Arm and Qualcomm are playing a dangerous game of chicken:
In its lawsuit filed against Qualcomm last August, Arm sought to block custom processor core technology acquired by the Snapdragon giant. The reason Arm believes this should happen? Qualcomm allegedly breached Arm's licensing agreements by continuing development of custom cores after the two parties failed to negotiate a new deal for the tech's use.
Qualcomm, by contrast, believes its existing agreements with Arm cover the custom cores it gained from the 2021 acquisition of chip startup Nuvia.
If Arm's lawsuit is successful, it would significantly stifle Qualcomm since the company plans to use the Nuvia technology for future chips in several major market segments. But to some analysts, Arm could suffer more than Qualcomm in the long term if it wins, making the lawsuit a risky move, even if it's based on sound legal reasoning.
[...] "The problem for Arm is that even if it wins the case, it loses," wrote Linley Gwenapp of TechInsights in an editorial published last month. "The company is seeking new revenue streams, and Windows PCs are a big opportunity. This market is served almost entirely by x86 processors from Intel and AMD, so any incremental share benefits Arm through royalty payments."
In the event that Arm decided to terminate all of Qualcomm's license agreements, that would do much greater damage to Arm's royalty and license revenues since the Snapdragon giant is such a major customer of the British chip designer.
[...] The greater risk posed by Arm's lawsuit, Gold and other analysts believe, is that it will push Qualcomm and other chip designers to consider the open-source RISC-V instruction set architecture (ISA) as an alternative to Arm, creating a greater existential threat to the British chip designer in the long run.
[...] The two companies may even try to settle earlier if new evidence that emerges during the lawsuit's discovery process, which begins January 13, convinces one side that its case is weak, Sangam said.
But when it comes down to who may benefit more from a settlement, Sangam thinks it will be Qualcomm and not Arm because the latter [former?)] can afford to let the legal process play out, which could include potential appeals beyond 2024. He also thinks it's unlikely for Arm to win an injunction that would stop Qualcomm from developing the custom Nuvia cores as the case plays out in court.
"There's nothing stopping Qualcomm from doing what they are planning to do while the case drags on, Whereas for Arm, if they are to do an IPO, and if you're an investor in the IPO, you will be worried about how the case will turn out," he said.
(Score: -1, Offtopic) by Runaway1956 on Tuesday January 10, @04:42PM (2 children)
123 testing I really have nothing to say. Tor browser on androi seems to work
Abortion is the number one killed of children in the United States.
(Score: 0, Offtopic) by Runaway1956 on Tuesday January 10, @04:43PM
but spell checker doesn't work!
Abortion is the number one killed of children in the United States.
(Score: 2, Informative) by janrinok on Tuesday January 10, @05:37PM
Test it on your journal please - not on the front page.
(Score: 4, Insightful) by RamiK on Tuesday January 10, @08:29PM (2 children)
ARM can't compete in a RISC-V market so their only long term option is to vendor lock where possible and patent troll everywhere else.
That's entirely in ARM's hands so where's the risk exactly?
ARM is very familiar with Nuvia's patents so I strongly doubt that's happening.
The difference between ARM winning or losing is simply more royalties since Qualcomm can't afford not to sign a new license yet. Maybe if they put their full workforce around migrating to RISC-V, they'll be ready to market in 2-3 years at a bare minimum... But right now, Qualcomm is stuck with ARM and ARM and their prospective investors know this.
compiling...
(Score: 3, Interesting) by Snotnose on Tuesday January 10, @10:29PM (1 child)
Actually they can. The Qualcomm license says QC can change the core ARM as long as the changes were sent back to ARM (disclaimer: I retired from QC 15 years ago so my info is 15 years old).
From the (I admit, since I retired I've only sorta followed things) things I've read RISC-V has issues with less dense op-codes. Which means that for a given clock rate and caching, RISC-V takes longer to "do something" than ARM.
Then again, those folks at Qualcomm are pretty smart, as are folks at other companies looking at this lawsuit and paid royalties and thinking "hmm, we also have a lot of pretty smart folks".
I just passed a drug test. My dealer has some explaining to do.
(Score: 3, Interesting) by RamiK on Wednesday January 11, @05:15AM
It's a (very old) myth. RISC-V's density matches Thumb-2 through compressed instructions and macro-op fusion: https://medium.com/codex/addressing-criticism-of-risc-v-microprocessors-803239b53284 [medium.com] https://riscv.org/announcements/2016/04/risc-v-offers-simple-modular-isa/ [riscv.org] (p.4)
If that wasn't the case, Espressif wouldn't have switched over after only introducing a single gen 1 product: https://www.eenewseurope.com/en/espressif-moves-exclusively-to-risc-v/ [eenewseurope.com]
Basically, their customers told them they throw everything from bare metal and FreeRTOS to MicroPython at the things and it was all good. They even stuck an H.264 decoder into their new ESP32-P4 so those things are being used in some really weird places due to how good they are: https://www.espressif.com/en/news/ESP32-P4 [espressif.com]
compiling...