Record Royalties, Borrowed Quality: Initiating Arm at Hold with Related Parties at 37% of Revenue

Results vs. Consensus

Arm reported its fiscal second quarter, ended September 30, 2025, after the close on November 5. It was a clean beat on both guided lines, and unusually it cleared the high end of each rather than merely the midpoint. The Street, however, was sitting almost exactly on the company's own guide: consensus revenue of roughly $1.06 billion against a guide midpoint of $1,060 million, and consensus non-GAAP earnings of $0.33 against a guide midpoint of $0.33. The beat is therefore a beat against Arm's own framing, not against an independently ambitious set of expectations.

Q2 FYE26 Scorecard

Year-Over-Year Comparison

Quarter-Over-Quarter Comparison

Prior-quarter figures below are the reported Q1 FYE26 results, which also tie to the six-month columns in this quarter's filing less the three-month columns.

First-Half Comparison

The half-year view is where the composition problem becomes impossible to read past. Total revenue for the six months grew 22.7%. Revenue from external customers, over the same six months, declined 2.0%.

Revenue Composition

Arm reports as a single operating segment. The chief operating decision maker reviews consolidated information only, and there is no reportable segment margin, no data center revenue line and no end-market revenue disclosure. What the filings do disaggregate is revenue by product line, by customer class and by geography, and for this quarter the customer-class cut is the analytically decisive one.

Revenue by Product Line

Within the licence line, the filing splits recognition timing: over-time revenue of $273 million against $120 million a year ago, and point-in-time revenue of $242 million against $210 million. Over-time recognition grew 128% and now accounts for 53% of the licence line, up from 36%. That is the signature of subscription-style and design-services arrangements rather than discrete IP deliveries, and it is consistent with where the incremental revenue came from.

Revenue by Customer Class

The related-party line decomposes cleanly from the notes. Arm China contributed $243.1 million under the intellectual property licence agreement, up 64.6%, and Arm holds a 10% non-voting interest in the entity that owns 48.2% of Arm China. The SoftBank Group affiliate contributed $177.9 million, up from $43.2 million, under a consulting agreement whose statements of work were amended during the half. Significant influence investees added $0.8 million. Those three sum to $421.8 million, which reconciles to the $422 million reported.

The two are not equivalent in quality. Arm China's revenue is a contractual share of a genuinely independent Chinese licensing business selling to third parties; it is related by ownership structure, not by being a captive customer. The SoftBank affiliate's revenue is a direct commercial arrangement with the controlling shareholder, undated, with the design-services component described by management as funded research and development at a lower margin. We treat the first as ordinary revenue with a governance wrapper and the second as a genuine quality-of-earnings item.

Assessment: A company can sustain a high related-party share indefinitely if the terms are disclosed and the pricing is arm's length and demonstrable. Arm has disclosed the amount and nothing else. With external-customer revenue growing 9.4% in the quarter and declining 2.0% across the half, the reported 34% growth rate is not a number an investor can underwrite at 93 times earnings, and the gap between the two is now the single most important open question in the story.

Revenue by Geography

Assessment: Geography independently corroborates the customer-class table. Japan, where the controlling shareholder is headquartered, contributed $132 million of the $291 million total increase, or 45.4% of all year-on-year revenue growth, from an 8.8% revenue base a year ago. This is not a story about Japanese end demand for Arm CPUs. It is the consulting agreement showing up in a second disclosure.

End Markets: What Management Said and What It Did Not Quantify

Smartphones

The CFO named smartphones as one of the two biggest royalty growth contributors, on higher royalty rates per chip, and said royalty revenue from smartphones "grew in order of magnitude faster than the market" as multiple original equipment manufacturers ramped chips based on Armv9 and compute subsystems. The structural point behind that sentence is the one that matters for the franchise: Arm can grow its smartphone royalty take without the handset market growing at all, because what it is monetising is architecture generation and content per chip. The quarter also brought the Lumex compute subsystem launch and the disclosure that the top four Android vendors are all now shipping subsystem-powered devices.

Assessment: Mechanically confirmed, and the single most important thing Arm proved this quarter. Not sized in dollars, so it can be graded only through the total royalty line.

Data center

Data center royalties more than doubled year on year, the second element of the CFO's two-part attribution, on continued deployment of Arm-based chips by hyperscalers. Cumulative Neoverse CPU cores deployed passed one billion across NVIDIA Grace, AWS Graviton, Google Axion and Microsoft Cobalt. Management expects Arm's share of CPUs deployed by the top hyperscalers to reach nearly 50% this year.

Assessment: The most valuable growth vector in the business and still the least disclosed. Cloud and networking together were sized only under direct questioning, and only as a range management accepted rather than stated (see Analyst Q&A). Until there is a reported dollar line, this pillar can be tracked but not modelled.

Automotive and IoT

Both "continued to grow year-on-year and contributed to our strong royalty performance." Neither was quantified. The quarter's supporting evidence is product-level: a flagship electric vehicle launched on Arm Automotive Enhanced with park assist, voice control and safety functions, and Tesla's next-generation AI5 chip, Arm-based, at up to 40 times the prior generation's AI performance.

Assessment: Real optionality on the same architecture and the same royalty mechanism, with a longer design-to-revenue lag than either smartphones or data center. Treat as a third royalty pool that is not yet a modelled line.

Key Operating Metrics

Assessment: The two contracted-revenue metrics point in opposite directions and management directs attention to the one that does not reconcile to revenue. Annualized contract value is a management-defined figure that annualises committed fees over an assumed contract life, three years for single-use and limited-use licences and for the Arm China aggregate fee. Remaining performance obligations is the accounting figure. One is up 28% and one is down 6%. The letter's narrative and the call both cited the first. The letter's narrative gave the second only as a 1% sequential increase, leaving the 6% annual decline to its metrics table; the call did not mention it, and no analyst raised it.

Key Topics & Management Commentary

Overall Management Tone: Confident and consistently forward-leaning, with the confidence anchored on demand for compute rather than on the composition of this quarter's revenue. Management was precise and volunteered numbers when asked about the parent's contribution and about the data center royalty mix, and categorically unwilling to discuss the silicon roadmap, offering process gates instead of substance. The one place the posture turned notably careful was the description of the parent-funded revenue as lower-margin funded research and development, which was disclosed accurately and without being connected to any of the margin framing elsewhere on the call.

1. Royalty Revenue: A Record, and the Rate Escalator Working as Designed

Royalty revenue of $620 million is a company record and grew 21% against a guide of mid-teens growth. The composition of that growth is what makes it the highest-quality number in the release: it came from rate and content per chip, not from unit volume, and from all four named end markets simultaneously.

"Royalty revenue exceeded our expectations, growing 21% year-on-year to a record $620 million versus our guidance of mid-teens. The biggest growth contributors were smartphones with higher royalty rates per chip and in data center where we continue to see share gains from custom hyperscaler chips. Royalty revenue from smartphones grew in order of magnitude faster than the market as multiple OEMs ramped smartphones based on Armv9 and CSS chips." — Jason Child, Chief Financial Officer

The quarterly report attributes the increase the same way: royalty revenue rose $106 million, or 21%, "driven by an improved mix of products with higher royalty rates per chip, such as Armv9 technology." What the filing does not do, and what prior disclosure had done, is give Armv9 as a percentage of royalty revenue. That number is absent from both documents this quarter.

Assessment: This is the bull case working, and it is the reason to own Arm rather than a semiconductor cyclical. The escalator monetises architecture transition and design content independently of industry units, which is a structurally better position than volume leverage. The grading problem is that the two disclosed drivers, smartphone rate and data center share, are both unsized, and the one metric that would let an outsider verify the rate mechanism, Armv9 penetration of royalties, went missing in the quarter it would have been most useful.

2. Data Center: One Billion Neoverse Cores, and Still No Dollar Line

The data center story advanced on every axis except disclosure. Cumulative Neoverse cores deployed passed one billion. Google has migrated more than 30,000 applications including YouTube and Gmail, with plans to move the majority of its 100,000-plus applications, on Axion chips delivering up to 65% better price performance at up to 60% less energy. Microsoft has taken Cobalt 100 to 29 regions. Five new Stargate sites were added in the quarter, all anchored on Arm. Management framed the driver not as performance but as a physical constraint.

"One thing that's become quite evident is that power has become the bottleneck for everyone and power not only means access to energy, but everything underneath it in terms of infrastructure build-out, turbines, transformers, everything associated with generating power. So in that environment, everyone wants to move to the most efficient compute platform as possible. Arm is about 50% more efficient than competitive solutions." — Rene Haas, Chief Executive Officer

Management also made a forward assertion that is specific enough to grade: "We expect Arm's share of CPUs deployed by top hyperscalers to reach nearly 50% this year." Separately, the letter noted that customers are increasingly deploying Neoverse CPUs as head nodes alongside AI accelerators, which is an architectural claim that the accelerator build-out pulls Arm CPU content with it rather than substituting for it.

Assessment: If the power-constrained framing is right, Arm's efficiency advantage converts from a talking point into a procurement requirement, and this becomes the most durable growth vector in the business. The thesis risk is not whether it is happening but whether an outsider can size it. Two fiscal years into the Neoverse inflection there is still no disclosed data center revenue, which forces every analyst to model the largest growth driver from qualitative statements and a milestone core count.

3. The SoftBank Agreement: $177.9 Million, "Lower Margin," and No Terms

This is the quarter the parent's contribution became large enough that it can no longer be treated as a footnote. Revenue from the SoftBank Group affiliate under the consulting agreement was $177.9 million, 15.7% of total revenue, against $43.2 million a year ago and about $126 million in the prior quarter. Management quantified it and characterised it only under direct questioning.

"In terms of the impact, it was about a $50 million increase from last quarter. So last quarter, we think we were about $126 million. It actually went up $52 million, so now about $178 million. and that's a good run rate to assume going forward. The only way it would change is if we have any additional deals. And again, these are license plus design services. So think of it as being licenses to our IP to work with SoftBank on exploring solutions. But then think of the design services being effectively a kind of a funded R&D model. And so that's a lower margin revenue, of course." — Jason Child, Chief Financial Officer

Three things in that answer are load-bearing. The run rate is now declared, which makes it permanent rather than episodic. The economics are lower margin, in a business whose entire multiple rests on a 98.2% non-GAAP gross margin, and no quantification of "lower" was offered. And the duration is undisclosed: asked how long the stream persists, the answer was "we're not at liberty to say yet."

The balance sheet adds the part management did not address. Contract assets from the affiliate stood at $351.5 million at September 30 against $145.5 million at March 31, with no accounts receivable and no contract liabilities. Across the six months in which $304.0 million of this revenue was recognised, the unbilled balance grew $206.0 million. Roughly two thirds of what has been booked has not been invoiced.

Management's defence was that the relationship would be unremarkable but for the ownership: "if SoftBank wasn't a related party, we would just be booking license and design services, and it wouldn't be a related party, but then the numbers would be pretty similar." That is true as far as it goes, and it is also the precise reason the disclosure standard is higher. The counterparty sets Arm's board, and the arrangement's pricing cannot be tested by an outsider against anything.

Assessment: We are not alleging the revenue is not real; the work is presumably being done and the letter's Stargate context makes the commercial logic plausible. We are saying it is not underwritable. A $178 million quarterly run rate, roughly $712 million annualised, at an undisclosed margin, on undisclosed terms, from the controlling shareholder, with two thirds unbilled, is the largest single reason this is a Hold and not an Outperform. The remedy is entirely in Arm's hands and costs nothing: disclose the term, the scope and the pricing basis.

4. External-Customer Licence Revenue Fell 7.5%

Licence and other revenue from external customers was $222 million against $240 million a year ago. Across the six months the decline is steeper: $477 million against $681 million, down 30%. The reported licence line grew 56%; its independent component shrank.

Two honest offsets apply. First, the prior-year first half was an unusually strong external licence period, so the six-month comparison flatters the decline. Second, the Artisan physical-IP business left the perimeter on August 26, five weeks before quarter-end, and Artisan revenue sat inside licence and other. Arm did not disclose Artisan's revenue contribution, in either document, so the like-for-like basis changed mid-quarter and the decline cannot be decomposed into divestiture and demand from the outside.

Management's framing of the licence line was that it is lumpy and that the right metric is annualized contract value, which grew 28%: "As always, licensing revenue varies quarter-to-quarter due to the timing and size of high-value deals. So we continue to focus on annualized contract value, or ACV, as a key indicator of the underlying licensing trend." That is a reasonable position on quarterly noise and a less reasonable one on a two-quarter trend in a specific customer class.

Assessment: The single most useful disclosure Arm could add, and the cheapest, is Artisan's trailing revenue. Without it, the most important question in the licence line, whether third-party demand is growing or being replaced by the parent, is unanswerable for at least three more quarters. We treat the external licence decline as partly explained and wholly unresolved, and we will not give the licence line credit for organic growth until the basis is clean.

5. Arm China at 21.4% of Revenue and One of Arm's Largest Licence Deals

Arm China contributed $243.1 million under the intellectual property licence agreement, up 64.6% from $147.7 million. That is 21.4% of total revenue from a single related counterparty whose financial reporting Arm depends on. The quarterly report is explicit about that dependency: "China's payments due to us are determined based on the financial information that Arm China provides to us. Accordingly, we are dependent on Arm China providing us with reliable and timely financial information."

"In terms of the China performance, yes, it definitely has done well. And I would just overall say the demand in China looks to be as strong as we've ever seen. We did have one of our largest license deals actually come out of China. And so I would say license was slightly more of a -- I'd say, more of the overperformance came from license." — Jason Child, Chief Financial Officer

The net receivable from Arm China rose to $315.0 million from $284.6 million at the fiscal year end, and the $16.0 million credit-loss allowance against it was unchanged with no new charge this quarter, against a $16.0 million charge in the prior-year half.

Assessment: Arm China is a materially better-quality related party than the consulting agreement, because the underlying revenue is third-party Chinese licensing and royalties rather than a bilateral arrangement with the controlling shareholder. It carries a different risk: a fifth of revenue determined by a counterparty's own reporting, inside a jurisdiction the filing names as exposed to trade and national-security policy. The quarter's strength here is a genuine positive and the concentration is a standing structural risk rather than a new one.

6. The $131 Million Artisan Gain: GAAP Net Income Did Not Double

GAAP net income of $238 million, up 122%, and GAAP diluted earnings of $0.22, up 120%, were the most eye-catching numbers in the release. They are also the least informative. The reconciliation shows a $131 million pre-tax gain on the disposal of a business, recognised in other non-operating income, and excluded from non-GAAP results. That is the Artisan sale to Cadence, which completed on August 26 and produced $127 million of cash proceeds.

Removing the pre-tax gain leaves GAAP net income of $107 million, identical to the $107 million reported in the prior-year quarter. The reconciliation does not attribute a separate tax effect to the disposal, so this is a pre-tax-gain-removed figure rather than a fully tax-adjusted one, but the direction is unambiguous. The prior-year comparative is distorted in the opposite direction by a $43 million income tax benefit, where this quarter carried a $64 million tax expense. Neither period's GAAP net income is a clean read on operating progress.

The reason the two frameworks diverge so violently is share-based compensation. It was $265 million, 23.3% of revenue, all equity-classified, plus $39 million of associated employer taxes net of research incentives. GAAP operating income was $163 million against $467 million on a non-GAAP basis.

Assessment: Arm's non-GAAP presentation is internally consistent and correctly excludes the disposal gain, so the $0.39 is the number to use. The point is for readers encountering the GAAP headline elsewhere: the 122% increase is a divestiture, not earnings growth, and the underlying GAAP result was flat. Separately, compensating employees at nearly a quarter of revenue in stock is a real economic cost that the 41.1% non-GAAP operating margin does not describe.

7. The Operating Expense "Underrun" Was a Related-Party Credit Reversal

Non-GAAP operating expense of $648 million came in $7 million below the ~$655 million guide, and the CFO cited "slightly lower OpEx" as one of the two drivers of the earnings beat. The notes explain where the favourability came from.

"During the three and six months ended September 30, 2025, the Company recognized a reversal for the current expected credit loss reserve with respect to its loan receivable from Arduino by $16.8 million based on updated information and risk profile as of September 30, 2025 in selling, general, and administrative operating expenses." — Arm Holdings plc, Form 6-K, Note 14

Arduino is a related party. The reversal reduces reported expense, and it is not among the items Arm excludes from its non-GAAP measures, which are share-based compensation, employer taxes on share-based compensation, other operating income and expense, disposal items, equity-method results and their tax effects. It therefore lands directly in non-GAAP selling, general and administrative expense and in non-GAAP earnings per share.

Adjusting for it: non-GAAP selling, general and administrative expense of $182 million becomes roughly $198.8 million, growing 15.6% rather than the reported 6%. Non-GAAP operating expense of $648 million becomes roughly $664.8 million, which is about $10 million above the guide rather than $7 million below it. The earnings effect is small, roughly $14.3 million after tax at the 15.1% non-GAAP rate, or about $0.013 per diluted share, so the $0.39 would have rounded to $0.38 and still cleared the $0.37 guide ceiling.

Assessment: The earnings beat survives this adjustment; the cost-discipline narrative does not. Selling, general and administrative expense growing 6% while research and development grows 45% reads as deliberate prioritisation, and at 15.6% it reads as ordinary scaling. Anyone modelling forward operating expense off a $648 million base is starting $17 million too low. This is the cleanest example in the quarter of a pattern worth naming: the reported figures are accurate and the flattering ones keep resolving to related-party items.

8. Research and Development Up 45%, and Guided Up Again

Non-GAAP research and development expense was $466 million, up 45%, against selling, general and administrative expense up 6%. The company called it another record in research investment. Management was explicit that the spending is tied to the possible move beyond licensing.

"These investments in R&D reflect ongoing engineering head count expansion to support customer demand for more Arm technology, including continued innovation in next-generation architectures, compute subsystems, and possibly chiplets or complete SoCs." — Jason Child, Chief Financial Officer

The forward guide is the part that changes the operating-leverage picture. Third-quarter non-GAAP operating expense is guided to approximately $720 million, an 11.1% sequential increase, against revenue guided up 7.9% sequentially at the midpoint. That is the first quarter in the recent sequence where the company's own guide has expense growing faster than revenue. Working the guide through a 98.2% gross margin implies roughly 39.4% non-GAAP operating margin, about 170 basis points below the 41.1% just delivered.

Asked directly about the trajectory, the CEO offered a general assurance rather than a bounded commitment: "as you've seen on the guidance going forward, our revenue go forward is higher than our OpEx increase, which is something we've been very careful to manage." On the guided quarter that is not what the numbers say.

Assessment: We are not troubled by research and development growing at 45% in a business where the product is architecture and the addressable compute market is expanding. We are troubled by the absence of a bounded commitment. Management asserted expense discipline in language the very next quarter's guide contradicts, and the spending is attached to an optionality management will not describe. Until there is either a stated ceiling or a described product, the investment cycle is an open-ended claim on the margin structure.

9. Compute Subsystems and Lumex: What the Escalator Costs to Build

Compute subsystems are the mechanism by which the royalty rate escalator actually escalates, and the quarter's data supports it. Three new subsystem licences were signed, one each in smartphones, tablets and data centers, bringing the total to 19 across 11 companies, with five customers already shipping and one on its second generation. The expanded Samsung collaboration for Exynos means the top four Android vendors are all shipping subsystem-powered devices. Demand "continues to exceed our expectations."

Lumex, launched in September, is the flagship: up to five times faster CPU AI performance through Scalable Matrix Extensions, up to three times the energy efficiency of the prior generation, with Gmail, YouTube and AliPay named as early adopters and MediaTek designing it into next-generation chips. It is already earning royalties, which management flagged as ahead of plan and explained by the customer being on its second subsystem rather than its first.

The cost of building it was disclosed for the first time, and it is the most useful number on the call for understanding the research and development line.

"over the past 4 years, we've invested heavily in developing the technology that makes up the Lumex Compute Subsystems for smartphones, which we announced in September. This project took around 1,000 man-years with a team size peaking over 450 engineers" — Jason Child, Chief Financial Officer

The dollar figure followed in the same sentence, after a slip the speaker corrected in real time: the investment required was sized at hundreds of millions of dollars.

Assessment: This is the clearest articulation yet of the bargain at the centre of Arm's model: hundreds of millions of dollars and four years per platform generation, recovered through a higher per-chip rate across a very large installed base over many years. The economics are attractive and the disclosure is welcome. It also frames the research and development escalation as structural rather than discretionary. Each generation costs more than the last, and the payback sits in royalty years that are not yet on the income statement.

10. Remaining Performance Obligations Down 6% While Contract Value Compounds at 28%

The letter reported annualized contract value up 28% to $1,600 million and remaining performance obligations up 1% sequentially to $2,246 million. The second figure is also down 6% year on year, from $2,385 million, a decline that appears in the letter's metrics table but not in its narrative. The call cited the first metric twice and the second not at all.

The distinction matters because the two are not the same kind of number. Annualized contract value is management-defined: total annualised committed fees for active agreements, with Total Access and architecture agreements annualised over the contract term and everything else, including single-use licences and the Arm China aggregate fee, annualised over an assumed three years "based on the historical licensing patterns of our customers." The filing states it "is not reconcilable to, nor a substitute for, revenue reported under GAAP." Remaining performance obligations is the accounting disclosure of the transaction price allocated to unsatisfied obligations. One is an assumption-weighted bookings proxy; the other is contracted revenue.

Arm expects to recognise approximately 29% of the $2,246 million within twelve months, 16% in the following twelve, and the remainder thereafter. Alongside this, Arm Flexible Access licences declined by one sequentially, to 312, the second consecutive quarterly decline in that count after 314 and 313, also uncommented.

Assessment: A single quarter of divergence is a recognition-timing artifact. A year of it is a signal. Contracted future revenue going backwards while the bookings proxy compounds at twice its historical rate is the kind of divergence that resolves in one of two ways: either remaining performance obligations reverse next quarter, or the annualized contract value growth rate was never describing the same thing. We put this on the watch list at the highest priority short of the SoftBank terms, and we note that the metric management chose to emphasise is the one that cannot be tied to revenue.

11. The Silicon Question: DreamBig, Chiplets, and Three Gates

The quarter turned the move beyond licensing from a remark into a signed transaction. The subsequent-events note discloses a definitive agreement, signed in October, to acquire DreamBig Semiconductor for approximately $265.0 million in cash, expected to close by the end of the fiscal fourth quarter. The CEO described the asset as Ethernet intellectual property and controllers for scale-up and scale-out networking. In prepared remarks he restated the broader ambition.

"As mentioned in our last call, we are continuing to explore the possibility of moving beyond our current platform into additional compute to subsystems, chiplets or complex SoCs. As a result, we continue to accelerate the investment in our R&D as we are seeing increased demand from our customers for our work from Arm." — Rene Haas, Chief Executive Officer

Pressed on timing, he declined entirely: "there's nothing I can talk to you about today in terms of time line, about products or technologies." The CFO supplied the disclosure framework instead, three gates that must all clear before Arm will say anything: tape-out, samples returned, and non-cancelable customer orders. He also volunteered, in the context of the SoftBank arrangement, that a product outcome would be "to some extent, cannibalistic of whatever the current license and design services."

Assessment: Arm is buying networking silicon capability and accelerating research spending toward an end state it will not describe, and management has now said on the record that the end state may cannibalise the revenue stream currently funding it. A licensing business with a 98% gross margin that moves toward selling silicon is changing the multiple it deserves, whatever it does to revenue. The three gates are a sensible governance posture and a poor one for an investor, because they mean the business model may be reset with a single announcement, after the capital has been committed. The DreamBig purchase price is immaterial against a $169 billion market capitalisation; its signal value is not.

Guidance & Outlook

The guide is internally coherent, which is worth confirming rather than assuming. Revenue of $1,225 million at a 98.2% gross margin gives roughly $1,203 million of non-GAAP gross profit; less $720 million of operating expense gives about $483 million of operating income, or 39.4% of revenue; adding $27 million of net interest income and applying the 15.1% non-GAAP tax rate across 1,069 million diluted shares produces about $0.41. The components tie to the total as well: royalty growth "just over 20%" on the $580 million of the prior-year third quarter is roughly $700 million to $710 million, and licence growth of 25% to 30% on $403 million is $504 million to $524 million, which brackets the $1,225 million midpoint.

Running the same arithmetic backwards through the quarter just reported shows how conservatively the second quarter was set. The $1,060 million revenue midpoint at a 98.2% gross margin, less the ~$655 million expense guide, embeds roughly $386 million of operating income, or 36.4% of revenue. Arm delivered 41.1%. That is a 470 basis point margin beat against its own guide, and it is the single clearest measure of how much room the company leaves itself.

Implied sequential ramp. Revenue up 7.9% and operating expense up 11.1%. This is the first guided quarter in the recent run where expense growth exceeds revenue growth, and it costs about 170 basis points of operating margin. Management's framing, that "our revenue go forward is higher than our OpEx increase," is not what the guided numbers show for this specific quarter, and the discrepancy was not reconciled on the call.

Where the Street sits. Consensus for the reported quarter was effectively the company's own guide midpoint on both revenue and earnings, which tells you something about how this stock is modelled: the Street takes the guide. No independent published third-quarter consensus was available as of this writing, so we compare the guide against filed prior-year actuals rather than assert a gap.

Guidance style. Conservative and consistently so. Arm guides one quarter at a time, has no published full-year revenue or earnings guidance, and cleared the high end of both guided metrics this quarter after guiding royalties to mid-teens growth and delivering 21%. The pattern is a company that sets a beatable bar. The corollary is that the licence deceleration embedded in the third-quarter guide, from 56% to 25% to 30%, is likely to prove conservative too, and that a licence line management describes as lumpy should not be extrapolated in either direction from one quarter.

Analyst Q&A Highlights

Arm's Position in the Data Center Build-Out

The call opened on the gigawatt-scale capacity announcements of the prior quarter and what they mean for Arm. The answer was framed around a physical constraint rather than a competitive one: power has become the binding limit, so efficiency becomes a procurement requirement rather than a preference. Management also disclosed a perspective advantage, sitting on the parent's board with direct visibility into the largest of those projects, which is an unusual thing for a chief executive to volunteer and a useful thing for a reader to know when weighing the confidence of everything else on the call.

Q: "Rene, I wanted to ask about the AI opportunity. There's been a seemingly nonstop stream of new data center deals announced over the last quarter, calling for tens of gigawatts of additional computing capacity to be stood up. How do you feel about Arm's strategic positioning with respect to these AI deals? And what do you view as the opportunity across the build-out?"
— Sebastien Naji, William Blair

A: "As a Board member of SoftBank and also given our heavy involvement there with Stargate and regular dialog with OpenAI, I believe I have a unique perspective in terms of visibility in terms of this market. One thing that's become quite evident is that power has become the bottleneck for everyone... So in that environment, everyone wants to move to the most efficient compute platform as possible. Arm is about 50% more efficient than competitive solutions. We've seen that across the board in benchmarks, but also more importantly, in real-life performance. And that's why we see NVIDIA, Amazon, Google, Microsoft, Tesla, all using Arm-based technology."
— Rene Haas, Chief Executive Officer

Assessment: The strongest strategic answer on the call, and unusually falsifiable. If power is the constraint, Arm's efficiency edge compounds; if capacity additions slow, the same logic unwinds. Worth noting that the visibility being cited flows through the controlling shareholder, which is also the counterparty on 15.7% of this quarter's revenue.

The Rationale for Buying Networking Silicon

An analyst spotted the acquisition in the filing rather than in the letter, which is itself informative about how it was presented. The response tied the target directly to data center interconnect: Ethernet intellectual property and shipping controllers for scale-up and scale-out networking. The framing was product-portfolio breadth for end customers, which sits somewhere between licensing more intellectual property and selling more complete systems, and management did not place it on that spectrum.

Q: "I noticed in the filing you announced your intention to acquire DreamBig Semiconductor. Curious just kind of what's behind that? And how does that kind of fold into your plans to potentially expand beyond your current kind of offering platform?"
— Joe Quatrochi, Wells Fargo

A: "So DreamBig is a great company. They've got a lot of interesting intellectual property particularly around the Ethernet area and already make controllers, which are very, very key for scale-up and scale-out networking. So when we look at the demand for what's going on inside the data center and particularly in the area of high-speed communications, that type of technology will be very helpful for us to broaden our offering to end customers."
— Rene Haas, Chief Executive Officer

Assessment: A $265 million cash acquisition is immaterial to the balance sheet and material to the thesis, because it is the first capital actually committed to the move beyond licensing. Note the phrase "already make controllers." Arm is acquiring a business that ships product.

The Step-Up in Related-Party Revenue

The most important question of the call, and the one that went least answered. The question named the disclosure precisely and asked for the nature of the relationship. The response addressed the opportunity set around the parent's data center projects, spanning compute, networking, power distribution and assembly, and did not touch the agreement itself: not its term, not its scope, not its pricing, not its margin, not its recognition. The quantification came later in the call, from the chief financial officer, in answer to a different analyst.

Q: "I noticed in your disclosures that you saw a material step-up in related party revenue. So I was wondering if you could maybe talk a little bit about -- there's also been many announcements related to Stargate and SoftBank since the last earnings call. Can you maybe give us any kind of color you can on the nature of that relationship and how things are changing in terms of design activities?"
— Jim Schneider, Goldman Sachs

A: "So one of the ways to think about Stargate and particularly given the relationship between Arm and SoftBank is a huge opportunity for Arm to partner with SoftBank and SoftBank's partners to provide technology into all those solutions. So without getting into too many of the specifics, but at a high level, if you think about what's associated with building out these data centers, you have the compute, obviously, you have the networking, you have everything associated with power distribution... So as a result of all the work that SoftBank and the SoftBank family of companies are doing, it provides huge opportunity for Arm to provide solutions into that space."
— Rene Haas, Chief Executive Officer

Assessment: The question was about an accounting disclosure and the answer was about a market opportunity. "Without getting into too many of the specifics" was the whole of it. On the quarter's single largest quality-of-earnings item, asked as directly as it can be asked, the response converted a governance question into a demand narrative.

When Exploration Becomes Return on Investment

A recurring line of questioning on the call pressed the connection between rising operating expense and the undefined silicon ambition, and asked for a monitorable path. The chief executive declined to give any timeline, product or technology detail. The chief financial officer then restated the three conditions that must all be met before Arm will discuss it, which is a governance answer rather than an investment one: it tells shareholders when they will learn, not what they are funding.

Q: "I wanted to go back to the OpEx side of things... You mentioned about exploring different sorts of go-to-market methodologies, chiplets, et cetera. When do you expect to give us more color on when that's going to go from exploration to return on investment or the actual strategy, how should we monitor that and expect to get more information from you?"
— Ross Seymore, Deutsche Bank

A: "The only thing I would add is, I think last quarter, we said, as soon as the way we think about when we announce something, if it were to be something related to full SoCs, it would be once there's tape-out, once there's samples back and once there's actually noncancelable customer orders, when we achieve all 3 of those milestones, that's when we would probably talk about something because this would be a new business and something we haven't done before."
— Jason Child, Chief Financial Officer

Assessment: The three gates are a real commitment and a late one. Non-cancelable customer orders means the capital is spent, the silicon exists and the customers are contracted before the market is told the business model has changed. Investors are being asked to fund an accelerating research budget toward an outcome they will learn about only once it is irreversible.

Sizing and Durability of the Parent-Funded Revenue Stream

This exchange produced the quarter's most valuable disclosure, and it came from a question that asked four things at once: the size, the forward assumption, the duration, and whether a move into physical product would cannibalise the stream. The answer quantified the first two precisely, refused the third, and conceded the fourth. It also introduced the characterisation that reframes the margin discussion: the design-services half is funded research and development at a lower margin.

Q: "I just wanted to clarify how much was the SoftBank contribution in Q2 versus what you thought? And then what is baked in for Q3 and hopefully, if you have the number for Q4? And the real question is how long can this quarterly rate persist? And if you do move into physical chips or chiplets or any other products as part of target, does it start to cannibalize this licensing stream?"
— Vivek Arya, Bank of America

A: "In terms of the impact, it was about a $50 million increase from last quarter. So last quarter, we think we were about $126 million. It actually went up $52 million, so now about $178 million. and that's a good run rate to assume going forward... And again, these are license plus design services... But then think of the design services being effectively a kind of a funded R&D model. And so that's a lower margin revenue, of course. So these -- in terms of how long these revenue streams will occur, we're not at liberty to say yet... And yes, I would think of that as being, to some extent, cannibalistic of whatever the current license and design services."
— Jason Child, Chief Financial Officer

Assessment: Credit where due. This is a candid, quantified, self-aware answer that gave analysts more than the letter did, and more than the earlier response to the same subject. It also confirmed every element of the concern: permanent run rate, lower margin, undisclosed duration, and cannibalistic of the existing licence economics if the product ambition lands. The only thing still missing is the thing only Arm can supply, which is the agreement's terms.

The Shape of the Parent's Data Center Opportunity, and an Early Lumex Royalty

A two-part question sought a one-, three- and five-year revenue shape for the parent-linked opportunity and clarification on the claim that a September-launched platform was already earning royalties. No forecast was offered on the first; the reply argued instead that the demand picture is larger now than when the $500 billion programme was announced eleven months earlier. The second half was the more concrete: the early royalty came from a customer on its second compute subsystem rather than its first, which shortened the design cycle to a couple of months after delivery.

Q: "Just following on, on the Stargate theme and the sites. Can you maybe just talk about the shape of what that revenue opportunity looks like on a sort of 1-, 3- and 5-year view just kind of when it's going to start having an influence on the revenue -- the annual revenue or quarterly revenue of the business?"
— Timm Schulze-Melander, Rothschild & Co Redburn

A: "Without giving you kind of a go-forward forecast of 1, 3, 5 years, maybe a way to think about it is, back in January of this year, OpenAI with Oracle and SoftBank announced Stargate, which was a $500 billion project to build out data centers over the next number of years. When we go back to where we are now 11 months later, I would say the demand picture for compute is greater than it was at that time... So if nothing else, I think the opportunity for compute has only grown since we made that Stargate announcement."
— Rene Haas, Chief Executive Officer

Assessment: The pattern of the call is now clear. Questions about how the largest new revenue source will behave are answered with the size of the market it sits inside. The Lumex royalty detail in the second half of the exchange was genuinely good news and the better-supported half of the answer: a second-generation subsystem customer converting to royalties within a couple of months of delivery is direct evidence that the adoption flywheel shortens each cycle.

Cloud and Networking as a Share of Royalty Revenue

The only exchange that put a number on the data center business. The question supplied its own estimate, noted that the figure had been roughly 10% of royalties at the last fiscal year end, and proposed 15% to 20% now. Management accepted the trajectory rather than stating a figure, promised a full update at the fiscal year end, and added that it is running ahead of internal expectations. This is how the most important growth driver in the company is currently disclosed: an analyst's range, confirmed as "not a bad assumption."

Q: "So last fiscal year, cloud and networking accounted for about 10% of royalty revenues. We're midway through this fiscal year. Maybe you guys could just true us up, I assume, this mix has increased. Is it approaching 15%, 20% of total royalty revenues for the team? Any color here would be great."
— Harlan Sur, JPMorgan

A: "Harlan, so on the royalties, yes, I mean, it ended the year at around 10-ish percent. And so we're certainly with the growth rate in infrastructure being double, I'd say, all the other categories in overall average royalty, you should expect it to continue to increase. We'll provide a full update at the end of the year. But your trajectory of somewhere in the 15% to 20% range is not a bad assumption and probably a reasonable expectation for where we expect to trend throughout the year."
— Jason Child, Chief Financial Officer

Assessment: Useful and insufficient. Against the $620 million royalty line, 15% to 20% implies something in the order of $93 million to $124 million a quarter from cloud and networking, which is the only way to put a dollar figure on the second bull pillar and it rests on a range the company confirmed rather than published. A reported data center revenue line would resolve this at a stroke, and the promise of a full update at the fiscal year end is now a gradeable commitment.

China Strength and the Licence Pipeline

The closing question pressed on China's 22% revenue share and on whether the licence pipeline supports confidence into the seasonally important fourth quarter. The answer disclosed that one of the company's largest licence deals came out of China, and that licence rather than royalty drove the regional outperformance. On the pipeline, management described a strong remainder of the year while explicitly declining to characterise the fourth quarter, and offered a structural point about deal behaviour instead.

Q: "I see China is maybe 22% of sales this Q. And I was just wondering what is driving that? Is it more licensing or royalties for strength in the quarter? And maybe just as you look at the licensing pipeline for the rest of the year, have you seen more reason to be confident in the growth this year for licensing, especially as you look to Q4, which, as I believe we said before, there's potential for good renewal deals this year."
— Lee Simpson, Morgan Stanley

A: "In terms of the China performance, yes, it definitely has done well... We did have one of our largest license deals actually come out of China... But as a reminder, we don't -- the deal cycles on large license deals are usually 6 to 9 months, and we don't really lose deals. It's really just about what exactly are the market needs for customers and when do they need it."
— Jason Child, Chief Financial Officer

Assessment: "We don't really lose deals" is a strong claim about competitive position and consistent with the absence of a credible architecture alternative at the high end. It also relocates the licence line's variability entirely to timing, which is the correct framing for quarterly noise and a convenient one for a two-quarter decline in external-customer licence revenue. The confirmation that China's outperformance was licence-led sits alongside the 64.6% increase in Arm China revenue, and both are third-party demand routed through a related party.

What They're NOT Saying

1. The SoftBank agreement's term, scope, pricing basis and recognition pattern. Four quarters into an arrangement now running at roughly $712 million annualised, none of it is disclosed. The shareholder letter does not mention SoftBank at all. The amount and a forward run rate came out only under questioning, and the duration was answered with "we're not at liberty to say yet."
2. Armv9 as a percentage of royalty revenue. Absent from both documents, in a quarter whose central claim is that royalties grew on Armv9 and subsystem rate mix. This is the disclosure that would let an outsider verify the rate escalator directly, and it was the single most useful number Arm had been providing.
3. Data center royalty revenue in dollars. Neoverse royalties rose to "more than double year-on-year", and the chief financial officer put infrastructure royalty growth at about double that of the other categories. It has never been sized. The only quantification this quarter was an analyst's 15% to 20% range that management called "not a bad assumption."
4. Artisan's revenue contribution. The physical-IP business left on August 26, inside the licence line, five weeks before quarter-end. Without its revenue, the 7.5% decline in external-customer licence revenue cannot be split between divestiture and demand, and no organic licence growth rate can be computed for at least three more quarters.
5. The year-on-year decline in remaining performance obligations. Down 6%, to $2,246 million. The letter's narrative reported the sequential increase of 1%, and the annual decline appeared only in its metrics table. The call cited annualized contract value twice and remaining performance obligations not once. No analyst asked.
6. The sequential decline in Flexible Access licences. Down one, to 312, the second consecutive sequential fall in the count, after 314 at the fiscal year end and 313 in the first quarter. Reported as a data point in the letter with no commentary, and not raised on the call.
7. How much lower "lower margin" is. The chief financial officer volunteered that the design-services component of the parent revenue is funded research and development at a lower margin. In a company whose valuation rests on a 98.2% non-GAAP gross margin, the magnitude is material and was not given.
8. Any capital expenditure guidance. Capital expenditure was $138 million against $53 million a year ago, up 160%, and net property and equipment rose 61.9% in six months to $573 million. A company exploring complete systems-on-chip has an obvious reason for rising capital intensity and offered no forward figure.
9. What the accelerating research budget is actually buying. Research and development grew 45% and is guided up again, explicitly attributed in part to "possibly chiplets or complete SoCs," a programme management declined to describe in any respect, including whether it exists.
10. The reconciliation between the stated expense discipline and the guide. Management said revenue growth exceeds operating expense growth going forward. The third-quarter guide has expense up 11.1% sequentially against revenue up 7.9%. The contradiction went unremarked by management and unchallenged on the call.
11. The credit-loss reversal inside the operating expense beat. The $16.8 million Arduino reversal is disclosed in the notes and was not mentioned in the letter's financial overview, in the prepared remarks, or in the discussion of operating expense landing below guide.

Market Reaction

• Pre-print setup: The shares closed at $160.19 on November 5, up 29.9% year to date against 15.6% for the S&P 500, up 13.9% over twelve months and 2.5% over thirty days. They entered the print weak, having fallen 7.4% across the prior six sessions from $173.09 on October 28, including a 4.7% decline on November 4, the session immediately before the release. The 52-week closing range was $85.82 to $179.93.
• Reaction session, November 6: Opened at $167.00, a 4.3% gap up, touched $167.48 for a 4.6% high, then sold down to $155.18 at the low, 3.1% below the pre-print close, and closed at $158.25, down 1.2%. High to low is a 7.3% round trip inside one session, on 7.1 million shares against a 5.1 million 30-day average, or 1.4 times normal volume.
• Relative: The S&P 500 fell 1.1% the same day, so the close was roughly flat against the market. The move that matters is not the 1.2%, it is the complete surrender of a 4.6% gap.
• Position versus range: The close leaves the shares 12.0% below the 52-week closing high and 84.4% above the 52-week closing low.

The opening gap tells you the initial read on the release and the call was positive, and it should have been: revenue and earnings both cleared the guided high end, royalties beat a mid-teens guide by six points, and the third-quarter guide implied 25% growth. That the entire gap was sold inside a single session, on elevated volume, in a market that was itself down about a point, is the more informative fact.

Our read is that the fade is about composition, not about the print. The headline is a 34% growth rate; the filing published alongside it shows related parties at 37.2% of revenue, licence revenue from external customers down 7.5%, and a GAAP earnings increase that is a disposal gain. The third-quarter guide then asks for operating expense growth of 11.1% against revenue growth of 7.9%, which is the first guided quarter in the recent sequence that spends operating leverage rather than banking it. Investors who read past the first two lines found a quarter whose quality is harder to hold than its headline, and a stock at roughly 93 times forward earnings has no cushion for that discovery.

The second contributor is positioning. The shares had already given back 7.4% in six sessions into a print they were expected to beat, and a business that has cleared its own guide repeatedly attracts owners who need more than a clearance to add. A beat against a consensus that simply equals the company's guide midpoint is not new information, and it is not a catalyst.

Street Perspective

Debate: Is the Royalty Rate Escalator Strong Enough to Carry the Multiple on Its Own?

Bull view: Royalties are 54.6% of revenue, grew 21% on rate and content rather than units, beat a mid-teens guide, and are compounding while the handset market does nothing. Compute subsystems are at 19 licences across 11 companies with the top four Android vendors shipping, and data center royalties are doubling annually off a base heading from 10% toward 20% of the royalty line. On this view the licence line is noise around a royalty annuity that grows at 20% for years, and the composition debate misses the asset.

Bear view: The royalty line is excellent and it is only half the company. The other half grew 56% on the strength of the controlling shareholder while its third-party component shrank, and at roughly 93 times forward earnings the multiple is priced on the consolidated growth rate, not on the royalty growth rate. Royalties compounding at 20% do not justify 93 times when 15.7% of revenue is parent-funded at an undisclosed margin and might be cannibalised by the company's own product ambitions.

Our take: The bulls are right about the asset and the bears are right about the price. The royalty franchise is genuinely worth a premium multiple, and the correct question is what premium a business deserves when its reported growth rate cannot be reproduced from its independent revenue. External-customer revenue grew 9.4% in the quarter and declined 2.0% across the half. Underwriting 93 times against 34% is a different act from underwriting it against 9%, and only one of those two numbers describes demand Arm won in the open market.

Debate: Is the Related-Party Revenue a Governance Discount or an Underappreciated Growth Channel?

Bull view: Arm is being paid to co-design the compute layer of the largest data center programme ever announced, and the parent's involvement is how it gets that seat. Management's point is fair: strip the ownership link and this is ordinary licence and design-services revenue that nobody would query. A declared run rate, growing, with the option of additional statements of work, is a channel the market is discounting because of an accounting label.

Bear view: The counterparty appoints the board, the terms are entirely undisclosed, the margin is admitted to be lower and unquantified, two thirds of the revenue recognised has not been billed, and the duration answer was a refusal. It also arrived in exactly the period when external-customer licence revenue went backwards, which makes it look less like incremental growth and more like substitution.

Our take: The bull argument is coherent and would be persuasive with disclosure attached. It has none. There is no term, no scope, no pricing basis, no renewal mechanic and no margin, and a $351.5 million unbilled contract asset with zero receivables against a $304 million half-year revenue recognition. We are not asserting the revenue is not real. We are saying that at 15.7% of revenue it has become too large to carry on management's characterisation alone, and that the cost of fixing this is a paragraph in the next shareholder letter. Until it appears, the appropriate response is a discount, and the fact that the chief executive answered a direct question on it with a market-opportunity narrative reinforces rather than relieves that view.

Debate: Does the Move Toward Silicon Create Optionality or Destroy the Multiple?

Bull view: Arm sits at the centre of the most valuable compute transition in a generation and is being pulled up the stack by customers who want more complete solutions faster. Subsystems already prove the pattern: more content, higher rate, stickier customers. A $265 million networking acquisition is a rounding error against a $169 billion market capitalisation and buys real Ethernet capability. If the chiplet work lands, the addressable revenue per design multiplies.

Bear view: Every step up the stack trades gross margin for revenue. A 98.2% gross margin licensing business that begins selling silicon is a different business with a different multiple, and management has now said a product outcome would be "to some extent, cannibalistic" of the revenue currently funding the work. Research and development is up 45% toward an end state shareholders will be told about only after tape-out, samples and non-cancelable orders, which is to say after the capital is irrecoverable.

Our take: The bears have the better of this, on the specific ground that the disclosure sequence is backwards. We have no objection to Arm pursuing complete systems, and the strategic logic of moving where the value is accumulating is sound. The objection is that shareholders are being asked to fund an accelerating expense line toward an unnamed product, on the explicit understanding that they will learn what it is only once it cannot be reversed, in a business whose entire valuation premium rests on the margin structure the change would dilute. The three gates protect commercial confidentiality at the cost of informed consent.

Debate: Valuation, Is Any Entry Point Visible After a 30% Year?

Bull view: The shares are 12.0% below their 52-week closing high and have de-rated while forward earnings rose, so the multiple has done some work already. With data center royalties more than doubling and royalties compounding in the twenties, a 20%-plus grower with a near-monopoly architecture position and no capital intensity to speak of does not get cheap, and waiting for a conventional multiple means never owning it.

Bear view: Roughly 94 times trailing and 93 times forward non-GAAP earnings, 34.3 times enterprise value to forward sales, and a 0.63% free cash flow yield, on earnings that exclude share-based compensation running at 23.3% of revenue. On GAAP, this quarter's operating margin was 14.4% and net income excluding a disposal gain was flat. Nothing in that set of numbers leaves room for the composition questions to resolve badly.

Our take: We agree that Arm will not be available at an ordinary multiple and we are not asking for one. We are asking to be able to reproduce the growth rate the multiple is paying for, and at present we cannot. That is why this is a Hold rather than an Underperform: the franchise is real, the royalty escalator is confirmed, and the risk is concentrated in disclosure gaps that Arm can close whenever it chooses. It is also why it is not an Outperform: at 93 times forward earnings, an investor is paid nothing for waiting to find out.

Model Update & Valuation Framework

This is initiation of coverage, so the table below sets our opening estimates rather than revising prior ones. Company guidance is used for the third quarter of fiscal 2026. Arm does not publish full-year guidance, so full-year figures are ours.

Valuation. At the November 6 close of $158.25 and 1,069 million diluted shares, Arm carries a market capitalisation of approximately $169.2 billion. Cash and equivalents of $2,518 million plus short-term investments of $740 million give $3,258 million of net cash, with no borrowings line on the balance sheet, for an enterprise value near $165.9 billion. Against trailing-twelve-month non-GAAP diluted earnings of $1.68 that is 94.2 times; against our fiscal 2026 estimate of $1.70 it is 93.1 times. Enterprise value is 34.3 times our fiscal 2026 revenue estimate. Trailing-twelve-month non-GAAP free cash flow of $1,073 million, itself up 126%, is a 0.63% yield on the equity. On GAAP, the trailing multiple is not a meaningful figure: this quarter's GAAP operating margin was 14.4% and share-based compensation ran at 23.3% of revenue.

Valuation impact. We are not publishing a price target with this initiation. Two of the three inputs a target would need are withheld by the company rather than uncertain in the market: the terms of a revenue stream that is 15.7% of the total, and the existence and economics of a product programme management has confirmed it is spending against but declines to describe. A third, the organic licence growth rate, is unobtainable until the Artisan divestiture laps. We would rather set a target on the other side of those disclosures than anchor one to a revenue mix and a business model that management has signalled are both in motion.

Thesis Scorecard Post-Earnings

This is initiation of coverage, so the table below establishes the pillars we will grade each quarter from here rather than scoring a standing thesis. Status tags follow our convention: bull pillars run ON TRACK, AT RISK or BROKEN; bear points run CONTAINED, EMERGING or MATERIALIZING.

Overall: Thesis established. Both verifiable bull pillars are confirmed on reported numbers in the same quarter that all four risk points are confirmed or opened, which is the defining tension of this initiation. Arm is a structurally advantaged franchise whose royalty engine is doing precisely what the bull case requires, reporting a growth rate that cannot be reproduced from revenue it won in the open market.

Action: Hold. Initiating at Hold rather than Outperform on price and disclosure, not on business quality. The upgrade path is specific and inexpensive for the company: disclosure of the SoftBank agreement's term, scope and pricing basis; a reported data center revenue line; reinstatement of Armv9 as a share of royalties; or a reversal in remaining performance obligations alongside continued royalty outperformance. A de-rating toward the low seventies on forward earnings would also do it without any new information. The path to Underperform would require the royalty line to decelerate below the guided low twenties while the parent-funded revenue plateaus and the licence line's external component keeps contracting, which would leave a business growing at single digits on an unchanged multiple.