Arm Holdings plc American Depositary Shares
ARM remains a real AI and CPU IP winner, but at $309.48 the stock discounts far more than the company has actually proven, with valuation at 284.13x P/E and 56.98x sales despite supply-constrained AGI revenue and only ~20% near-term growth guidance.
ARM is expensive enough that being right on the business is not enough. The company is clearly winning in the right places. Q4 FY26 revenue was $1.49B, up 20%, licensing rose 29%, data center royalties more than doubled, and EBITDA margin hit 37.0%. Management also guided Q1 FY27 to $1.26B ± $50M of revenue and $0.40 ± $0.04 of EPS, with FY27 royalty growth expected in the 20-ish percent range. Those are strong numbers. They do not support 284.13x earnings and 56.98x sales.
The core mispricing is that the market is capitalizing ARM on the FYE31 story today. Management's long-term target is ambitious, $15B of AGI CPU revenue plus $10B of IP revenue and more than $9 of EPS by FYE31. That can justify substantial value if execution and supply align. But the same call also made clear that realization is the bottleneck: AGI CPU demand is now >$2B across FY27-FY28, yet the company is still only carrying $1B in outlook because of constraints in memory, wafers, packaging, and test equipment. That gap matters. When the stock already values the company at $294.4B, deferred revenue is not a detail, it is the thesis.
The second issue is that current profitability is still mixed. Non-GAAP quarterly economics are improving sharply, but FY2026 reported operating margin was only 12.0%, down from 20.6% in the prior year. Investors are paying a premium usually reserved for businesses with cleaner earnings conversion and a shorter path from demand to recognized profit. ARM's balance sheet is excellent, with Debt/EBITDA 0.34 and a 5.25 current ratio, so solvency is not in question. The question is simply what price makes sense for a great asset whose earnings are still catching up to its narrative.
What the case depends on is straightforward: first, that AI and hyperscaler royalty growth stays close to or above the guided ~20% range, second, that AGI supply constraints begin to clear, and third, that margin expansion becomes durable enough to close the gap between strategic value and accounting earnings. The company may achieve all three. At $309.48, investors are already paying as if it has.
Data center royalties accelerate faster than expected, AGI supply opens sooner, and investors continue capitalizing FYE31 targets years ahead of delivery.
Execution stays solid and AI royalties grow, but valuation normalizes toward the engine's earnings-growth blend as the market demands proof beyond long-term targets.
AI enthusiasm cools, supply constraints persist, and the market compresses ARM toward still-rich software-like multiples on sub-25% growth.
Quant call (SELL) diverges from analyst consensus (Strong Buy: 12 buy / 4 hold / 1 sell) — resolve explicitly.
We suffer more often in imagination than in reality.
Forward earnings power is improving, but not fast enough to justify the current stock. Against FY2026 EPS of $0.85, the market is paying 284.13x trailing earnings while management's near-term guide is only $0.40 ± $0.04 for Q1 FY27 and roughly 20% royalty growth for the year. The moat is stronger today than it was two years ago because Armv9, CSS, and hyperscaler server adoption are broadening the royalty base, but over the next 3-5 years the question is not whether ARM wins AI, it is whether those wins arrive quickly enough and with enough margin capture to support a $294.4B equity value. AI is a tailwind, regulation looks secondary here, and the real gating factor is physical supply and monetization pace, not demand.
- Valuation is the core issue: 284.13x P/E and 56.98x P/S versus peer averages of 38.9x and 9.2x.
- Near-term fundamentals are good but not enough for the price: Q1 FY27 guide is $1.26B ± $50M revenue and $0.40 ± $0.04 EPS, with management calling for roughly 20% royalty growth in FY27.
- AI upside is real but partly deferred: management sees >$2B of AGI CPU demand across FY27-FY28, yet is holding the outlook at $1B because of supply constraints.
- Execution has improved materially, with 7 straight quarterly EPS beats since the -60.0% miss in 2024-11-06, including $0.45 vs $0.40 on 2026-07-29 and $0.60 vs $0.58 on 2026-05-06.
- Q4 FY26 was genuinely strong: revenue reached $1.49B, up 20%, licensing grew 29%, and EBITDA was $0.55B, a 37.0% margin, up 1,360 bps sequentially.
- Data center is scaling fast, with royalty revenue more than doubled YoY in Q4 and management expecting data center royalties to double again YoY in FY27.
- Balance sheet risk is low, with Debt/Equity 0.06, Debt/EBITDA 0.34, and a 5.25 current ratio, which gives ARM room to invest through volatility.
- The stock price already capitalizes an aggressive end-state, while current FY2026 earnings are only $0.85 per share, leaving ARM at 284.13x trailing earnings.
- Reported FY2026 operating margin was just 12.0%, down from 20.6% in the prior year, showing that GAAP earnings power is still well below what the valuation implies.
- AGI CPU demand is not the same as revenue: management has visibility to >$2B across FY27-FY28 but kept the outlook at $1B because of bottlenecks in memory, wafers, packaging, and test equipment.
- Mobile remains a large but weak end market, with management describing it as flattish to slightly negative, while licensing is still back-half weighted, which increases quarterly volatility.
| Metric | Value | Context |
|---|---|---|
| Current price | $309.48 | Reference price for the rating and scenario framework |
| Quant price target | $139.17 | -55.0% vs current, Low confidence, basis: earnings_growth_blend |
| FY2026 revenue | $4.9B | +14.3% YoY |
| FY2026 EPS | $0.85 | +165.5% YoY, but still leaves valuation extremely stretched |
| Operating margin | 12.0% | Down from 20.6% in the prior FY |
| Valuation | P/E 284.13x | P/S 56.98x | P/B 34.04x | Vs peer averages of 38.9x P/E and 9.2x P/S |
| Q4 FY26 revenue and EBITDA | $1.49B revenue | $0.55B EBITDA | EBITDA margin 37.0%, up 1,360 bps QoQ |
| Q1 FY27 guidance | $1.26B ± $50M revenue | $0.40 ± $0.04 EPS | Implies continued roughly 20% growth |
| Balance sheet | Debt/Equity 0.06 | Debt/EBITDA 0.34 | Current ratio 5.25 | Very low leverage and strong liquidity |
| Free cash flow | $0.7B | Capex $0.2B, no buybacks, no dividend |
| Price action | -31.6% vs 52-week high | +48.7% vs 200-day average | 1W +29.27%, 1M +23.91%, 3M -7.57% |
| Quality and risk | Quality 73/100 | Risk 42/100 | Moderate risk, with valuation risk sub-score at 22.5 |
The biggest risk to the bearish call is that ARM converts its AI roadmap faster than expected, because management is targeting $25B of combined AGI CPU and IP revenue by FYE31 and more than $9 in EPS, which would make today's price less extreme on long-duration earnings power.
The next key catalyst is the November 4, 2026 earnings report, especially whether royalty growth stays near the guided ~20% range and whether AGI supply constraints begin to ease.
Positioning is not washed out. Insiders own 86.5%, the top 3 institutions only 8.2%, and short interest is just 1.57% of shares, or 16.7M shares with 2 days-to-cover, down 8.3%, so there is limited fuel for a squeeze-driven rerating. Options positioning is described as positive gamma in ≤45DTE, which can dampen volatility rather than create forced upside, and the stock already trades 48.7% above its 200-day average after a 29.27% one-week move.
Q4 licensing grew 29% YoY, ACV grew 22% YoY, and management said Q1 FY27 should see roughly 20% YoY license growth, though revenue remains back-half weighted.
Q4 data center royalty revenue more than doubled YoY and management expects royalty growth in the 20-ish percent range for FY27, with data center royalties set to double again YoY.
Management sees >$2B of demand across FY27-FY28 but is only guiding $1B because supply is constrained; first production chip revenue is still expected in Q4 FY27.
No geographic revenue split was provided, but hyperscaler cloud and networking ramps are the main growth vectors in the current cycle.
Record Q4 and Q1 guide imply continued ~20% growth while AGI CPU demand doubled.
Management's record since the 2025 Q1 call is mostly good, with one important nuance. In that 2025 Q1 call, the company reiterated FY25 revenue at $3.8B-$4.1B, trimmed royalty growth to the low 20% range from mid-20%, and guided Q2 FY25 revenue to $780M-$830M, calling it the low point of the year. Since then, ARM has delivered a strong beat pattern, 7 straight quarterly EPS beats after the 2024-11-06 miss, and FY2026 revenue reached $4.9B, up 14.3%, while Q4 FY26 revenue hit a record $1.49B and beat the guidance midpoint. The strategy has not changed materially, it has broadened: the company is still monetizing Armv9 and cloud royalties, but now the narrative is much more centered on AGI CPU and hyperscaler demand. Credibility is good on execution and product traction, but less complete on timing because the company now has >$2B of AGI demand across FY27-FY28 while holding the outlook at $1B due to supply, so the commercial opportunity is real but not fully under management's control.
For Q1, we expect revenue of $1.26 billion, plus or minus $50 million. We expect our non-GAAP EPS to be $0.40 plus or minus $0.04.
We expect that by FYE31, we will be generating $15 billion in AGI CPU revenue and $10 billion in IP revenue for a total of $25 billion. We expect this to translate to more than $9 in EPS.
Arm delivered a record quarter and record fiscal year. Revenue this quarter was $1.49 billion, up 20%, our highest quarterly revenue quarter ever and above the midpoint of our guidance.
We now have line of sight to more than $2 billion of demand across fiscal '27 and '28. However, we are maintaining our outlook of $1 billion while we pursue supply chain capacity.
I expect that the royalty growth is going to be in the roughly 20-ish percent for the year. And so across each of the quarters, plus or minus depending on the quarter, but will all be relatively close to the 20s percent range.
ARM's moat is its CPU instruction-set ecosystem, deep software compatibility, and entrenched position across mobile and increasingly cloud, automotive, and edge devices. The best evidence of durability is that royalties are now being lifted not just by units but by richer content, including Armv9 and CSS adoption, while hyperscaler custom silicon is broadening the customer set. The moat looks durable over the next several years, but monetization remains partly dependent on customer ramps and foundry supply, so moat strength does not fully eliminate cyclicality in realized revenue.
Capital allocation is conservative and sensible. ARM generated $0.7B of FCF on $4.9B of revenue, spent $0.2B on capex, and used $0.0B for buybacks and $0.0B for dividends, which is reasonable given the need to fund roadmap and ecosystem investment. With only $0.5B of debt and low leverage, the balance sheet is not the problem, but investors are not being paid back with capital returns at a valuation that would warrant them.
ARM rates SELL. This is not a call against the company, it is a call against the price. The business is executing, AI demand is real, and the moat is strengthening in data center and custom silicon, but investors are paying years in advance for revenue and earnings that management itself says are partly supply constrained. A more constructive view would require either a much lower entry price or clear evidence that AGI revenue, margins, and EPS are inflecting fast enough to make $309.48 look reasonable on nearer-term numbers.
≣ Full research note tap to expand
Executive Summary
SELL, Medium conviction. Rated SELL, Medium conviction. ARM is executing well operationally, with record results, repeated EPS beats, and clear data center royalty momentum, but the stock still embeds a heroic outcome that current guidance does not support. The setup is especially difficult with shares 48.7% above the 200-day average, still worth $294.4B after a 31.6% drawdown from the high, and a quant base value of $139.17.
Investment Thesis
ARM is expensive enough that being right on the business is not enough. The company is clearly winning in the right places. Q4 FY26 revenue was $1.49B, up 20%, licensing rose 29%, data center royalties more than doubled, and EBITDA margin hit 37.0%. Management also guided Q1 FY27 to $1.26B ± $50M of revenue and $0.40 ± $0.04 of EPS, with FY27 royalty growth expected in the 20-ish percent range. Those are strong numbers. They do not support 284.13x earnings and 56.98x sales.
The core mispricing is that the market is capitalizing ARM on the FYE31 story today. Management's long-term target is ambitious, $15B of AGI CPU revenue plus $10B of IP revenue and more than $9 of EPS by FYE31. That can justify substantial value if execution and supply align. But the same call also made clear that realization is the bottleneck: AGI CPU demand is now >$2B across FY27-FY28, yet the company is still only carrying $1B in outlook because of constraints in memory, wafers, packaging, and test equipment. That gap matters. When the stock already values the company at $294.4B, deferred revenue is not a detail, it is the thesis.
The second issue is that current profitability is still mixed. Non-GAAP quarterly economics are improving sharply, but FY2026 reported operating margin was only 12.0%, down from 20.6% in the prior year. Investors are paying a premium usually reserved for businesses with cleaner earnings conversion and a shorter path from demand to recognized profit. ARM's balance sheet is excellent, with Debt/EBITDA 0.34 and a 5.25 current ratio, so solvency is not in question. The question is simply what price makes sense for a great asset whose earnings are still catching up to its narrative.
What the case depends on is straightforward: first, that AI and hyperscaler royalty growth stays close to or above the guided ~20% range, second, that AGI supply constraints begin to clear, and third, that margin expansion becomes durable enough to close the gap between strategic value and accounting earnings. The company may achieve all three. At $309.48, investors are already paying as if it has.
Key Metrics
| Metric | Value | Context | ||
|---|---|---|---|---|
| Current price | $309.48 | Reference price for the rating and scenario framework | ||
| Quant price target | $139.17 | -55.0% vs current, Low confidence, basis: earnings_growth_blend | ||
| FY2026 revenue | $4.9B | +14.3% YoY | ||
| FY2026 EPS | $0.85 | +165.5% YoY, but still leaves valuation extremely stretched | ||
| Operating margin | 12.0% | Down from 20.6% in the prior FY | ||
| Valuation | P/E 284.13x | P/S 56.98x | P/B 34.04x | Vs peer averages of 38.9x P/E and 9.2x P/S |
| Q4 FY26 revenue and EBITDA | $1.49B revenue | $0.55B EBITDA | EBITDA margin 37.0%, up 1,360 bps QoQ | |
| Q1 FY27 guidance | $1.26B ± $50M revenue | $0.40 ± $0.04 EPS | Implies continued roughly 20% growth | |
| Balance sheet | Debt/Equity 0.06 | Debt/EBITDA 0.34 | Current ratio 5.25 | Very low leverage and strong liquidity |
| Free cash flow | $0.7B | Capex $0.2B, no buybacks, no dividend | ||
| Price action | -31.6% vs 52-week high | +48.7% vs 200-day average | 1W +29.27%, 1M +23.91%, 3M -7.57% | |
| Quality and risk | Quality 73/100 | Risk 42/100 | Moderate risk, with valuation risk sub-score at 22.5 |
Financial Strength
Financial risk is low. ARM carries only $0.5B of debt, leverage is just 0.34x Debt/EBITDA, and liquidity is strong with a 5.25 current ratio. Cash generation is positive, with $0.7B of FCF and modest $0.2B capex, which supports continued heavy investment in IP and ecosystem development. The issue is not balance sheet weakness. It is the mismatch between a sturdy but still maturing earnings profile and a market value of $294.4B.
Competitive Position
ARM's moat is its CPU instruction-set ecosystem, deep software compatibility, and entrenched position across mobile and increasingly cloud, automotive, and edge devices. The best evidence of durability is that royalties are now being lifted not just by units but by richer content, including Armv9 and CSS adoption, while hyperscaler custom silicon is broadening the customer set. The moat looks durable over the next several years, but monetization remains partly dependent on customer ramps and foundry supply, so moat strength does not fully eliminate cyclicality in realized revenue.
Management & Guidance
Current guidance is solid and mostly credible. CFO Jason Child said, "For Q1, we expect revenue of $1.26 billion, plus or minus $50 million. We expect our non-GAAP EPS to be $0.40 plus or minus $0.04." That follows a Q4 in which ARM exceeded its midpoint and beat EPS partly because OpEx was about $10M below guidance. Management credibility has improved because the company has now posted 7 straight quarterly EPS beats, but the AGI outlook still carries timing risk since the company acknowledged >$2B of FY27-FY28 demand while keeping the recognized outlook at $1B due to supply limits.
Segment Analysis
- Licensing and other revenue (Not disclosed in provided data): Q4 licensing grew 29% YoY, ACV grew 22% YoY, and management said Q1 FY27 should see roughly 20% YoY license growth, though revenue remains back-half weighted.
- Royalty revenue (Not disclosed in provided data): Q4 data center royalty revenue more than doubled YoY and management expects royalty growth in the 20-ish percent range for FY27, with data center royalties set to double again YoY.
- AGI CPU and long-duration AI programs (Early-stage contribution): Management sees >$2B of demand across FY27-FY28 but is only guiding $1B because supply is constrained; first production chip revenue is still expected in Q4 FY27.
- Geography (Not disclosed in provided data): No geographic revenue split was provided, but hyperscaler cloud and networking ramps are the main growth vectors in the current cycle.
Capital Allocation
Capital allocation is conservative and sensible. ARM generated $0.7B of FCF on $4.9B of revenue, spent $0.2B on capex, and used $0.0B for buybacks and $0.0B for dividends, which is reasonable given the need to fund roadmap and ecosystem investment. With only $0.5B of debt and low leverage, the balance sheet is not the problem, but investors are not being paid back with capital returns at a valuation that would warrant them.
Management Execution & Track Record
Management's record since the 2025 Q1 call is mostly good, with one important nuance. In that 2025 Q1 call, the company reiterated FY25 revenue at $3.8B-$4.1B, trimmed royalty growth to the low 20% range from mid-20%, and guided Q2 FY25 revenue to $780M-$830M, calling it the low point of the year. Since then, ARM has delivered a strong beat pattern, 7 straight quarterly EPS beats after the 2024-11-06 miss, and FY2026 revenue reached $4.9B, up 14.3%, while Q4 FY26 revenue hit a record $1.49B and beat the guidance midpoint. The strategy has not changed materially, it has broadened: the company is still monetizing Armv9 and cloud royalties, but now the narrative is much more centered on AGI CPU and hyperscaler demand. Credibility is good on execution and product traction, but less complete on timing because the company now has >$2B of AGI demand across FY27-FY28 while holding the outlook at $1B due to supply, so the commercial opportunity is real but not fully under management's control.
Positioning & Flows
Positioning is not washed out. Insiders own 86.5%, the top 3 institutions only 8.2%, and short interest is just 1.57% of shares, or 16.7M shares with 2 days-to-cover, down 8.3%, so there is limited fuel for a squeeze-driven rerating. Options positioning is described as positive gamma in ≤45DTE, which can dampen volatility rather than create forced upside, and the stock already trades 48.7% above its 200-day average after a 29.27% one-week move.
Bull Case
- Execution has improved materially, with 7 straight quarterly EPS beats since the -60.0% miss in 2024-11-06, including $0.45 vs $0.40 on 2026-07-29 and $0.60 vs $0.58 on 2026-05-06.
- Q4 FY26 was genuinely strong: revenue reached $1.49B, up 20%, licensing grew 29%, and EBITDA was $0.55B, a 37.0% margin, up 1,360 bps sequentially.
- Data center is scaling fast, with royalty revenue more than doubled YoY in Q4 and management expecting data center royalties to double again YoY in FY27.
- Balance sheet risk is low, with Debt/Equity 0.06, Debt/EBITDA 0.34, and a 5.25 current ratio, which gives ARM room to invest through volatility.
Bear Case
- The stock price already capitalizes an aggressive end-state, while current FY2026 earnings are only $0.85 per share, leaving ARM at 284.13x trailing earnings.
- Reported FY2026 operating margin was just 12.0%, down from 20.6% in the prior year, showing that GAAP earnings power is still well below what the valuation implies.
- AGI CPU demand is not the same as revenue: management has visibility to >$2B across FY27-FY28 but kept the outlook at $1B because of bottlenecks in memory, wafers, packaging, and test equipment.
- Mobile remains a large but weak end market, with management describing it as flattish to slightly negative, while licensing is still back-half weighted, which increases quarterly volatility.
Valuation & Price Target
Engine price target $139.17 (-55% vs current), Low confidence.
- P/E Multiple: $81.5 (7% weight)
- PEG (growth-adjusted): $27.74 (20% weight)
- P/S Multiple: $43.71 (6% weight)
- Quality-adjusted: $329.78 (20% weight)
- DCF: $7.58 (4% weight)
The rating aligns with the quant baseline SELL, but with higher respect for the business quality than the blended target alone suggests. The engine's $139.17 target is low confidence because peer dispersion is extremely wide at 4.6x, and ARM plainly deserves a premium on quality, ecosystem position, and AI relevance, which is why the quality-adjusted method reaches $329.78. Even so, the current stock still asks investors to underwrite a lot of FYE31 upside today, despite FY2026 EPS of only $0.85, a reported operating margin of 12.0%, and near-term guidance centered on roughly 20% growth rather than hypergrowth. Sell-side consensus remains Strong Buy at 12 buy / 4 hold / 1 sell, but that is low-signal context here because the disagreement is fundamentally about valuation timing, not whether ARM is a good company.
Risk Assessment
Risk score 42/100 (Moderate). The biggest risk to the bearish call is that ARM converts its AI roadmap faster than expected, because management is targeting $25B of combined AGI CPU and IP revenue by FYE31 and more than $9 in EPS, which would make today's price less extreme on long-duration earnings power.
The Bottom Line
ARM rates SELL. This is not a call against the company, it is a call against the price. The business is executing, AI demand is real, and the moat is strengthening in data center and custom silicon, but investors are paying years in advance for revenue and earnings that management itself says are partly supply constrained. A more constructive view would require either a much lower entry price or clear evidence that AGI revenue, margins, and EPS are inflecting fast enough to make $309.48 look reasonable on nearer-term numbers.
Go deeper in the app
Live charts, options analytics, the full earnings-call breakdown, price alerts, and every other stock — updated in real time.
Get Tickr