News Intel might be too big to fail — Washington policymakers are already discussing potential solutions if the chipmaker cannot recover

[UniFan]

Hopefully, someone can provide you with a better answer than this, which actually ranks the various factors in terms of impact, but I can list some of the bigger problems they're facing.

Their cash cow has long been the datacenter market. Most of their biggest customers have turned to using custom ARM-based CPUs, and AMD has now taken a huge chunk of the x86 server market that remains. This not only reduces their volume, but also their negotiating power on server CPU prices. Plus, China is trying to make its own server CPUs, though I don't know how much real decline they've seen in that market.

Meanwhile, newer manufacturing processes are more expensive and they're having to include more CPU die area to compete, which squeezes them on the cost side.

As if that weren't bad enough, they've been investing in fabs for future manufacturing nodes, which is getting ever more expensive. I think a new fab takes something like 5 years from start to finish, before it can start churning out product and recouping the investment. If you stop building fabs, you risk falling behind in the future, so options are limited on that end.

On the desktop products front, they had to cancel Meteor Lake, which left them selling defective Raptor Lake CPUs for two years. They claim now to have fixed the issue via firmware/microcode, but are still on the hook for an unspecified number of warranty claims.

It gets worse: their latest desktop and laptop CPUs are now made almost entirely by TSMC, which has made them less profitable at a time when Intel could really use the money.

One of their biggest customers, Apple, switched to making its own CPUs starting about 4 years ago. The ARM crowd is also gunning for the lucrative Windows laptop market, with Qualcomm leading the charge.

Oh, and their GPU and AI businesses have yet to turn a profit. Their GPUs missed the crypto boom, by a couple years, and had middling performance (at best) and could only be sold at bargain prices. Their AI solution, Gaudi, has yet to find a good market niche.

Here's a historical plot I found of their revenue, though it only includes data through 2023:

​

Intel revenue 1999-2023 | Statista

In 2023, Intel's revenue amounted to approximately 54.2 billion U.S.

www.statista.com

​

Here's the article discussing their latest quarterly report, which goes into some of these areas in a little more detail:

​

https://www.tomshardware.com/pc-com...lion-in-q3-reports-usd13-3-billion-in-revenue​

​

thanks for the reply!! I only knew about their GRU failing and Apple making their own CPUs. Hoping they come back swinging.

 

[bit_user]

thanks for the reply!! I only knew about their GRU failing and Apple making their own CPUs. Hoping they come back swinging.

Intel can still design good CPUs, but the biggest problem they face is their fabs. Those are like a big boat anchor around their neck, weighing them down. I believe they will be spun off, sooner or later, but the foundry business would not be able to keep innovating, if it were split off right now - it hasn't yet reached critical mass.

Crazy idea: maybe OpenAI (or some other group headed by Sam Altman) will buy the fabs? At the beginning of the year, I recall him trying to raise $Trillions in funding to build a global network of leading edge fabs!

 

[bit_user]

@UniFan , there are some nice graphs showing AMD vs. Intel marketshare trends, here:

https://www.tomshardware.com/pc-com...ndal-amd-makes-biggest-leap-in-recent-history
​

I'm sure that last one is hurting Intel the most, since it's concerning high-value server CPU parts. What's probably worse is the decline in pricing power that tends to accompany the nearly complete market dominance they had before. In fact, if you look at total revenue, AMD actually just passed Intel in the Datacenter market!

https://www.tomshardware.com/pc-com...er-amd-outsells-intel-in-the-datacenter-space
​