Intel foundry business to make custom chip for Amazon, chipmaker's shares jump
reuters.com
reuters.com
Another point I think Amazon is willing to partly / dual source ( I dont think Amazon is making a one way jump ) to Intel is TSMC is now constrained on leading node. It is always Apple or Nvidia. And the lead time for them to get any capacity spared is getting longer now Nvidia is also in the game.
Intel can sell design and manufacturing.
"U.S. Govt pushes Nvidia and Apple to use Intel's foundries — Department of Commerce Secretary Raimondo makes appeal for US-based chip production"
https://www.tomshardware.com/tech-industry/us-govt-pushes-nv...
It'll be curious how the financials on this look. I assume with Intel breaking foundry out into a subsidiary, they'll also break out revenue, and if this is their marque first big customers, it should show up in the numbers.
On Server ( for now ). Cutting edge laptop is no longer on Intel but TSMC.
There's no reason for Intel to target that money pit when there are higher value options like Server processors.
The bleeding edge nodes built in Chandler have much better uses.
This is part of the longer term IDM 2.0 strategy who's endgame was always going to be a split or potentially even spinoff of Intel Foundary services from Intel Chip Design services - their goals are just too diametrically opposed and undermining each other.
knowing Amazon its not US goverment to force them but Amazon can make the 2 of them outbidding each other in more desperate times
For open source software loads. If you're running licensed software in the cloud you don't often get to make this choice.
Even if Microsoft does release such as thing, enterprise software lives in an ecosystem of thousands of third-party vendors that provide components that are often co-installed on the same box.
Also, products like MSSQL are licensed per CPU core, so you want the fastest possible cores, not the cheapest. Currently that's the AMD EPYC X86-compatible CPUs, which run rings around everything else.
The only problem is efficiency vs redundancy tradeoff. The organization needs to be antifragile to cope with that.
Yeah, it sucks working for one of their suppliers but every single major company does this: auto companies, console manufacturers, smartphone makers, white goods/appliances, etc.
If you were running such a business, you'd do it to. It's how capitalism works.
Just three days ago: Poland's Intel plant gets EU green light for $1.9B in state support [1].
Germany is subsidizing Intel with $11bn.
I'd be fuming.
This stuff rarely takes the form of the government writing a cheque, especially not upfront.
I wonder whether there are any "outs". DE and PL subsidized because they want fabs ASAP; if Intel can't hold up their end, why not cancel the deal and go with TSMC or Samsung instead.
There are a few threads on this now, so maybe they can be merged and linked to this primary source instead of news articles summarizing it.
https://www.reddit.com/r/wallstreetbets/comments/1ehjuzj/i_b...
I have been hearing in the last week a lot of "Intel is doomed" and I wondered why. Now I Know.
1. if you buy it, and it goes down, dammit I shouldn't have bought it. I'm so stupid.
2. if you buy it, and it goes up, dammit I should have bought more. I'm so stupid.
It's nothing but regrets!
Buying individual stocks is certainly much riskier, but index investing is still stock investing.
Trading in individual stocks is how you slay countless grandmas, please don't do it for your grandma's sake: https://www.reddit.com/r/wallstreetbets/comments/1ehjuzj/i_b...
7x, not 10x
> Even adjusting for inflation, its still a massive return.
Adjusted for inflation it's 4.7x
> If thats not a win, I dont know what is.
You compared the low with the current price. Pick a random stock and pick its lowest point and compare with now - many/most handily beat the S&P 500. Think AMD at $4.
The proper comparison is to pick two random points 10-15 years apart. Don't cherry pick.
5624/677 = 8.3x
Yes, I rounded up, and yes I picked a low to make a point. But go ahead and pick a random point 10-15 years ago if you'd like and compare it to today and the returns are still very large.
That said I don't advise to play the trading game.. it's kinda soulless and empty.
1. I buy, it goes down. “Meh. Can’t win ‘em all. Might as well hold see if it comes back.”
2. I buy, it goes up. “Nice.”
At some point all of your stocks are going to be in recovery mode if you never sell at a loss.
But in all my years investing, I never get used to the downturns and have a hard time being sanguine about it.
your investment going down means the whole economy went down: you are no worse off than the aggregate, and you still own your piece of the economy.
if the whole economy goes down a lot, the widespread largescale unemployment means you are a lot better off than a lot of people, and you still own your segment of the economy.
it's more of an opportunity cost type thing: you can't afford not to be fully invested. In general, the economy grows, you want to grow with it.
nobody can predict downturns; anybody who can is busily correcting the market to make money from it; in any case, that's not us, so we do what we can do.
More discussion: https://news.ycombinator.com/item?id=41560359
Are we to believe that Amazon's workload is so unique that an existing design can't fit the bill?
This is just a vanity exercise.
POWER from this perspective, does way too much. It's too much chip for AWS's needs, and a big waste of money.
That's what optimization means! You get the smallest/ simplest/cheapest thing that does exactly what you want, and ignore the big bad "do everything" chips that go for megabucks.
Then you can devote chip transistor count toward the instructions you want, and ditch all the dead weight that comes with a traditional chip like power, that tries to make a balanced general purpose cpu.
This results in a chip that isn't good for anything but the specific need, but it's super cost effective and higher performing.
You can easily benchmark this by comparing the Graviton variants with the Intel specific instances.
Graviton 4 has also not been made available for EC2 and RDS instance types in most regions two years after release. My guess is that they are unable to scale production.
You need to run actual workloads that fully utilize all cores and not benchmarks to understand the value of Graviton. People don’t run benchmarks in production; they run real business applications. Graviton is optimized for horizontal scale, not straight line single core performance.
Lol graviton has been in production with fantastic margins for like 7 years. Me thinks maybe you don't know as much as you think you do.
Could have been Jassys equivalent of Beso’s AWS, but unfortunately he’s a lousy leader with no imagination beyond the quarterly results.
And Intel is building an "AI fabric" chip for AWS, though its not clear who designed it.
Apple designs their own cores and Qualcomm is starting to. Most people just grab ARM's reference designs. They're sorta in-house in that they're putting it together, but they aren't doing the design work of actually making their own cores.
Qualcomm already did for many years! The Snapdragon line from 2008-2016 used custom Scorpion/Krait/Kryo cores before they went to Cortex A73 cores in the Snapdragon 835.
Only now, after buying the Nuvia CPU design team, they have begun to use again their own CPU core designs, first in the laptops introduced a few months ago and then presumably also in the smartphones of next year.
While many companies have designed Arm cores, only Apple has consistently outperformed the Arm company and now Qualcomm might have a chance to do that, with a team that includes many former Apple designers.
The server Arm cores that have been demonstrated until now by the Ampere company are not impressive, being weaker than the Arm-designed cores used by Amazon or Microsoft.