I’ve been a fan of ARM Holdings for a very long time…
There are many specialized chip makers and the industry is organized as a value chain of designers and merchant fabs, but two names stand out: Intel and ARM Holdings. They are the Gorillas: Intel of servers and PCs and ARM of portables. I first got interested in ARM Holdings (the company, not the stock) when Apple selected them to power the Newton. Ever since I have heard people say that Intel would eat their lunch. But it’s ARM Holdings that’s nibbling on Intel. The chip business was Intel’s to lose and they are slowly losing it, at least in mobile.
SoftBank reportedly seeking to raise up to $10B via ARM IPO
Thanks for the heads-up! This gives ARM util the end of September to IPO.
The accounts reveal that another subsidiary of Arm parent SoftBank entered into an agreement with financial intuitions in March 2022 to borrow $8 billion, providing 75.01 percent of Arm’s shares as collateral.
“A springing guarantee, naming Arm as the guantor, becomes effective on the triggering event of material breach or misrepresentation, an IPO not occuring within 18 monthgs of 31 March 2022 or such intention being announced. On trigger, future events of default fall to the company.”
BTW, I don’t buy IPOs. It’s better to wait until the stock settles down.
They’ve hit $126 today. I used to work there. Honestly, this surprises the cr@p out of me. I never expected this. Just goes to show you, just because you work at (or used to) a company does not mean you really know how well the company is going to do!
One thing I also think this means is the RISC-V threat to Arm is being over-blown.
I used to be a great fan of ARM Holdings until they sold out to SoftBank. You might be able to clear up a notion I have about ARM that might explain why RISC-V is not a threat. In addition to their design expertise I believe they have some kind of a deal with the fabs to optimize the silicon. Any ideas?
I can read later but will say this now. SiFive was the leading RISC-V start-up the last several years but recently had big layoffs and killed future product development. That there is the biggest reason to think the threat is not as big as once thought. The other is going to be the software ecosystem that exists for ARM. That would all have to be ported. I don’t know how easy that would be however. But it would not be trivial. And that is a real moat. Just look at NVDA compared to AMD in that regard. (Can’t say more, I work at NVDA and used to work at AMD years back).
Now, regarding software. Back when I worked at Oracle on SPARC there were rumors we might get cancelled. Back then SPARC only ran on Solaris, but there was a push to get SPARC on Linux. Rodrigo Lang was trying to ease our fears but he said something that concerned me. He mentioned how it is relatively simple to port software to a new ISA, but much harder to port it to a new OS. Want to change processor architecture? Then you use a different compiler. Most of the problem is solved. But to move to a new operating system, that really does require code re-writes, sometimes in big ways. So what he was telling us was he was actively getting the company and our customers through “the hard part” of porting all the SPARC applications to Linux. After that, moving everyone to x86 would be much simpler. Turns out I was right, in roughly a year they cancelled SPARC.
I don’t know anything specifically about any deals Arm has with fabs. If they did I would not have known. Agreed about the Soft Bank vibes. I joined shortly after they went private. It was a good time as an employee. I went to England twice a year, as did most people. I could have probably gone more often and got approved. Just how extreme was our travel? I’ll give you an idea. British Airways had to add a daily non-stop from Austin to Heathrow because of our demand for travel. (note: that demand no longer exists by the way). In other words, Massa showered Arm with money. And at first he wanted us to rule the Internet of Things. Turns out that was a misguided direction to go. The SoftBank years for Arm were probably a bad diversion, unfortunately. But I love the stamps in my Passport from those years.
You talked about server chips, and Arm has done very well there, just look at Amazon’s Graviton, it’s a fantastic server chip. Fast, cheap, power efficient. Most people will get better value from AWS using Graviton than Intel, by a large margin. And I’m tying this currently on a M1 Pro computer, and the battery life of this thing is unreal. And Nvidia’s super computer ties their (our, I work there) GPUs with an Arm processor, not an x86 one. 'nough said. :).
There are three main processor companies now: INTC, AMD, ARM. Of those three I find only ARM a compelling buy. (I own none of them however).
Build AI inference chips to run our Full Self-Driving software, considering every small architectural and micro-architectural improvement while squeezing maximum silicon performance-per-watt. Perform floor-planning, timing and power analyses on the design. Write robust tests and scoreboards to verify functionality and performance. Implement drivers to program and communicate with the chip, focusing on performance optimization and redundancy. Finally, validate the silicon chip and bring it to mass production in our vehicles.
At a high level, the chip is a full system-on-a-chip capable of booting a standard operating system. It is manufactured on Samsung’s 14-nanometer process at their Austin, Texas fab, packing roughly six billion transistors on a 260 millimeter squared silicon die. The FSD chip meets AEC-Q100 Grade-2 automotive quality standards. The choice to go with a mature 14 nm node instead of a more leading-edge node boiled down to cost and IP readiness. There are twelve 64-bit ARM cores organized as three clusters of quad-core Cortex-A72 cores operating at 2.2 GHz which are used for general purpose processing. There is also relatively light GPU primarily designed for light-weight post-processing. It operates at 1 GHz, capable of up to 600 GFLOPS, supporting both single-precision and double-precision floating point operations.
From early on ARM’s list of customers was a Who’s Who in technology.
Yet Musk does NOT trust his own vehicle to use the “FSD” all the time–or even much of the time. Kind of like “Long Island Iced Tea”–not from Long Island, not made there, no relation to NY in any way. It is “just a name”.
Does that mean Musk has enough common sense to only use the tool when it is the “right tool”?
Would you, dear reader, just “engage” the app and go to the back of the car to make a sandwich, as the old joke about “cruise control” goes?
Who here thinks that Musk actually spends much time “driving himself around”?
I suspect when he drives it is an intentional effort, so that he can advertise and actually be cognizant of the state of the development.
The majority of the time, he’s being driven, and escorted by his security team.
At least I hope so. I got $ riding on his health.
ralph
Included only for clarity.
Old joke from 1980’s, 90’s…
<Ethnicity/gender/age of choice> driving on highway in New RV has wreck.
Cop asks “why?”.
Ethnicity answers s/he put it on “cruise control” and stepped into the back of the RV to make a sandwich.
It’s just an adhominem attack and not actually funny.
It morphed in the 2000’s n 10’s into “target-for-ridicule put Maps/GPS on and just blindly followed instructions” right into a wall, a lake or other obstacle.
It’s still just an adhominem attack.
From what I hear, Tesla n Musk strongly recommend staying in the driver’s seat, and alert, keep hands near to the steering wheel, and feet near the brake n accelerator.
That’s what Level 2 means, BTW.
And so far, Tesla FSD is still Level 2?