumes are flat. These four observations are where most bull cases are underestimated.

The CSS royalty rate math is not in sell-side models
Structural — Q3 FY26 onward
CSS v2 commands royalties above 10% of chip ASP. A high-end server chip with a $500 ASP means a ~$50 royalty to Arm per chip — versus ~$3-5 for a typical mobile Cortex license. Arm has 21 CSS licenses and 5 customers shipping. The street models CSS as a "volume story" but not as a "rate story" (dramatically higher per-chip royalty). When the 21 CSS licensees reach production at scale, the royalty rate mix shifts meaningfully. This is the unlabeled variable in every sell-side DCF on ARM.
ACV is a 12–18 month leading indicator nobody tracks properly
Structural — Q3 FY26 baseline
Arm's ACV (Annualized Contract Value) grew 28% YoY in Q3 FY26. ACV captures the forward commitment of long-term licensing agreements before they show up in revenue. Most sell-side models don't include ACV in their royalty forecasts because the lag between license signing and chip production is 12–24 months. The ACV trend right now is leading a royalty acceleration in FY28–FY29 that isn't in most 2-year price targets.
The v9 mix inflection creates a non-linear royalty trajectory
Structural — ongoing
The math: assume royalty base of $2.9B (FY26 run rate). If v9 is 30% of royalty rev at 2x rate, the blended rate premium is ~0.43x on top of v8 baseline. As v9 goes from 30% to 50%, the blended rate premium goes from 0.43x to 0.67x — a 55% increase in rate premium, adding ~$700M+ to annual royalty run rate with zero volume growth. The market models "v9 is growing" without doing this specific rate-uplift math. It's where most bull cases are underestimated.
Data center royalty ≠ AI accelerator royalty
Structural — widely misunderstood
Arm collects royalties on data center CPUs (Neoverse). It does NOT collect royalties on NVIDIA GPUs, Google TPUs, or AWS Trainium — those don't use Arm CPU cores. But: AI inference at the edge, AI on device (every iPhone, Android, laptop), and the CPU orchestration layer in data centers ALL use Arm. The "ARM benefits from AI" thesis is accurate but the mechanism is different from what most retail investors assume. Arm gets paid per CPU in every device that touches AI, not per GPU.
📋
07 / 07
Q4 FY26 Pre-Earnings Preview
⚡ PRE-PRINT — Arm reports Q4 FY26 (quarter ended March 31, 2026) on May 6, 2026 after market close. Q4 guidance issued on the Q3 FY26 call (February 4, 2026): Revenue $1.47B ±$50M, Non-GAAP EPS $0.58 ±$0.04.
MetricCompany GuidanceStreet ConsensusActual (Post-Print)
Total Revenue$1.47B ±$50M~$1.47BTBD
Non-GAAP EPS$0.58 ±$0.04~$0.57–0.59TBD
Royalty Revenueup low-teens YoY~$650–670MTBD
License Revenueup high-teens YoY~$815–830MTBD
FY26 Full Year Revenue— (no guide given)~$4.87–4.95BTBD
FY27 Revenue Guide—~$5.8–6.0B est.TBD
// Key watch items on the May 6 call
  • FY27 guidance: the #1 market mover. Arm declined to give FY26 full-year guidance in February citing macro uncertainty. Street consensus models ~$5.8-6B (20%+ growth). If Haas provides explicit FY27 guidance at or above that level, the stock re-rates immediately. If he remains non-committal or guides below $5.5B, expect a significant selloff.
  • License revenue: beat vs. miss vs. Q3 pattern. Q3 license $505M vs. $519M consensus triggered a 7% drop. Q4 implies ~$815-830M (high teens YoY). Any miss will be treated as a leading indicator problem. Watch for large ATA agreements slipping from Q4 into Q1 FY27.
  • v9 royalty mix update. Q3 confirmed v9 at ~30% of royalty revenue. Any update showing acceleration toward 35%+ is a structural tailwind signal.
  • CSS license count and first revenue disclosure. 21 CSS licenses at Q3, 5 customers shipping. Any increase in license count or CSS royalty quantification validates the rate uplift thesis.
  • Data center royalty as % of total. Q3: >100% YoY growth. Management guided surpassing smartphone in 2-3 years. Any acceleration of this timeline is meaningful.
  • Qualcomm dispute update. Any settlement commentary would be significant. A settlement announcement is a major positive catalyst.
  • China revenue commentary. ~22% of revenue in Q2. US-China trade tensions are a structural headwind worth tracking on every call.

The single most important number on this call: the combination of total royalty guidance for Q1 FY27 and any quantification of FY27 framework revenue. If Haas says "we expect FY27 revenue of $6B+" with specific drivers, the stock re-rates immediately. If he hedges, expect disappointment regardless of Q4 numbers.

📬 Post-Print Update — ARM prints May 6 AMC

Get the post-print update. We'll email you within 24h of the print with actuals, thesis verdict, and an updated bull/bear range.

Data current as of April 28, 2026. Financial figures sourced from Arm Holdings Q3 FY26 earnings release (February 2026), Q2 FY26 earnings release (November 2025), Q1 FY26 earnings release (August 2025), and Arm Investor Relations. Q4 FY26 guidance as provided by management on the February 4, 2026 earnings call. Consensus estimates aggregated from sell-side research. Section 7 actuals to be populated following the May 6, 2026 earnings call.
// Also in semiconductors & AI infrastructure
// Get the full brief for any ticker

Every print. Every quarter.
Zero noise.

Vektor delivers the full 7-section brief — bull/bear ranges, consensus tables, call watch items — for any ticker in 40 seconds. $29/month.

Get Access — $29/mo — or —
You're on the list. We'll email you before the next print.

No spam. Unsubscribe anytime.