With the end of testimony on Day 3 of the trial, the only items remaining were the closing arguments, instructions for the jury, and jury deliberations. A verdict was not reached after several hours, so jury deliberations will continue today. In addition, testimony for the countersuit about Arm’s use of Nuvia design suggestions was not completed and will continue after the holiday break, which would be on or after court resumes on January 3, 2025. If the jury does not reach a verdict today, while not confirmed, it is assumed that deliberations will continue on January 3rd. So, I am taking this opportunity to cover the Day 4 proceedings, as well as to reflect on some of the insights that have arisen through the trial.

Disclosure: My company, Tirias Research, has consulted for Arm, Qualcomm and other companies mentioned in this article.

The Jury Is Out

Procedurally, Arm, the plaintiff, made its closing arguments, followed by Qualcomm, the defendant. An Arm rebuttal followed the closing arguments, and Qualcomm was allowed to provide a brief rebuttal because the third question being decided by the jury is related to Qualcomm’s counterclaim. The three statements the jury was instructed to rule on were (note that these are summarized):

  1. Did Arm prove that Nuvia breached Section 15.1(a) of the Nuvia architecture license agreement (ALA)? Section 15.1a is the termination clause of the agreement that requires the use of any Arm technology to stop and all related materials and designs to be destroyed.
  2. Did Arm prove that Qualcomm breached the Nuvia ALA?
  3. Did Qualcomm prove that Qualcomm CPUs are covered under the Qualcomm ALA? Qualcomm’s counterclaim

Arm’s closing argument focused on the Nuvia ALA agreement. Arm went line by line through several sections of the Nuvia ALA agreement and matched it with expert testimony, primarily from Qualcomm and Nuvia staff. Arm also made arguments against Qualcomm’s claims made throughout the trial.

Qualcomm’s closing argument was a more impassioned argument about what Arm is requesting, the destruction of intellectual property (IP) that took considerable engineering effort to develop, and a timeline of events relative to the legal actions and the Nuvia and Qualcomm CPU designs. Qualcomm argued that the designs and information transferred to Qualcomm at the time of the transaction were Nuvia IP that only included only 1% of Arm IP, which were the opcodes that link the hardware microarchitecture to the Arm software instruction set architecture (ISA). Qualcomm also argued that these opcodes are regenerated regularly, and the original code was not included in the design at the time the termination notice was received or future Qualcomm CPU designs that began after the acquisition. Qualcomm also argued that the Qualcomm CPU designs are covered by the Qualcomm ALA, which is the basis of the counterclaim. Qualcomm also used testimony, primarily from Arm, to support its argument.

Both sides also used very select testimony to try to undermine the credibility of some of the expert witnesses and demonstrate the opposing party’s intent in their closing arguments.

Immediately after closing arguments, the judge read the jury instructions, and the jury was sequestered under guard. As there is little to do while waiting for a verdict, this is a good time to highlight some of the insights that have arisen from the trial.

Insights From The Trial

The most interesting thing about a civil trial between two companies is the confidential information that is made public, and this case is no different. Some of the information brought into public view included details of Arm contracts, Qualcomm roadmaps and design information, and behind-the-scenes communication of the parties through letters, emails, internal presentations, and chats.

The following are some of the insights gleaned from the testimony and documents:

  • Arm has, or at least had, an official program called Picasso to try to raise royalty rates. The rate increases to Qualcomm, Arm’s largest licensee, were in the 300% to 400% range. As a result, it would be assumed that Arm is trying to get similar increases from other licensees.
  • Qualcomm started four CPU design efforts after the Nuvia acquisition: Hamoa for computing (primarily laptop PCs), Pakala for mobile (primarily smartphones), Nordschleife for automotive, and Pegasus for future computing and mobile applications. The Hamoa and Pakala-based SoCs are shipping now.
  • There was disagreement within Arm about how to react to the acquisition of Nuvia. Some of the old guard, including then CEO Simon Segars and Chief Architect Richard Grisenthwaite, took a more supportive stance, and others took a combative approach.
  • Arm appeared to act aggressively in its dealings with Qualcomm following the acquisition of Nuvia. This included pushing for the assumption of the Nuvia ALA, renegotiation of the Qualcomm ALA, and a campaign to 37 current and potential Qualcomm customers with inaccurate statements and information. This “pressure” on Qualcomm was promoted even by the Chairman of the Board, Masayoshi Son.
  • Qualcomm appeared to operate in good faith through and after the Nuvia acquisition, including paying royalties in the Hamoa and Pakala-based chipsets. Arm, however, has returned the funds. There was no evidence presented that Qualcomm tried to deceive Arm. Qualcomm appeared not have direct access to the Nuvia ALA prior to closing the acquisition and had no reason to believe that it was significantly different from the Qualcomm ALA, which provides broad coverage of any custom IP developed by or for Qualcomm and no requirements tied to acquisitions of Qualcomm.
  • Qualcomm’s ALA license extends through 2028, and the company has the option to extend it another 5 years for an additional fee of $1M/year.
  • Arm has been considering building complete processors for some time and has discussed building a processor for Samsung. This is presumably to compete against internal development projects from its ALA licensees. It is presumed that Arm would not sell chips in the open market because they would compete directly with Arm’s TLA customers, the most profitable portion of Arm’s business.
  • Arm did not indicate any harm to its business until just before the trial, and at that point, the judge would not allow it to be included in the trial. According to Arm, there was no harm until units began shipping, which refers to the Hanoa-based Snapdragon SoCs for PCs and Pakala-based Snapdragon mobile SoCs for next-generation Android smartphones.
  • The Nuvia ALA appears to have conflicting clauses and atypical requirements that ultimately engendered the trial with Qualcomm. Had the Nuvia ALA been clearer with common terms and conditions, the potential for many of the disputed points would have been greatly reduced.

With the relationship between the two companies deteriorating, Qualcomm may have to consider using another architecture before its ALA expires in 2033. Even if Arm offers a new license, the terms are likely to be less favorable given the recent history between the two companies and the goals of Arm’s management. Note that Qualcomm does have a license to use the latest Armv9 architecture.

Analyst Insights

Thus far, I have sought to keep my own analyses and opinions out of my reports on the trial, but there are a few points worth making.

  • I am not an attorney, but I am a bit surprised that the validity of the “right to approve an acquisition clause” was not questioned or deemed unenforceable. That clause could impact the valuation and operation of the business, especially a startup, with no value to the company itself. According to the testimony, certain provisions, which I am assuming this was one of them, were granted to Arm for a reduced license fee. However, according to the testimony, Nuvia paid what it thought was the full price for a license fee.
  • Another point that was missing in the testimony, or I missed it, was that a microarchitecture is not hard-coded to an ISA and vice-versa. While it may require some optimizations to the microarchitecture (the silicon design), you can switch from supporting one ISA to another and there several industry examples that could have been sighted.

Final Thoughts

As a technologist, I can appreciate the challenge of this case. It is more than just about the terms of the contract, it is fundamentally tied to the related technology and chip designs. In this case, attorneys are presenting a case using technical experts and attempting to communicate their technical knowledge to a group of people that probably have little to no technical knowledge. As a result, both parties relied on analogies, Arm leveraged a piano analogy, and Qualcomm leveraged a building analogy. Neither was completely accurate or effective, but any analogy is subject to interpretation.

As I have maintained, the outcome of this case will depend upon how the jury interprets the contract and the information presented, and it is impossible to predict it. However, regardless of how the scales of justice tip in this case, it is likely to impact technology license agreements going forward and cause concern throughout the broader Arm ecosystem about potential competition with Arm, especially for ALA licensees. It could also accelerate development of alternative ISA ecosystems such as RISC-V which are already starting to gain momentum.

Share.
Exit mobile version