Fast anchor · not a valuation

Arm Holdings — Bull & Bear Memo

Why this could work, how I lose money, and what to watch — in a form I can reopen in six months.

company
Arm Holdings plc (Nasdaq: ARM)
sources used
Company: FY2026 20-F (year ended 31 Mar 2026); FY2027 Q1 interim 6-K (quarter ended 30 Jun 2026); shareholder letters of 6 May 2026 and 29 Jul 2026; press release of 24 Mar 2026 (Arm AGI CPU launch); FY2025 20-F and the FY2026 auditor’s report, for the related-party and audit-matter wording comparison. Outside sources are used only in section 5 and listed there.
latest period
Q1 FY2027 — quarter ended 30 June 2026
missing or stale data
No earnings-call transcripts, investor-day materials or usable insider-trading data in the source set. Armv9 share of royalty, royalty per chip, AGI CPU pricing, gross margin and wafer commitments are not disclosed, nor are the loan-to-value and ADS price thresholds in the SoftBank margin facility. Whether the $300m fixed payment due in FY2027 sits inside the $645.8m contract asset or is incremental cannot be determined from the filings. Company sources stop at 29 July 2026; external challenge run 10 September 2026.

1 · Business in one line

Arm licenses CPU architecture and processor designs to chipmakers for an upfront fee, then collects a per-chip royalty — a percentage of the chip's selling price, or a fixed fee per unit — on substantially every chip shipped, for as long as it ships. Since March 2026 it also sells a data-center CPU of its own. FY2026 revenue: $4,920m — $2,613m royalty, $2,307m license and other, at a 98% gross margin.

Type: fast grower — three consecutive years above 20% revenue growth since the IPO — sitting on a cyclical royalty base whose units track the semiconductor cycle. The classification matters because Arm has no unit share left to win in its largest market: the bull case must be a pricing story, not a volume story.

2 · Bull case — Peter Lynch pitch

The simple reason this stock could work is that Arm gets paid more each year for the same chip, and has just opened two new ways to get paid for chips it previously only rented cores to.

B1Royalty per chip rises faster than the unit base decays
What must happen
Mix shift toward Armv9 and Arm Compute Subsystems lifts average royalty per chip fast enough that royalty compounds above 20% on a roughly flat unit base.
Why Arm specifically
Royalty rates step up with how much Arm content sits in the chip, and every key mobile operating system depends on Arm cores. A customer who wants a cheaper royalty has to move a software ecosystem, not swap a core — the switching cost sits with the customer, the price increase with Arm.
Evidence
Royalty of $2,613m in FY2026, +21%, attributed in the 20-F to “an improved mix of products with higher royalty rates per chip, such as Armv9 technology”; +22% to $715m in Q1 FY2027. Mobile applications processors, where Arm has held above 99% share for years, are still 43% of royalty — that share cannot grow, so the growth is rate.
What to monitor
Royalty growth versus industry unit growth; mobile's falling share of total royalty.
B2The data center becomes a second royalty pool of comparable size
What must happen
Hyperscaler in-house Arm CPUs keep displacing x86 sockets, each carrying a far larger royalty than a phone.
Why Arm specifically
Royalty is a percentage of chip ASP or a fee that scales with content, so a 100-core server CPU pays a multiple of a smartphone SoC — and every AI accelerator still needs a host CPU.
Evidence
Data-center royalties more than doubled year over year in Q1 FY2027. Cumulative Neoverse shipments passed 1.5 billion cores — the most recent 500 million in nine months, against six years for the first billion. Named programs: NVIDIA Vera in full production, Google Axion hosting its TPU systems, AWS Graviton5 under a multi-year Meta agreement, Azure Cobalt 200, Qualcomm Dragonfly C1000.
What to monitor
Data-center royalty growth rate; Neoverse cumulative-core milestones; any named hyperscaler program slipping.
B3Arm AGI CPU converts dollars of royalty into thousands of dollars of silicon revenue
What must happen
The chip launched in March 2026 ships in volume from end-calendar-2026 and holds its multi-generation roadmap with Meta and the other named customers.
Why Arm specifically
Arm already owns the cores, interconnect and software stack, so the incremental cost of selling the whole chip is physical design and wafers, not architecture — chip-level revenue in a socket that previously paid only a royalty.
Evidence
Committed demand rose from a stated “$1 billion opportunity” across FY2027–28 to “exceeds $2 billion” within one quarter, capacity secured for the $1bn case, initial product delivered to multiple customers. Meta is lead partner and co-developer across multiple generations. Claimed more than 2x performance per rack versus x86.
What to monitor
Conversion of the >$2bn demand into recognized revenue; the first disclosed gross margin on the silicon line; inventory and wafer commitments appearing on the balance sheet.
B4Today's spending is an investment phase that ends
What must happen
R&D growth falls below revenue growth and GAAP operating margin recovers from single digits.
Why Arm specifically
Arm funds one core design and amortizes it across every licensee, so a fixed R&D base spread over a rising royalty stream is the operating leverage — nothing in the licensing model requires cost to scale with revenue.
Evidence
R&D was $2,776m in FY2026 — 56% of revenue — up 34% while revenue grew 23%. GAAP operating margin was 18% for FY2026 and fell to 7.1% in Q1 FY2027 from 10.8%, even as non-GAAP operating margin rose to 41.2%.
What to monitor
R&D growth versus revenue growth; GAAP — not non-GAAP — operating margin.

Why the market might be missing it: royalty is recognized a quarter in arrears on estimated shipments, so a data-center inflection reaches Arm's reported numbers materially later than it happens in its customers' fabs.

Each assumption beside the risk that tries to kill it. If an attack reads weaker than the assumption, that is where the thesis actually rests.

B1 The assumption

Royalty per chip rises faster than the unit base decays

Mix shift toward Armv9 and Arm Compute Subsystems lifts average royalty per chip fast enough that royalty compounds above 20% on a roughly flat unit base.

Evidence. Royalty of $2,613m in FY2026, +21%, attributed in the 20-F to “an improved mix of products with higher royalty rates per chip, such as Armv9 technology”; +22% to $715m in Q1 FY2027. Mobile applications processors, where Arm has held above 99% share for years, are still 43% of royalty — that share cannot grow, so the growth is rate.

The attacks

R1 Arm now competes with the customers who pay its royalties

The AGI CPU and Arm's direct hyperscaler relationships put it in the same socket as Broadcom, Marvell, Qualcomm and other licensees. Arm's own 20-F says these customers may “seek alternative architectures”, “withhold sensitive roadmap information” or “demand more favorable commercial terms”. A licensee cannot punish Arm this quarter — it starts a RISC-V program that removes a socket in five years.

Damage. Permanent: the royalty base is the entire durable asset, and an architecture that leaves does not come back.

R2 Arm cannot enforce the licence terms that make architecture licensing safe

Arm sued Qualcomm and Nuvia to force destruction of cores developed under a terminated architecture licence. The jury found the technology licensed under Qualcomm's own ALA, and in September 2025 the court granted Qualcomm judgment as a matter of law that Nuvia had not breached. Arm's appeal is pending; Qualcomm's counter-suit goes to trial in Q4 CY2026. If that stands, an architecture licence becomes a cheap option: acquire a design house, keep the cores, pay the architecture rate rather than the implementation rate.

Damage. Qualcomm alone was 9% of FY2026 revenue; the larger loss is the price ceiling — implementation royalties hold only if the cheaper route stays legally closed.

R6 The two largest revenue concentrations are entities Arm does not control

Arm China — an independent distributor Arm neither owns nor manages — is Arm's single largest customer at 16% of FY2026 revenue. Arm depends on Arm China's own reported figures to determine what it is owed, and recognized $12.3m of expected credit losses against it in FY2026, taking the allowance to $28.3m. PRC revenue was 18% of total; the top five customers, including Arm China and SoftBank Group, were 57%.

Damage. A sixth of revenue reaches Arm through a counterparty it cannot audit and a government that can close the channel.

B2 The assumption

The data center becomes a second royalty pool of comparable size

Hyperscaler in-house Arm CPUs keep displacing x86 sockets, each carrying a far larger royalty than a phone.

Evidence. Data-center royalties more than doubled year over year in Q1 FY2027. Cumulative Neoverse shipments passed 1.5 billion cores — the most recent 500 million in nine months, against six years for the first billion. Named programs: NVIDIA Vera in full production, Google Axion hosting its TPU systems, AWS Graviton5 under a multi-year Meta agreement, Azure Cobalt 200, Qualcomm Dragonfly C1000.

The attack

R1 Arm now competes with the customers who pay its royalties

The AGI CPU and Arm's direct hyperscaler relationships put it in the same socket as Broadcom, Marvell, Qualcomm and other licensees. Arm's own 20-F says these customers may “seek alternative architectures”, “withhold sensitive roadmap information” or “demand more favorable commercial terms”. A licensee cannot punish Arm this quarter — it starts a RISC-V program that removes a socket in five years.

Damage. Permanent: the royalty base is the entire durable asset, and an architecture that leaves does not come back.

B3 The assumption

Arm AGI CPU converts dollars of royalty into thousands of dollars of silicon revenue

The chip launched in March 2026 ships in volume from end-calendar-2026 and holds its multi-generation roadmap with Meta and the other named customers.

Evidence. Committed demand rose from a stated “$1 billion opportunity” across FY2027–28 to “exceeds $2 billion” within one quarter, capacity secured for the $1bn case, initial product delivered to multiple customers. Meta is lead partner and co-developer across multiple generations. Claimed more than 2x performance per rack versus x86.

The attacks

R1 Arm now competes with the customers who pay its royalties

The AGI CPU and Arm's direct hyperscaler relationships put it in the same socket as Broadcom, Marvell, Qualcomm and other licensees. Arm's own 20-F says these customers may “seek alternative architectures”, “withhold sensitive roadmap information” or “demand more favorable commercial terms”. A licensee cannot punish Arm this quarter — it starts a RISC-V program that removes a socket in five years.

Damage. Permanent: the royalty base is the entire durable asset, and an architecture that leaves does not come back.

R4 Production silicon destroys the financial model the multiple is paid for

Arm's 98% gross margin exists because it ships files, not wafers. The 20-F warns the silicon business brings “margin compression, greater quarter-to-quarter revenue volatility, increased capital intensity and working capital requirements”, plus foundry commitments that cannot flex, inventory obsolescence and a longer cash conversion cycle. At $2bn of silicon against roughly $5bn of IP revenue, blended gross margin falls mechanically however well Arm executes.

Damage. Structural, and realized by the bull case working rather than failing — which is why it is the easiest risk to underwrite away.

B4 The assumption

Today's spending is an investment phase that ends

R&D growth falls below revenue growth and GAAP operating margin recovers from single digits.

Evidence. R&D was $2,776m in FY2026 — 56% of revenue — up 34% while revenue grew 23%. GAAP operating margin was 18% for FY2026 and fell to 7.1% in Q1 FY2027 from 10.8%, even as non-GAAP operating margin rose to 41.2%.

The attacks

R3 Most of FY2026's growth came from the controlling shareholder

Revenue under the consulting agreement with a SoftBank Group affiliate was $704.4m in FY2026 against $145.5m in FY2025 — $559m of the $913m total revenue increase, and 14% of FY2026 revenue, from an entity controlled by the shareholder that owns 86.4% of Arm. It is recognized over time on Arm’s own estimate of engineering labour hours against total expected effort, and billed on a schedule written into a contract with its own parent — so both the amount and the timing rest on judgment exercised inside the group. Strip it out and FY2026 revenue grew about 9%; revenue from external customers alone grew 7.4%, and their license and other revenue fell 9% — which the 20-F states outright as “a $123 million, or 9%, decrease”. $645.8m sat unbilled at 31 March 2026 — 66% of all current contract assets and half the total. Deloitte’s FY2026 critical audit matter on long-term revenue contracts newly specifies “including related parties”; the FY2025 wording did not.

Damage. Less to cash flow than to the multiple: the market has been capitalizing a 23% grower whose arm's-length business grew about 9%.

R5 Stock compensation already consumes the operating profit

Equity-settled share-based compensation was $1,052m in FY2026 against GAAP operating income of $900m; with employer taxes, $1,212m, or 25% of revenue. The 41.2% non-GAAP operating margin quoted in Q1 FY2027 was a 7.1% GAAP margin. Diluted shares rose only from 1,044m to 1,068m over two years — the cost lands in the income statement rather than the share count, real in either place.

Damage. Caps the operating leverage in B4 — R&D can decelerate and GAAP margin still not recover if equity comp scales with engineering headcount.

3 · Bear case — Munger invert

The most likely way I lose money is that Arm's reported growth turns out to have been bought from its own parent, while the licensees it now competes with quietly fund the alternative.

Ordered by permanence, not by mirroring the bull sequence.

R1Arm now competes with the customers who pay its royalties

Attacks B1, B2, B3 · Permanent

How it could fail
The AGI CPU and Arm's direct hyperscaler relationships put it in the same socket as Broadcom, Marvell, Qualcomm and other licensees. Arm's own 20-F says these customers may “seek alternative architectures”, “withhold sensitive roadmap information” or “demand more favorable commercial terms”. A licensee cannot punish Arm this quarter — it starts a RISC-V program that removes a socket in five years.
What would confirm failure
An architecture licensee announcing a non-Arm core for a segment where it currently ships Arm; licence renewals on visibly worse terms.
Damage
Permanent: the royalty base is the entire durable asset, and an architecture that leaves does not come back.
R2Arm cannot enforce the licence terms that make architecture licensing safe

Attacks B1 · Largely permanent

How it could fail
Arm sued Qualcomm and Nuvia to force destruction of cores developed under a terminated architecture licence. The jury found the technology licensed under Qualcomm's own ALA, and in September 2025 the court granted Qualcomm judgment as a matter of law that Nuvia had not breached. Arm's appeal is pending; Qualcomm's counter-suit goes to trial in Q4 CY2026. If that stands, an architecture licence becomes a cheap option: acquire a design house, keep the cores, pay the architecture rate rather than the implementation rate.
What would confirm failure
The Third Circuit affirming; an adverse Q4 CY2026 verdict; other licensees restructuring toward architecture-only terms.
Damage
Qualcomm alone was 9% of FY2026 revenue; the larger loss is the price ceiling — implementation royalties hold only if the cheaper route stays legally closed.
R3Most of FY2026's growth came from the controlling shareholder

Attacks B4 and the headline growth rate · Fixable, but it has already repriced the growth rate

How it could fail
Revenue under the consulting agreement with a SoftBank Group affiliate was $704.4m in FY2026 against $145.5m in FY2025 — $559m of the $913m total revenue increase, and 14% of FY2026 revenue, from an entity controlled by the shareholder that owns 86.4% of Arm. It is recognized over time on Arm’s own estimate of engineering labour hours against total expected effort, and billed on a schedule written into a contract with its own parent — so both the amount and the timing rest on judgment exercised inside the group. Strip it out and FY2026 revenue grew about 9%; revenue from external customers alone grew 7.4%, and their license and other revenue fell 9% — which the 20-F states outright as “a $123 million, or 9%, decrease”. $645.8m sat unbilled at 31 March 2026 — 66% of all current contract assets and half the total. Deloitte’s FY2026 critical audit matter on long-term revenue contracts newly specifies “including related parties”; the FY2025 wording did not.
Counter-evidence
The balance is converting to cash. The affiliate contract asset fell from $645.8m to $577.2m over Q1 FY2027 despite $192.9m of new revenue recognized — implying roughly $261.5m billed in the quarter (arithmetic from disclosed figures). Loss allowances on contract assets are “immaterial for all periods presented”, and Arm states it has “an enforceable right to payment for performance completed to date”. This is money Arm cannot yet bill, not money it may never collect.
What would confirm failure
The consulting line flattening while ACV growth stays in the low teens. It has not flattened: Q1 FY2027 consulting revenue of $192.9m was up 53% on Q1 FY2026 and above the FY2026 quarterly average of $176m — so the headline is still being flattered. ACV meanwhile decelerated from +22% at March 2026 to +13% at June 2026. The other tell is the contract asset ceasing to convert — rising, or flat while new consulting revenue is booked.
Damage
Less to cash flow than to the multiple: the market has been capitalizing a 23% grower whose arm's-length business grew about 9%.
R4Production silicon destroys the financial model the multiple is paid for

Attacks B3 · Permanent if the silicon business succeeds

How it could fail
Arm's 98% gross margin exists because it ships files, not wafers. The 20-F warns the silicon business brings “margin compression, greater quarter-to-quarter revenue volatility, increased capital intensity and working capital requirements”, plus foundry commitments that cannot flex, inventory obsolescence and a longer cash conversion cycle. At $2bn of silicon against roughly $5bn of IP revenue, blended gross margin falls mechanically however well Arm executes.
What would confirm failure
The first quarter reporting inventory, wafer purchase commitments and a gross margin below the mid-90s.
Damage
Structural, and realized by the bull case working rather than failing — which is why it is the easiest risk to underwrite away.
R5Stock compensation already consumes the operating profit

Attacks B4 · Ongoing; fixable only by slowing hiring

How it could fail
Equity-settled share-based compensation was $1,052m in FY2026 against GAAP operating income of $900m; with employer taxes, $1,212m, or 25% of revenue. The 41.2% non-GAAP operating margin quoted in Q1 FY2027 was a 7.1% GAAP margin. Diluted shares rose only from 1,044m to 1,068m over two years — the cost lands in the income statement rather than the share count, real in either place.
What would confirm failure
GAAP operating margin staying in single digits as silicon cost of sales lands on top.
Damage
Caps the operating leverage in B4 — R&D can decelerate and GAAP margin still not recover if equity comp scales with engineering headcount.
R6The two largest revenue concentrations are entities Arm does not control

Attacks B1 · Structural, outside Arm's control

How it could fail
Arm China — an independent distributor Arm neither owns nor manages — is Arm's single largest customer at 16% of FY2026 revenue. Arm depends on Arm China's own reported figures to determine what it is owed, and recognized $12.3m of expected credit losses against it in FY2026, taking the allowance to $28.3m. PRC revenue was 18% of total; the top five customers, including Arm China and SoftBank Group, were 57%.
What would confirm failure
A China export-control or end-use change touching the IPLA; a widening credit allowance against Arm China; PRC revenue turning negative after +17% in FY2026.
Damage
A sixth of revenue reaches Arm through a counterparty it cannot audit and a government that can close the channel.
R7Most of the parent’s stake is pledged against a margin loan

Attacks no bull assumption · A supply risk, not a business risk · Outside Arm’s control

How it could fail
A SoftBank subsidiary borrowed $8.5bn under a margin loan facility “initially secured by a pledge of 769,029,000 of our ordinary shares, which, as of May 21, 2026, represents a 72.0% equity interest in us” — about 83% of SoftBank’s own 86.4% holding. The facility requires prepayment on certain change-of-control events “or in the event the trading price of our ADSs declines below certain thresholds”, and a margin call if loan-to-value passes a threshold, in which case the SoftBank subsidiary “may decide to sell some of the pledged shares to provide such funds”. A forced sale lands in a float of roughly 14%.
What would confirm failure
The loan-to-value or ADS price thresholds being disclosed; the pledged share count increasing; a registered secondary offering; any SoftBank refinancing that references the facility.
Damage
None to the operating business — Arm states it has “no material obligations with respect to the SoftBank Group Facility”. The exposure is to share supply, and because the thresholds are undisclosed its size and timing cannot be assessed from the filings.

4 · Signals to monitor

Every signal carries the assumption or risk it tracks. Values are as of the latest source listed.

Royalty revenue growth, year over year +22% ($715m)
Sustained above ~20% supports B1; a drop toward unit growth breaks it
Shareholder letter, 29 Jul 2026 B1
Data-center royalty growth and Neoverse cores shipped More than doubled y/y; 1.5bn cores (last 500m in 9 months)
Data-center royalty must stay well above total royalty growth, and the interval per 500m cores must keep shortening, for B2
Shareholder letter, 29 Jul 2026 B2
Arm AGI CPU committed demand, FY2027-28 Exceeds $2bn (was $1bn)
Conversion into recognized revenue confirms B3; a flat or cut figure kills it
Shareholder letter, 29 Jul 2026 B3
ACV, and the SoftBank consulting line behind it ACV $1,732m, +13% y/y (was +22%); consulting $192.9m in Q1, $577m unbilled
ACV stuck in the low teens while the consulting line flattens confirms R3 — organic growth nearer 9% than 22%. Not yet: consulting grew 53% y/y in Q1 FY2027, so the headline is still flattered
Shareholder letters 6 May and 29 Jul 2026; FY2027 Q1 6-K R3
GAAP operating margin 7.1% (vs 10.8% a year earlier)
Recovery supports B4; further decline as headcount and equity comp scale confirms R5
Shareholder letter, 29 Jul 2026 B4R5
Consolidated gross margin 97.2% GAAP
The first move below the mid-90s, alongside inventory and wafer commitments, marks R4 arriving
Shareholder letter, 29 Jul 2026 R4
Qualcomm litigation Appeal pending; trial Q4 CY2026
Judgment as a matter of law went to Qualcomm in Sep 2025; an adverse appeal or verdict confirms R2. Qualcomm was 9% of FY2026 revenue
FY2026 20-F R2
Top-five customer concentration 57% of FY2026 revenue (Arm China 16%)
A licensee responding to Arm as a competitor, or a China channel disruption, shows up here first - R1 and R6
FY2026 20-F R1R6
SoftBank affiliate contract asset $577.2m (was $645.8m at 31 Mar 2026)
Must keep falling while consulting revenue is recognized. Rising, or flat against new revenue, confirms R3; a loss allowance recorded against it would be worse
FY2027 Q1 interim 6-K; FY2026 20-F R3
Arm shares pledged under the SoftBank facility 769,029,000 shares = 72.0% equity interest
Thresholds are undisclosed, so the watchable items are an increase in the pledged count, a registered secondary, or a SoftBank refinancing referencing the facility — R7
FY2026 20-F, Items 3.D, 7.A and 7.B R7

5 · External challenge notes

Run 10 September 2026. Everything below comes from outside Arm's filings and is kept separate from the sections above, so the reader can tell later which claims came from Arm and which came from its critics, competitors and regulators.
E1US FTC opened an antitrust probe into Arm's licensing, reported May 2026.
External point
US FTC opened an antitrust probe into Arm's licensing, reported May 2026. Regulators are examining whether Arm would refuse or degrade its CPU blueprints for third-party licensees while accelerating its own chip business. It follows scrutiny in Europe and South Korea.
How it affects the economics
Turns R1 from a commercial risk into a regulatory one, and cuts both ways. A remedy that forces Arm to license on non-discriminatory terms would cap the pricing power B1 rests on; the investigation alone gives large licensees leverage in renewal negotiations, and constrains how far Arm can favour its own silicon.
What would confirm it
A formal complaint or consent decree; licensing-term concessions disclosed in a renewal; the probe appearing as a named legal proceeding in the next 20-F.
Source (outside the filings)
Bloomberg via Investing.com, 15 May 2026
E2Qualcomm is developing RISC-V CPU cores with Ventana for high-core-count, datacentre-class chiplets, alongside its Arm-based Oryon line; Meta is expanding RISC-V work around its MTIA accelerators.
External point
Qualcomm is developing RISC-V CPU cores with Ventana for high-core-count, datacentre-class chiplets, alongside its Arm-based Oryon line; Meta is expanding RISC-V work around its MTIA accelerators. RISC-V penetration is put near 25% by units and about 10% by revenue (SHD Group).
How it affects the economics
This is R1's confirmation signal arriving early, and from the two names most exposed to Arm's move into silicon — Qualcomm is in litigation with Arm, Meta is the AGI CPU's lead partner. A shipping RISC-V server CPU from a current architecture licensee removes royalty-bearing sockets permanently.
What would confirm it
A RISC-V server CPU shipping in volume from a current Arm architecture licensee; RISC-V share rising in revenue terms, not only in unit terms.
Source (outside the filings)
eeNews Europe, Jan 2026; EE Times Asia; RISC-V International
E3Morgan Stanley cut Arm to Equal-Weight on 7 April 2026, arguing merchant silicon carries 40-50% gross margins against Arm's historic 95%+, and that the R&D surge behind the AI CPU will weigh on operating income for at least eight quarters.
External point
Morgan Stanley cut Arm to Equal-Weight on 7 April 2026, arguing merchant silicon carries 40-50% gross margins against Arm's historic 95%+, and that the R&D surge behind the AI CPU will weigh on operating income for at least eight quarters. DRAM supply constraints were cited as a possible cap on FY2027 growth.
How it affects the economics
Puts an outside number on R4 and R5. If AGI CPU revenue carries a 40-50% gross margin, then roughly $2bn of silicon against ~$5bn of IP revenue pulls consolidated gross margin from ~97% toward the low 80s — before any inventory or warranty cost. That is the mix arithmetic the memo flags, quantified.
What would confirm it
The first disclosed gross margin on the silicon line landing in the 40-50% band; consolidated gross margin falling below the mid-90s as AGI ships.
Source (outside the filings)
Morgan Stanley via Benzinga, Seeking Alpha and Investing.com, 7 Apr 2026
E4SoftBank has been reported in talks for a roughly $5bn margin loan secured on its Arm shares to fund its OpenAI investment, and separately secured a $10bn margin loan against its OpenAI stake in August 2026.
External point
SoftBank has been reported in talks for a roughly $5bn margin loan secured on its Arm shares to fund its OpenAI investment, and separately secured a $10bn margin loan against its OpenAI stake in August 2026.
How it affects the economics
Superseded in part by the filings. This was carried here as external reporting; the FY2026 20-F in fact discloses an $8.5bn SoftBank margin loan secured on 769,029,000 Arm shares — a 72.0% equity interest — with prepayment triggers tied to the ADS price and a margin-call mechanism. It is now a company-sourced risk and is carried as R7. What the external reporting still adds is how much further SoftBank may borrow against the stake, which the filings do not address. The economics are unchanged: minority holders bear a forced-seller risk created outside the business.
What would confirm it
The loan-to-value or ADS price thresholds being disclosed — the pledged share count and the encumbrance are already confirmed in the FY2026 20-F. Beyond that: an increase in the pledged count, or a registered secondary offering.
Source (outside the filings)
Bloomberg via Seeking Alpha, 2026; Bloomberg, 6 Aug 2026
E5NVIDIA sold its remaining Arm shares in February 2026 — about 1.1 million shares, roughly $140m — while deepening the technical partnership through Vera.
External point
NVIDIA sold its remaining Arm shares in February 2026 — about 1.1 million shares, roughly $140m — while deepening the technical partnership through Vera.
How it affects the economics
Minor in dollars and not a thesis input on its own. It matters only as evidence that ownership and partnership have decoupled: Arm's most important ecosystem partner has no equity reason to protect it, which is worth remembering when reading partnership announcements as if they were commitments.
What would confirm it
Nothing to confirm — treat as context, not a signal.
Source (outside the filings)
Bloomberg, CNBC and Tom's Hardware, 17-18 Feb 2026

6 · Bottom line

The stock could work because Arm is repricing a socket it already owns — royalty per chip rising above 20% a year on a flat unit base — while the data center adds a structurally higher-value royalty pool and the AGI CPU turns some sockets from a royalty into a whole chip sale.

For that, Armv9 and CSS mix must keep lifting royalty per chip, Neoverse adoption at AWS, Google, Microsoft, NVIDIA and Qualcomm must keep compounding, the >$2bn of AGI CPU demand must convert to shipped revenue from end-2026, and R&D growth must eventually fall below revenue growth.

The thesis most likely breaks not in a bad quarter but through licensees responding to Arm becoming their competitor — funding RISC-V or in-house architectures for sockets they currently pay royalties on — a decision invisible for years and irreversible once made.

I would change my mind if the SoftBank consulting line flattened while ACV growth stayed in the low teens, putting organic growth nearer 9% than 22% — it has not, growing 53% year over year in Q1 FY2027, so the headline is still flattered and the test is still ahead — if a major architecture licensee announced a non-Arm core in a segment it ships Arm in today, or if the AGI ramp reset consolidated gross margin below the mid-80s.

Sources

Annual reports 2
  • FY2026 20-F, year ended 31 March 2026 (filed 26 May 2026) — revenue split, customer concentration, Arm China and SoftBank related-party notes, Qualcomm litigation, production-silicon risk factors, R&D and share-based compensation; revenue recognition policy (Note 1), contract balances (Note 4), the SoftBank Group Facility share pledge (Items 3.D, 7.A, 7.B) and Deloitte’s critical audit matter
  • FY2025 20-F, year ended 31 March 2025 — prior-year related-party wording and critical audit matter, for the year-on-year disclosure comparison
Interim report 1
  • FY2027 Q1 interim 6-K, quarter ended 30 June 2026 — Q1 revenue split, SoftBank consulting revenue and contract assets, segment costs
Shareholder letters 2
  • 6 May 2026 (Q4 FY2026) — full-year revenue, ACV $1,660m +22%
  • 29 Jul 2026 (Q1 FY2027) — Q1 metrics, ACV $1,732m +13%, data-center royalties, Neoverse cores, AGI CPU demand
Press release 1
  • 24 March 2026 — Arm AGI CPU launch: specifications, Meta as lead partner, ecosystem list
Outside the filings — section 5 only 5
  • Bloomberg via Investing.com, 15 May 2026 — FTC antitrust probe into Arm licensing
  • eeNews Europe (Jan 2026), EE Times Asia, RISC-V International — Qualcomm/Ventana and Meta RISC-V programmes, RISC-V penetration estimates
  • Morgan Stanley via Benzinga, Seeking Alpha and Investing.com, 7 Apr 2026 — downgrade to Equal-Weight on merchant-silicon margins and R&D
  • Bloomberg via Seeking Alpha (2026) and Bloomberg, 6 Aug 2026 — SoftBank margin loans secured on Arm and OpenAI shares
  • Bloomberg, CNBC and Tom's Hardware, 17-18 Feb 2026 — NVIDIA exits its remaining Arm stake