The merge of the Qualcomm architecture with the Nuvia IP they acquired, which was created under a far-reaching license ARM granted to Nuvia. Combining both creates a custom architecture different from ARMs consolidated and harmonized designs offered to licensees (i.e. Blackhawk or Cortex-X).
The IP of Nuvia was not supposed to be used in all the use-cases that Qualcomm intends to deploy it in (and moreover there is still the ongoing legal dispute that Qualcomm is actually not allowed to use it)
The wrinkle is that nuvia's IP supposedly wasn't used here. Qualcomm set up an IP firewall during the acquisition and immediately switched the team over to doing new work under the existing Qualcomm license. It's about the right timeframe for the v2 chips at the very least.
That's not Qualcomm's position in court. Do you have any legitimate source for that statement?
Also, the foundation of Qualcomm's "Oryon" is clearly Nuvia's "Phoenix" core, which is based on Arm’s v8.7-A ISA.
After Acquisition, Qualcomm formed a team to redesign Phoenix for use in consumer-products instead of servers, creating Oryon.
That's the issue they have. Qualcomm was/is confident to resolve this IP issue of the technical QCT-division via their licensing strong-arm QTL, forcing ARM into accepting Qualcomm's view.
However, they possibly overstepped a bit, as they also expect that they don't need to license newer CPU-designs from ARM because (like Apple) they built a custom design under their architecture license.
But in reality the core design of Oryon was in parts built under the license agreement of Nuvia, which has explicit limitations in transferability (only Nuvia as-is) and usage (only servers).
In court, Qualcomm doesn't even dispute that, they argue that this contract should not be enforced and hope that the court agrees.
In a way, the legalese details are less relevant to the industry perception of ARM as the move is seen as lighting the house on fire in attempt to win an argument.
Qualcomm is the supplier of high-performance ARM-based SoCs in consumer segment, with the best performing core design. ARM is not doing damage to Qualcomm here but to ARMv8/9s long-term survival.
I, for one, am greately unhappy about it because RISC-V is a disgusting design that happened to be in the right place at the right time, becoming yet another example of the industry pushing for abysmally inferior choice due to circumstantial factors. I sincerely hope it fails in all possible ways (especially the RVV extension) and a completely redone, better design that is very close to ARMv8-A takes its place.
But, in the meantime, we have ARMv8/9-A, which is the best general-purpose ISA, even with the shortcomings of SVE/2, where AVX family, with especially AVX512VL extension, is just so much better.
The alternative for ARM is to accept that the supplier of high-performance ARM-based SoCs is taking control over the future roadmap of ARM architecture not just for the consumer segment, but a wide range of industries, just like they announced publicly. And they do that by taking the cash-cow of consumer products to push their custom architecture into those other industries.
As of now, ARM is largely in control of the evolution of ARM architecture, because even by those with an ALA (like Qualcomm), ARM's CPU-designs are the reference for the evolution in the respective industries. Straying too much from those designs turned out to not be economically feasible for most players since the move to 64bit, which is a beneficial development for ARM as they can drive a harmonized ecosystem in different industries.
Now ARM gave Nuvia a very permissive license to cooperate on the creation of ARM-based architecture, for a segment where ARM was very weak: server-architecture. With the licensing contract explicitly limiting Nuvia to use the resulting IP only for servers and only to Nuvia.
Now regardless of the legal dispute, Qualcomm now plans to use this IP to create a design roadmap parallel to that of ARM, with a market-position in consumer smartphone SoC's funding a potential hostile takeover of several other industries where ARM carefully works to establish and maintain a competitive landscape.
Qualcomm's plan is to achieve something similar to Apple, but with the plan to sell the resulting chipset.
So while ARM is building and maintaining an ecosystem of ARM as a vendor-agnostic architecture-option in several industries, Qualcomm is on a trajectory to build up a consolidated dominant position in all those industries (which may end up forcing ARM to actually follow Qualcomm in order to preserve the ecosystem, with Qualcomm having little vested interest to support an ecosystem outside of Qualcomm).
The IP of Nuvia was not supposed to be used in all the use-cases that Qualcomm intends to deploy it in (and moreover there is still the ongoing legal dispute that Qualcomm is actually not allowed to use it)