Back to Portfolio
ARM

ARM Holdings

$197 209B market cap April 15, 2026
Arm Holdings plc ARM BUFFETT / MUNGER / KLARMAN SUMMARY
1 SNAPSHOT
Price$197
Market Cap209B
2 BUSINESS

ARM Holdings is arguably the most important technology franchise created in the past 40 years -- its CPU architecture powers 99% of smartphones and is rapidly approaching 50% of hyperscaler data center deployments through NVIDIA Grace Blackwell, AWS Graviton, Google Axion, and Microsoft Cobalt. The royalty-plus-licensing model generates 95%+ gross margins with powerful recurring economics. CSS is doubling per-chip royalty rates and expanding ARM's value chain. However, at 262x trailing and 96x forward earnings, the stock prices in a decade of flawless execution with no margin of safety. SBC of $820M-$1B/year means true free cash flow is negative. RISC-V is a credible long-term threat, especially in China (24% of revenue). SoftBank's 90% ownership creates governance distortion and float scarcity that artificially inflates valuation. This is a wonderful business at a terrible price. WAIT for $105 or below, which would provide a 40-50x forward P/E entry on a franchise-quality IP licensor.

3 MOAT WIDE

Architecture standard (99% smartphone, 50% approaching data center), 22M+ developer ecosystem (80% global base), CSS value-add doubles royalty rates, 280B+ cumulative chips shipped, decades of IP development

4 MANAGEMENT
CEO: Rene Haas

Moderate - R&D investment is aggressive ($2B+/year) and appropriate for the opportunity, but SBC levels are excessive. No dividends. $202M buyback initiated but trivial vs $1B SBC.

5 ECONOMICS
20.6% Op Margin
9.5% ROIC
11.3% ROE
262x P/E
0.18B FCF
Net Cash Debt/EBITDA
6 VALUATION
FCF Yield0.09%
DCF Range110 - 130

Overvalued by 52-79%. Current price requires euphoria scenario (30% growth, 40% margins for 5+ years).

7 MUNGER INVERSION
Kill Event Severity P() E[Loss]
RISC-V competition closing performance gap by 2027, capturing IoT/embedded/China. 25% global processor market share and growing at 47% CAGR. HIGH - -
SoftBank 90% control creates governance risk, float scarcity, and strategic drift toward chiplets/ASICs. China 24% revenue at risk from geopolitics. MED - -
8 KLARMAN LENS
Downside Case

RISC-V competition closing performance gap by 2027, capturing IoT/embedded/China. 25% global processor market share and growing at 47% CAGR.

Why Market Right

RISC-V performance parity by 2027 could erode pricing power; China RISC-V mandate reducing TAM in second-largest market; SoftBank could use ARM as funding vehicle for AI ambitions; SBC dilution running at $1B/year (~0.5% of market cap annually)

Catalysts

NVIDIA Grace Blackwell ramp driving data center royalty surge (50% hyperscaler share); CSS adoption doubling royalty rates per chip, next-gen CSS delivering even higher rates; AI edge deployment wave (Apple, Samsung, MediaTek integrating AI acceleration on ARM); Automotive CSS license signed with global EV leader -- new market expansion

9 VERDICT WAIT
A- Quality Moderate - Debt-free with $2.8B cash, but FCF depressed by $820M+ SBC and rising CapEx. True FCF-after-SBC is negative.
Strong Buy$80
Buy$105
Fair Value$130

Monitor for entry at $105 (Accumulate) or $80 (Strong Buy). Set price alerts. The stock touched $80 in April 2025 and $100 in January 2026 -- opportunities recur given 3.34 beta.

🧠 ULTRATHINK Deep Philosophical Analysis

ARM Holdings -- Deep Philosophical Analysis

The Core Question: What Makes This Business Special?

There is a story Charlie Munger used to tell about the few businesses in history that achieved something almost magical: they became the invisible infrastructure of civilization. He would point to Coca-Cola -- not because the formula was so remarkable, but because the distribution system made it impossible to dislodge. ARM Holdings occupies an analogous position in computing, but at a scale Munger never analyzed.

Consider this: approximately 70% of the world's population interacts with an ARM-powered device every single day. Over 280 billion ARM-based chips have been manufactured since the company's founding. When you pick up your smartphone, when a Tesla navigates a highway, when AWS processes your cloud workload, when NVIDIA trains a large language model on its latest supercomputer -- ARM's architecture is the common thread. It is the instruction set that taught silicon how to think efficiently.

What makes ARM truly special is not just ubiquity but the nature of its economic model. ARM does not manufacture chips. It does not even design complete chips for most customers. It designs the architectural blueprints -- the instruction sets, CPU cores, and increasingly complete compute subsystems -- and then licenses this intellectual property to the companies that do the manufacturing: Apple, Qualcomm, Samsung, NVIDIA, Broadcom, MediaTek, and hundreds of others. For every chip shipped with ARM inside, ARM receives a royalty. This is not a one-time sale. This is a perpetual toll on the world's computing infrastructure.

The gross margins tell the story: 95-97%. There is virtually no cost of goods sold because the product is pure intellectual property. This is the economics of a law firm or consulting practice, not a semiconductor company. And yet the TAM grows every year as computing permeates more devices, more markets, more of life itself.

Moat Meditation: The Ecosystem is the Moat

The natural question from a Buffett-Munger perspective is whether this dominance is durable. And here, I think, we must be both honest and nuanced.

ARM's moat is not primarily the architecture itself. RISC-V has proven that you can design competitive instruction sets -- Tenstorrent's Ascalon-X core is approaching parity with ARM's best. The moat is the ecosystem. Twenty-two million developers. Every major operating system -- Android, iOS, Windows, Linux -- deeply optimized for ARM. Decades of compiler toolchains, libraries, security patches, and software stacks. When a chip designer starts a new project, choosing ARM means inheriting all of this. Choosing RISC-V means building much of it from scratch.

This is what Munger would call a "lollapalooza" -- multiple reinforcing advantages creating a self-perpetuating cycle. More developers leads to more software leads to more demand for ARM chips leads to more developers. CSS (Compute Subsystems) strengthens this further by giving customers pre-verified, production-ready chip designs that save two to three years of development time.

But here is where intellectual honesty demands that we confront the limits of this moat. The ecosystem advantage, while formidable, is not permanent. Linux took two decades to challenge Windows. Android took a decade to challenge iOS in developer attention. RISC-V is following a similar arc. It will not displace ARM in smartphones or data centers by 2030. But by 2035? The question is legitimate.

China adds a particularly troubling dimension. When a nation of 1.4 billion people, with the world's largest semiconductor market, makes a strategic decision to abandon your architecture for a free alternative, that is not a competitive threat -- it is an industrial policy shift. ARM cannot compete against sovereign mandates. The 24% of revenue from China is not secure on a 10-year horizon.

The Owner's Mindset: Would Buffett Own This for 20 Years?

This is the central question, and I believe the answer is: yes, he would own the business -- but no, he would not own the stock at this price.

Buffett has always drawn a sharp distinction between a wonderful business and a wonderful investment. Coca-Cola was a wonderful business in 1998 at $88 per share, but it was a terrible investment that took 16 years to recover. ARM at $197 with a 262x trailing P/E is Coca-Cola 1998 on steroids.

The fundamental problem is not the business. The business is magnificent. It is the translation of business quality into shareholder returns. Consider:

  • Stock-based compensation of $820M-$1B per year means ARM is giving away roughly 0.5% of its market cap annually to employees
  • True free cash flow after SBC is negative -- the company is destroying economic value for shareholders even as revenue grows 25%
  • The 90% SoftBank ownership means the company is not run with minority shareholders as the primary constituency
  • No dividends, trivial buybacks -- the cash generated goes to R&D expansion, much of it at SoftBank's strategic direction

Buffett would admire this franchise. He would study it. He would put it on his watchlist. And he would wait for the market to give him a price that provides a margin of safety. That price is somewhere between $80 and $105 -- roughly where the stock traded as recently as early 2026.

Risk Inversion: What Could Destroy This Business?

Inverting the question -- as Munger always insisted -- what would have to happen for ARM to fail?

  1. RISC-V achieves ecosystem parity in high-performance computing (probability: 20-30% by 2035). If the open-source ecosystem reaches a tipping point where developer tools, OS support, and security infrastructure match ARM's, the licensing model's pricing power erodes. ARM would still survive but might look more like a utility than a growth franchise.

  2. China fully decouples from ARM (probability: 30-40% by 2030). If the Arm China entity is nationalized, or if Chinese chipmakers migrate wholesale to RISC-V, ARM loses its second-largest market. This is perhaps the most likely of the catastrophic scenarios.

  3. A paradigm shift beyond traditional CPU architectures (probability: 10-15% by 2035). If neuromorphic computing, optical computing, or quantum computing displaces classical CPUs for major workloads, ARM's entire value proposition becomes irrelevant. This is remote but non-zero.

  4. SoftBank mismanagement (probability: 15-20%). SoftBank pushes ARM into capital-intensive businesses (chiplets, ASICs, full solutions) that destroy the asset-light model and dilute returns on invested capital. Early signs of this are already visible.

None of these individually is likely to destroy ARM. But in combination, they could erode the moat from a "wide" to "narrow" classification over a decade.

Valuation Philosophy: Is Price Justified by Quality?

In a word: no.

ARM at $197 is priced at 44x revenue, 262x trailing earnings, and 96x forward earnings. To put this in perspective, the most expensive major technology franchise before the 2000 bubble -- Microsoft at its peak -- traded at roughly 70x forward earnings. ARM today is more expensive than peak-bubble Microsoft on a forward basis.

The comparisons to Synopsys (40x forward) and Cadence (55x forward) are instructive. These are also IP-heavy, high-margin, semiconductor-adjacent businesses with wide moats. ARM grows faster (25% vs 12-16%), which justifies some premium. But a 2x premium on forward P/E? That is not justified by 10 percentage points of additional growth.

The float scarcity (13% public float) creates an artificial valuation premium. If SoftBank's 90% stake were freely tradeable, ARM would almost certainly trade 30-50% lower simply due to normalized supply-demand dynamics. Paying a premium for artificial scarcity is not value investing -- it is the opposite.

The Patient Investor's Path: When and How to Act

ARM is a business worth owning for 20 years. But the entry price determines whether those 20 years are profitable.

At $80 (Strong Buy), ARM trades at approximately 40x forward earnings -- a meaningful premium to the market but justified by the franchise quality and growth trajectory. At this price, even if growth slows to 15% and RISC-V captures the IoT segment entirely, you earn a respectable return.

At $105 (Accumulate), ARM trades at approximately 52x forward earnings. This is expensive by traditional value standards but reasonable for a wide-moat IP licensor growing 20%+. This is where patient capital should begin to accumulate.

At $197 (current), there is no margin of safety. Period. Every dollar invested at this price is a bet that the euphoria scenario materializes. And in four decades of market history, euphoria scenarios almost never fully materialize.

The path is clear: add ARM to the watchlist. Set alerts at $105 and $80. Wait. The stock touched $80 just twelve months ago. With a 3.34 beta, it will visit those levels again. When it does, act decisively. Until then, admire from a distance. This is what Buffett calls "sitting on your hands" -- and it is the hardest, most profitable thing an investor can do.

ARM Holdings (ARM) - Investment Analysis

PHASE 1: RISK ASSESSMENT (Kill or Continue)

1.1 SoftBank Control and Lockup Risk (CRITICAL)

SoftBank Group owns approximately 90% of ARM's outstanding shares (~923M shares). This creates several structural risks:

  • Controlled company exemption: ARM operates under NASDAQ's controlled company rules, meaning SoftBank can override minority shareholders on governance, board composition, and strategic direction
  • Margin loan collateral: SoftBank has pledged ARM shares as collateral for an $8.5B margin loan requiring $34B in collateral value at 25% LTV. Breach occurs at ~$37/share -- remote but existential in a crash
  • Float scarcity: Only ~137M shares in public float (13% of total). This creates artificial scarcity that inflates the stock in bull markets but could cause violent drawdowns if SoftBank ever sells even modest amounts
  • Strategic misalignment: SoftBank's AI ambitions (Project Izanagi, $100B+ data center plans) may conflict with ARM's optimal capital allocation. Already expanding into design services/chiplets/ASICs at SoftBank's direction, adding execution risk to a previously asset-light model

Verdict: SoftBank control is a permanent governance risk but not a dealbreaker. The margin loan risk is distant at current prices. The real concern is strategic drift.

1.2 China Revenue Risk (HIGH)

  • China accounted for ~24% of ARM's revenue (FY2023), primarily through the Arm China joint venture entity
  • Arm China operates independently -- neither ARM nor SoftBank controls its operations
  • Revenue from China grew only ~7.5% YoY despite AI-driven demand elsewhere, suggesting geopolitical headwinds are already biting
  • China's aggressive pivot to RISC-V is a direct response to US export controls and ARM licensing restrictions
  • If US-China tensions escalate further, ARM could lose its second-largest revenue source -- and unlike most companies, it cannot simply exit and re-enter

Verdict: China risk is real and worsening. The Arm China entity structure is a governance black hole. However, the data center and AI growth elsewhere is partially compensating.

1.3 RISC-V Competition (MODERATE-TO-HIGH, LONG TERM)

  • RISC-V has captured ~25% of the global processor market as of early 2026, primarily in IoT and embedded segments
  • Tenstorrent's "Ascalon-X" core achieves ~22 SPECint2006/GHz, approaching parity with AMD Zen 5 and ARM Neoverse V3
  • Performance gap expected to close by end of 2026 or early 2027
  • RISC-V SoC market projected to grow from $6.1B (2023) to $92.7B (2030) at 47.4% CAGR
  • China is the primary driver -- essentially building a parallel semiconductor ecosystem on RISC-V
  • ARM's response: CSS (Compute Subsystems) that offer "plug-and-play" solutions the RISC-V ecosystem lacks

Verdict: RISC-V is ARM's most significant long-term competitive threat. Currently contained to IoT/embedded/China, but the trajectory toward data center competition by 2027-2028 is concerning. ARM's software ecosystem (22M+ developers) is its primary moat defense, but ecosystems can shift.

1.4 Royalty Rate Compression Risk (MODERATE)

  • ARM has successfully raised royalty rates through Armv9 and CSS
  • CSS delivers double the royalty of standard Armv9
  • But customers like Apple, Qualcomm, and Samsung have historically pushed back on pricing
  • Qualcomm litigation (resolved in ARM's favor) demonstrated that customers will fight rate increases
  • If RISC-V becomes viable in premium segments, it creates a credible alternative that strengthens customer negotiating position

1.5 Valuation Risk (EXTREME)

  • Trailing P/E: 262x
  • Forward P/E: ~96x
  • EV/Revenue: 44x
  • EV/EBITDA: 186x
  • Price/Book: 26.8x
  • Price/Sales: 44.7x
  • Beta: 3.34 (extreme volatility)

This is the most expensive major company in the semiconductor universe. Any disappointment in growth rates will produce violent corrections -- as the 52-week range ($100-$211) demonstrates, a 52% drawdown occurred within a single year.

Phase 1 Verdict: PROCEED WITH CAUTION. No single risk is fatal, but the combination of extreme valuation + China exposure + RISC-V threat + SoftBank governance creates a risk profile that demands an extraordinary business to justify.

PHASE 2: FINANCIAL ANALYSIS

2.1 Revenue Trajectory

Fiscal Year (Mar-end) Revenue ($M) YoY Growth Royalty ($M) License ($M)
FY2021 (Mar 2021) 2,027 -- ~1,200 (est) ~827 (est)
FY2022 (Mar 2022) 2,703 +33.3% ~1,500 (est) ~1,203 (est)
FY2023 (Mar 2023) 2,679 -0.9% ~1,500 (est) ~1,179 (est)
FY2024 (Mar 2024) 3,233 +20.7% ~1,800 (est) ~1,433 (est)
FY2025 (Mar 2025) 4,007 +23.9% ~2,100 ~1,907
TTM (Dec 2025) 4,671 +26.3% -- --

Revenue inflected strongly from FY2024 onward, driven by:

  • Armv9 adoption in flagship smartphones (30% of royalty mix, up from 25%)
  • CSS (Compute Subsystems) entering volume production
  • Data center penetration (NVIDIA Grace Blackwell, AWS Graviton, Google Axion, Microsoft Cobalt)
  • AI infrastructure buildout creating secular tailwind

Most recent quarterly run rate: ~$1.2B/quarter ($4.8B annualized). Q2 FY2026 guidance: $1.01-1.11B (midpoint +25% YoY).

2.2 Profitability and Margins

Metric FY2021 FY2022 FY2023 FY2024 FY2025 TTM
Gross Margin 88.4% 91.9% 92.7% 92.8% 94.9% 97.5%
R&D as % Rev 37.9% 35.0% 40.3% 59.8% 50.1% ~55%
SG&A as % Rev 40.7% 33.2% 28.4% 30.4% 24.6% ~22%
GAAP Op Margin 11.9% 25.2% 25.3% 3.1% 20.6% 15.4%
Net Income ($M) 388 549 524 306 792 791
Net Margin 19.1% 20.3% 19.6% 9.5% 19.8% 17.2%

Key observations:

  • Gross margins are extraordinary at 95-97% -- this is a pure IP licensing business with virtually no COGS
  • R&D spending has surged from $768M (FY2021) to $2,009M (FY2025), now absorbing 50%+ of revenue
  • This R&D ramp is intentional: building CSS, next-gen architectures, and the chiplet/ASIC opportunity at SoftBank's direction
  • SG&A leverage is improving as revenue scales
  • GAAP profitability is depressed by massive stock-based compensation (~$820M in FY2025, 20% of revenue)

2.3 Stock-Based Compensation (CRITICAL WARNING)

Period SBC ($M) As % Revenue As % Net Income
FY2021 54 2.7% 13.9%
FY2022 26 1.0% 4.7%
FY2023 79 2.9% 15.1%
FY2024 1,037 32.1% 338.9%
FY2025 820 20.5% 103.5%
TTM ~991 ~21.2% ~125.3%

SBC exploded post-IPO and remains at alarming levels. At $820M-$1B annually, ARM is effectively transferring ~0.5% of market cap per year to employees. This makes GAAP earnings nearly meaningless -- true economic earnings are much lower than reported.

2.4 Cash Flow Analysis

Fiscal Year OCF ($M) CapEx ($M) FCF ($M) SBC ($M) FCF ex-SBC ($M)
FY2021 1,233 165 1,068 54 1,014
FY2022 458 75 383 26 357
FY2023 739 93 646 79 567
FY2024 1,090 143 947 1,037 -90 (!)
FY2025 397 219 178 820 -642 (!)

On a "true" free cash flow basis (FCF minus the real economic cost of SBC), ARM has been cash-flow-negative for the past two fiscal years. The business is essentially giving away more equity to employees than it generates in free cash. At a $209B market cap, this is troubling.

Recent quarterly OCF has improved (Q2/Q3 FY2026: $567M and $365M respectively), but CapEx is also rising ($138-179M/quarter) as ARM invests in physical infrastructure for its expanding chiplet/ASIC ambitions.

2.5 Balance Sheet

Metric (Mar 2025) Value
Total Assets $8.9B
Cash & Short-Term Investments $2.8B
Current Receivables $1.9B
Goodwill $1.6B
Total Liabilities $2.1B
Long-Term Debt $0 (only $356M capital leases)
Shareholders' Equity $6.8B
Book Value/Share $7.34

The balance sheet is clean -- essentially debt-free with $2.8B in cash. This is exactly what you'd expect from a capital-light IP licensor. No financial distress risk.

2.6 EPS Growth and Analyst Beats

Quarter Reported EPS Estimate Surprise
Q3 FY2026 (Dec 2025) $0.43 $0.41 +4.9%
Q2 FY2026 (Sep 2025) $0.39 $0.33 +18.2%
Q1 FY2026 (Jun 2025) $0.35 $0.35 0%
Q4 FY2025 (Mar 2025) $0.55 $0.52 +5.8%
Q3 FY2025 (Dec 2024) $0.39 $0.34 +14.7%
Q2 FY2025 (Sep 2024) $0.30 $0.26 +15.4%

ARM has beaten or met estimates in every quarter since its IPO. Annual EPS: $1.25 (FY2024) -> $1.64 (FY2025) -> $1.17 annualized through 3 quarters of FY2026.

Phase 2 Verdict: The business economics are magnificent at the gross profit level -- 95%+ margins on pure IP licensing. But the picture deteriorates dramatically below that line due to massive R&D investment ($2B+/year), enormous SBC ($820M-$1B), and rising CapEx. True owner earnings are far lower than GAAP suggests.

PHASE 3: MOAT ASSESSMENT

3.1 Moat Sources

Architecture Standard Lock-In (WIDE)

  • ARM architecture is the global standard for mobile processors. 99%+ of smartphones use ARM CPUs
  • 280+ billion ARM-based chips shipped cumulatively
  • ~70% of the world's population uses ARM-based products daily
  • This level of ubiquity creates an ecosystem moat comparable to Windows on the desktop

Software Ecosystem (WIDE)

  • 22+ million developers build on ARM -- 80%+ of the global developer base
  • Decades of compiler optimizations, toolchains, libraries, and OS support
  • Android, iOS, Windows on ARM, Linux -- all major operating systems have deep ARM integration
  • This is the #1 barrier to RISC-V disruption. Hardware architectures are replaceable; software ecosystems are not -- or at least, not quickly

Network Effects (MODERATE-TO-WIDE)

  • More ARM chips = more developers = more software = more demand for ARM chips
  • Every new licensee (now expanding beyond traditional semis to hyperscalers, automakers, governments) strengthens the ecosystem
  • 70,000+ enterprises running AI workloads on ARM Neoverse (up 40% YoY, 14x since 2021)

IP Portfolio and R&D Depth (WIDE)

  • Decades of architectural development that cannot be replicated from scratch
  • CSS (Compute Subsystems) offer "plug-and-play" solutions that save licensees 2-3 years of design time
  • CSS royalty rates are double standard Armv9, and next-gen CSS platforms deliver even higher rates
  • No comparable offering exists in RISC-V

Switching Costs (MODERATE)

  • Once a chip design is built on ARM architecture, switching requires complete redesign
  • Software compiled for ARM must be recompiled and tested for alternatives
  • Multi-year design cycles mean today's design wins lock in 3-5 years of future royalties

3.2 Moat Erosion Risks

RISC-V in IoT/Embedded: Already losing share in low-end applications where ARM's ecosystem advantage matters least.

RISC-V in China: China's strategic pivot to RISC-V is essentially a national industrial policy decision. ARM cannot compete against sovereign mandates.

RISC-V in Data Center (2027+): If Tenstorrent or others achieve Neoverse-class performance, hyperscalers may dual-source.

Customer Consolidation: Apple designs its own ARM-based chips. If more customers do custom silicon (which CSS enables), customers gain negotiating leverage.

3.3 Moat Width and Durability

Width: WIDE -- The combination of architectural ubiquity, software ecosystem, and CSS value proposition creates a formidable moat.

Durability: 10-15 years with moderate confidence -- ARM's moat faces a credible long-term threat from RISC-V that x86 did not represent (because RISC-V is open-source). The moat is widest in high-performance applications and narrowest in IoT/embedded and China.

Moat Trend: STABLE but with risks -- CSS is widening the moat in the near term, but RISC-V's ecosystem maturation is a secular headwind.

PHASE 4: VALUATION AND SYNTHESIS

4.1 Comparable Company Analysis

Company P/E (TTM) P/E (Fwd) EV/Rev Gross Margin Growth
ARM 262x 96x 44x 95%+ 24-26%
Synopsys (SNPS) ~48-80x ~40x ~12x 80% 10-15%
Cadence (CDNS) ~71x ~55x ~18x 89% 12-16%

ARM trades at a significant premium to its closest IP licensing comparables. Even adjusting for ARM's higher growth, the premium is extreme. SNPS at 40x forward P/E with 12% growth implies PEG ~3.3x. ARM at 96x forward P/E with 25% growth implies PEG ~3.8x.

4.2 DCF Valuation

Assumptions:

  • FY2026E revenue: ~$5.0B (+25%)
  • Revenue growth years 1-5: 20%, 18%, 16%, 14%, 12%
  • Terminal growth: 4% (premium for IP licensing model)
  • Operating margin trajectory: 25% -> 30% -> 35% (as R&D investment normalizes)
  • Tax rate: 15% (IP licensing jurisdiction optimization)
  • SBC-adjusted margin: subtract 5-8% for ongoing dilution
  • WACC: 12% (high beta, no debt benefit)
  • Terminal EV/EBITDA: 30x (premium IP licensor)
Scenario Fair Value/Share Implied P/E (FY2028)
Bear (15% growth, 25% margin) $75-85 ~30x
Base (20% growth, 30% margin) $110-130 ~45x
Bull (25% growth, 35% margin) $155-175 ~60x
Euphoria (30% growth, 40% margin) $200+ ~75x+

The current price of ~$197 is above even the bull case and requires the euphoria scenario. The market is pricing in 25-30% annual revenue growth for 5+ years, margins expanding to 35-40%, no RISC-V erosion, and no China deterioration.

4.3 Entry Price Calculation

Strong Buy: $80 (~40x forward P/E on ~$2.00 FY2028E EPS)

  • Implies a 60% decline from current levels
  • Would require a severe market correction or RISC-V panic
  • At this price, you own a wide-moat IP licensor at a reasonable growth premium
  • The stock touched $80 in April 2025 -- these opportunities occur

Accumulate: $105 (~52x forward P/E on ~$2.00 FY2028E EPS)

  • Implies a 47% decline from current levels
  • Premium to Synopsys/Cadence but justified by higher growth and AI exposure
  • Touched this level in January 2026

Current Price Assessment: ~$197

  • At 96x forward P/E, price already reflects 5+ years of optimistic execution
  • No margin of safety exists at this level
  • Owning ARM at $197 is a bet on continued momentum, not value investing

INVESTMENT THESIS

The Bull Case (Not at This Price)

ARM is one of the greatest technology franchises ever created. Its architecture is the computing standard for the AI age -- from smartphones (99% share) to data centers (approaching 50% of hyperscaler deployments). The royalty model means ARM earns recurring revenue on every chip shipped, and CSS is doubling per-chip value. With 22M+ developers, the ecosystem moat is wider than any competitor can realistically attack in the medium term.

The Bear Case (Why We WAIT)

The market has already priced in all of this -- and more. At 262x trailing earnings and 96x forward, ARM must execute flawlessly for a decade to justify its valuation. Meanwhile:

  • SBC consumes more cash than the business generates in true FCF
  • RISC-V is closing the performance gap and capturing China strategically
  • SoftBank's 90% ownership distorts both governance and float
  • R&D spending is accelerating (50%+ of revenue), with uncertain returns
  • The stock has a 3.34 beta -- drawdowns of 40-50% are routine

This is a wonderful business at a terrible price.

The Path to Ownership

ARM will become attractive during:

  1. A broad market correction that brings ARM to 40-50x forward P/E
  2. RISC-V fear event that creates temporary panic
  3. SoftBank forced selling
  4. An earnings miss that resets growth expectations

The stock touched $80 in April 2025 and $100 in January 2026. Patience is rewarded.

KEY DATA SOURCES

  • AlphaVantage: Financial statements, Company Overview, Earnings, Monthly Prices
  • AlphaVantage: Earnings Call Transcripts Q4 FY2025 and Q1 FY2026
  • Web research: RISC-V competitive positioning, SoftBank ownership structure, SNPS/CDNS valuations

=== VERDICT: ARM | WAIT | SB:$80 | Acc:$105 | Current:$197 ===