This article is written in conjunction with Kevin Hein of Tirias Research, who is attending the trial proceedings.

The dispute over Arm v10 is revealing a larger question about who holds the leverage when one company controls technology another needs.

The second day of Qualcomm’s trial against Arm in Wilmington started to expose the issue sitting underneath many of the individual claims in the case: leverage. On the surface, testimony moved through what seemed like separate subjects. Qualcomm and Arm discussed future architecture licenses, royalty rates, processor verification, technical support and RISC-V. But looking at these in the context of leverage, these issues increasingly appear connected.

Qualcomm depends on Arm technology across important parts of its product portfolio. Arm depends on Qualcomm and other semiconductor companies to turn its architecture into products and generate licensing and royalty revenue. For years, that relationship helped create one of the industry’s most successful technology ecosystems.

Now this trial is providing an unusual look at what happens when the interests of those companies stop lining up, or even worse, start to conflict.

One of the most important subjects in Day 2 was Arm v10, the next major generation of the Arm architecture. Qualcomm already licenses Arm v8 and v9 under its Architecture License Agreement, or ALA. The dispute is over extending that agreement to v10 and whether Arm negotiated with Qualcomm in good faith. That issue is being considered in a bench trial running alongside the jury trial.

Qualcomm began asking about v10 years ago. Evidence discussed in court showed Qualcomm raising the issue in 2020 and asking Arm about its plans for v10. Lynn Couillard, Arm’s account manager responsible for Qualcomm at the time, testified that she pursued the requests internally but did not receive a response.

The timing is important. Evidence presented by Qualcomm shows Arm entered into a v10 license agreement with Microsoft in 2020 and another with Google in 2021. Qualcomm argues that Arm was therefore negotiating v10 access with other major customers while Qualcomm was trying to determine when it could begin its own negotiations.

An email admitted into evidence provided a particularly revealing look at the internal discussion. On May 5, 2020, then-Arm Vice President of Sales Todd Lepinski wrote:

“So what is the advantage to negotiate now vs later for v10? Only the fact that we have to negotiate in good faith, vs in the future we have the option to negotiate in bad faith?”

Couillard rejected the idea that the email reflected Arm’s actual negotiating strategy. She described Lepinski as being sarcastic and said the suggestion that Arm would negotiate in bad faith was so ridiculous that it was “obviously a joke.”

But the email does matter because good-faith negotiation is precisely one of the issues before the court. There also appears to be no simple contractual deadline requiring the v10 negotiations to begin or conclude by a particular date. Qualcomm can continue building products using v8 and v9 today. Technology, however, continues to move forward.

That changes the economic balance over time. As v10 becomes relevant to future processors and markets built around the Arm software ecosystem, Qualcomm eventually has to reach an agreement with Arm if it wants access to the new architecture. Due to this dynamic, any delays in beginning the negotiations with Qualcomm have a higher potential of strengthening Arm’s negotiating position rather than weakening it.

The Economics Behind The Fight

The v10 dispute also cannot be separated from a larger disagreement over Arm’s licensing economics. Arm argues that the semiconductor market has changed dramatically since Qualcomm’s current agreement was signed in 2013. Its attorneys have questioned whether licensing economics established around smartphone processors more than a decade ago should apply to today’s much larger processors and higher-value markets.

Qualcomm, however, argues that it is not demanding that Arm simply freeze 2013 pricing. Testimony Tuesday indicated that Qualcomm offered to pay as much as 5% above the best v10 rate Arm gives another licensee, with different rates possible for mobile, wearables, compute, automotive and data centers.

The courtroom also heard comparisons between Qualcomm’s existing economics and those of other major Arm licensees. Those agreements are not necessarily directly comparable, but the discussion illustrates how dramatically architecture-license economics can vary among Arm customers.

Arm CEO Rene Haas’s testimony added another important piece. He explained that earlier internal references to “unwinding ALA” meant revising ALA rates rather than abandoning architecture licenses. He also testified that Arm wanted ALA and Technology License Agreement economics to become more similar.

That puts the v10 negotiations in a broader context. Arm wants to capture more of the value created by its architecture as that architecture moves into PCs, data centers, automotive systems and other higher-value markets. Qualcomm argues that Arm cannot use control over the next architecture to escape existing contractual obligations or negotiate a replacement in bad faith.

This makes the disagreement not simply about what v10 is worth. It is about who has the power to determine that value.

Control Extends Beyond The ISA

The same leverage question appeared in the technical fight over Arm’s Architecture Compliance Kit, or ACK. The ACK helps companies verify that custom processors correctly implement the Arm architecture. Qualcomm argues that Arm stopped providing certain partner-specific out-of-band (OOB) updates, patches and engineering assistance after the Nuvia dispute escalated.

Arm witnesses argued the underlying ACK remained available and quarterly releases continued and that the additional OOBs, patches and engineering support were not mandatory contractual deliverables.

Qualcomm’s questioning focused on what happens in practice when engineers encounter a compliance problem. Its witnesses argued that design-specific OOBs and patches can be important to determining whether the problem lies in the processor or in the compliance tools themselves.

That creates a question extending beyond this particular contract: When does withholding support effectively become withholding part of the product? Again, the issue is the ecosystem surrounding the architecture, not merely the instruction set itself.

RISC-V Is The Alternative — Eventually?

RISC-V provides the obvious mitigation to Qualcomm’s dependence on Arm.

Arm highlighted Qualcomm’s own plans for RISC-V during Tuesday’s testimony. Those plans showed the open architecture moving from areas such as wearables toward automotive, data center, smartphones and other higher-performance markets over time.

But the testimony also highlighted the difference between having an alternative instruction set and having an alternative ecosystem.

Operating systems, applications, development tools, software compatibility and years of engineering investment all matter. Replacing an ISA is much easier than replacing everything that has grown around it.

That creates an interesting situation for Arm. Arm benefits in this case from demonstrating that Qualcomm has alternatives. A credible RISC-V weakens the argument that Qualcomm has little practical choice but to remain dependent on Arm.

But the stronger that argument becomes, the more RISC-V looks like a credible long term competitor to Arm itself. Haas’s testimony added another dimension. He acknowledged displeasure with Qualcomm’s promotion of RISC-V and testified that the “best Arm chips should come from Arm.”

As Arm moves beyond supplying architecture and CPU designs toward delivering more complete silicon itself, the company wants to capture more of the value created by its architecture while some of its largest customers are also trying to control more of their own technology.

Haas’s v10 testimony eventually reached an unusually sensitive point: which customers already have v10 licenses. Haas did not want those customer relationships disclosed publicly because the information was commercially sensitive. Judge Maryellen Noreika cleared the courtroom and sealed that portion of the proceedings before the testimony continued.

Other than to satisfy curiosity, there is no reason to even speculate about what was said behind closed doors. What happened publicly was revealing enough.

Arm wants to capture more of the value created by an ecosystem that has expanded far beyond the smartphone market of 2013. From a business perspective, that goal makes sense. Arm’s market capitalization has roughly tripled in just the past two years, rising from around $100 billion to more than $300 billion. Evidence presented in court showed the company targeting a roughly $2 trillion valuation by 2031. Getting there requires Arm to capture substantially more of the economic value generated by its architecture as it expands into PCs, data centers, automotive systems and other markets.

Qualcomm wants assurance that its long-standing architecture rights cannot become leverage for dramatically different economics every time the architecture advances. Qualcomm is hardly a small company, with a market capitalization approaching $200 billion, but that is what makes the imbalance so striking. Even a semiconductor company of Qualcomm’s scale has limited options when negotiating with the owner of an architecture and ecosystem it has spent decades building products around.

RISC-V gives Qualcomm another path. Custom CPU development gives it more control. But neither instantly replaces decades of software, tools, applications and engineering built around Arm. That ecosystem is what makes the Arm architecture so valuable. It is also what gives Arm leverage.

After two days in Wilmington, the larger question emerging from the courtroom is becoming clearer: How much negotiating power does a customer really have when walking away means leaving an ecosystem it spent decades helping to build?

Tirias Research tracks and consults for companies throughout the electronics ecosystem from semiconductors to systems and sensors to the cloud. Members of the Tirias Research team have consulted for IBM, Nvidia, Qualcomm, AMD and other companies throughout the mobile, data center, AI and Quantum ecosystems.