During the height of the covid tech bubble, Nvidia offered to buy ARM for $40b. Right now, Softbank wants a $60b - $70b valuation when many tech stocks have lost 90%+ and some large caps have lost 30% - 50%.
What justifies such crazy increase in value? Also, I believe Nvidia overpaid for the company regardless. Nvidia was willing to pay $40b because they want a world-class CPU design team to integrate their GeForce IP into SoCs and service chips.
But a standalone ARM is not very valuable. The reason is that ARM's business model (licensing core designs and ISA) makes peanuts compared to Qualcomm, Apple, Intel, AMD, etc. In addition, ARM's biggest customers are also their biggest competitors. For example, Apple competes with stock ARM designs with Apple Silicon. Qualcomm will be competing with ARM designs via Nuvia chips. Ampere Computing just designed a custom ARM core of their own.
When ARM only license the ISA (Apple Silicon, Nuvia, and Ampere One), they make peanuts. When they license ARM core designs, they make slightly more than peanuts.
It's generally not a good business to invest in. I find it hard to justify the $60b - $70b valuation. No doubt Softbank will try to sell ARM as an AI company. It's not.
>Arm reported $524 million in net income on $2.68 billion in revenue in its fiscal 2023, which ended in March
>i find the current valuation ludicrous, and it seems like it's pushed more by Softbank's Vision Fund than a firm grasp in reality.
Just for comparison, AMD, which is still quite small, gets about $2b - $4b annual net profit. Intel, before their recent disaster quarters, had as much as $24b in annual net profit.
$524m in net income is peanuts compared to the big boys. This is what I was saying in my original post. ARM is a more valuable company if they were acquired by Nvidia. As a standalone company, it's not that great. Again, a weird quirk of ARM is that their biggest customers are also their biggest competitors. This puts a cap on how much profit they can make. If ARM decides to raise licensing fees exponentially, which is likely not simple due to long-term contracts, then companies will seriously look to RISC-V.
Because ARM is the smallest fish in the pond, it can't pay for the best engineers. The best chip engineers will go to AMD, Intel, Nvidia, Apple, Qualcomm, and startups. ARM is where these companies go to poach.
If it was too successful on its own it wouldn't be in all those devices. Whether or not it has the best chip designers its ARM that has changed the world.
It seems a bit like Linux to me - you could say Linus Torvalds is disappointing because he's not as rich as Elon Musk and yet one might argue both that he has done more and that if he had tried to get super rich out of it he would not have achieved so much because every effort to extract significant money would have lessened the breadth of his impact. People wouldn't have co-operated.
If you think about it, should any of the major players buy ARM and shut it down to the competition, the other players would be in deep trouble (Samsung, Apple, AWS, Google all depend fairly deeply on ARM). That has the potential to drive valuation up a bit, no?
The strength of ARM is that it's universal. If you shut down ARM to the public, then they will switch to RIS-V and suddenly, software will no longer support ARM.
If it's that simple, they would have done it a long time ago. In reality, many companies have very long-term contracts. For example, Apple may have an indefinite architectural license.
I haven't looked into all their changes in detail but they're not just going to suddenly increase profit without their customers fighting back.
What justifies such crazy increase in value? Also, I believe Nvidia overpaid for the company regardless. Nvidia was willing to pay $40b because they want a world-class CPU design team to integrate their GeForce IP into SoCs and service chips.
But a standalone ARM is not very valuable. The reason is that ARM's business model (licensing core designs and ISA) makes peanuts compared to Qualcomm, Apple, Intel, AMD, etc. In addition, ARM's biggest customers are also their biggest competitors. For example, Apple competes with stock ARM designs with Apple Silicon. Qualcomm will be competing with ARM designs via Nuvia chips. Ampere Computing just designed a custom ARM core of their own.
When ARM only license the ISA (Apple Silicon, Nuvia, and Ampere One), they make peanuts. When they license ARM core designs, they make slightly more than peanuts.
It's generally not a good business to invest in. I find it hard to justify the $60b - $70b valuation. No doubt Softbank will try to sell ARM as an AI company. It's not.