Intel Suffers Worst Year Since 1971 IPO (cnbc.com) 34
Intel's market value plunged 61% in 2024, marking its worst performance since going public in 1971, while rival chipmaker Broadcom saw shares surge 111% on AI advances. Broadcom, now valued at $1.1 trillion, leverages its custom XPU chips and networking gear for major cloud providers including Google, helping companies build AI infrastructure at lower costs than Nvidia's GPUs.
Further reading: Intel Weighed $20 Billion Nvidia Takeover in 2005.
Further reading: Intel Weighed $20 Billion Nvidia Takeover in 2005.
Intel's STOCK (Score:2, Offtopic)
Intel's STOCK Suffers Worst Year, for those of us who think a company is more than just money
Re: (Score:2)
Re:Intel's STOCK (Score:5, Interesting)
think about the issues they had with the Gen 13 and 14 CPUs
And being stuck in 14nm for about 6 years while 10nm failed endlessly.
Intel has something notable now. Their B580 GPU is actually good. They're still making up ground in software/firmware refinements, but testing has this GPU running current games well. There is a legitimate alternative to AMD+NVidia today in discrete GPUs. Too bad they were not prepared for the demand: you can't find one in stock at MSRP.
Re: (Score:3)
Re: Intel's STOCK (Score:2)
Re:Intel's STOCK (Score:5, Interesting)
Intel laid a bunch of employees off, fired their CEO, continued missing goals and deadlines, exited markets and product lines. Their recent GPU release was well-received as a solid budget offering by the PC gaming press, but otherwise I can't think of many positive headlines from Intel this year. While we can definitely objectively say that this was the worst year for the stock, we can probably also subjectively say that this was one of the worst years (if not the worst) for the company.
Re: (Score:3)
The one good thing we can say is their budget GPU is the first step towards building the kinds of parallel systems needed for matrix math like AI systems.
But their management culture has been rotten since the 90's, letting sales and marketing "business" people torpedo their engineering staff. Pentium 4, anyone?
Shareholder alarm bells should have gone off when Apple found it to be better to build their own chip than rely on Intel. And then they caught up and perhaps exceeded Intel on some metrics.
Then the
Re: Intel's STOCK (Score:1)
Re: Intel's STOCK (Score:4, Informative)
Dec Alpha was a Reduced Instruction Set Computer (RISC) that was "really" reduced and 64-bit. Alpha could not even load a single byte. String operations needed to load 8 bytes and the shift bits to extract individual bytes. The transistors Alpha saved by not decoding complex instructions were used to supply lots of registers. If I recall correctly, Alpha processors were not pipelined.
Pentium 4 was 32-bit. Pentium 4 and its successors have the most complex variable length instructions ever attempted in a CPU. Pentium 4 and is successors translate complex instructions into multiple simpler instructions and then execute the simpler instructions. Pentium used a seventeen stage pipeline! Transistors are more plentiful today but the waste of so many transistors delayed the Intel move to 64-bit by a decade and still hampers the architecture to this day.
It was the many-stage pipeline needed to obtain high clock rates that limited the Intel design and caused so much power consumption.
Re: (Score:2)
https://www.chessprogramming.o... [chessprogramming.org]
If I read that correctly, the first ones had no pipeline for decoding instructions.
But had pipelines for integer executions and floating point executions.
Re: (Score:2)
Dec Alpha was a Reduced Instruction Set Computer (RISC) that was "really" reduced and 64-bit. Alpha could not even load a single byte. String operations needed to load 8 bytes and the shift bits to extract individual bytes. The transistors Alpha saved by not decoding complex instructions were used to supply lots of registers. If I recall correctly, Alpha processors were not pipelined.
The original Alpha was pipelined with in-order execution. Later out-of-order execution was implemented.
Out-of-order execution is more complicated but *allows* higher clock speeds because it increases the allowable load-to-use delay, meaning that the first level cache does not have to be as fast for a given clock speed. Clock speed is limited by latency to the first level cache.
The lack of smaller than 32-bit operands ended up being a mistake, so later models added byte and word operands.
Transistors are more plentiful today but the waste of so many transistors delayed the Intel move to 64-bit by a decade and still hampers the architecture to this day.
Intel chose not to
Re: (Score:2, Informative)
This is a company that makes some $20 billion in annual profit. Something tells me that wasn't happening in 1971.
Re: (Score:2)
Re: (Score:2, Insightful)
Almost like stock prices aren't a measurement of how well a business is doing!
Re: (Score:2)
This is a company that used to makes some $20 billion in annual profit. Something tells me that wasn't happening in 1971.
My edit - Intel has been losing a ton of money in recent quarters - this is no longer the case. Hense the dying stock price.
Re: (Score:2)
Intel and Broadcom (Score:5, Insightful)
Intel is trying very hard to be a long-haul, sustainable, top-tier chipmaker with a full suite of product lines and manufacturing facilities. Yeah, they’re struggling, but mostly because the chipmarket is viciously competitive and this little unknown company in Taiwan has polished it’s game to perfection..
Broadcom doesn’t develop new products. They buy companies, lay off 50% of the workforce, cancel all R&D projects, and increase product prices by 10 times. They lose 80% of their customers, but the 20% that cant easily dump the product pays 10x the cost, and that amounts to a 4x increase in revenue if you do the math. Currently, they’re making money hand over fist and the stock market loves them.
But, with a tiny, pissed-off customer base and zero r&d, each broadcom business is coasting on profit from a dead product line. It’s where companies go to die and be scavenged for resources. This is fine and all - efficient capitalism requires recycling of corporate resources. But companies don’t like 10x price hikes. Pretty soon, companies will flee a product if there’s even a hint that it might be bought by broadcom.
I’ve read articles that compare broadcom to GE under Welch, except on a much larger scale.
Re: (Score:2)
Broadcom doesn’t develop new products. They buy companies, lay off 50% of the workforce, cancel all R&D projects, and increase product prices by 10 times. They lose 80% of their customers, but the 20% that cant easily dump the product pays 10x the cost, and that amounts to a 4x increase in revenue if you do the math. Currently, they’re making money hand over fist and the stock market loves them.
But, with a tiny, pissed-off customer base and zero r&d, each broadcom business is coasting on profit from a dead product line. It’s where companies go to die and be scavenged for resources. This is fine and all - efficient capitalism requires recycling of corporate resources. But companies don’t like 10x price hikes. Pretty soon, companies will flee a product if there’s even a hint that it might be bought by broadcom.
Just to add emphasis, that is exactly what Broadcom does.
In addition, Broadcom, as it is known now, has killed several profitable but small product lines leading to an exodus of people and industries from the US. There are many cutting edge technical products which can no longer be designed and manufactured in the US because of Broadcom's actions, and these represent technologies which cannot be supported in the US any longer.
Are you a betting person? (Score:4, Interesting)
Looking at the lifetime (MAX) stock price graph on the Google shows that, over time, Intel has a period of low cap followed by periods of increased growth. Current price is about $20 a share and with typical periods going up to the high twenties low thirties - seem like every 20 years or so there's a "wave" where the cap goes insane, but the overall value of the company appears to be somewhat stable.
So you gotta ask yourself, are you a betting person? Do you buy now, or wait a little longer in the hopes it goes even lower, then hold on to it long term waiting for AMD (or another competitor) to make a mistake so they can take on more market share before selling your stake?
All the "doom and gloom" is stupid, recent share price trends came off the back of the pandemic where people were quarantining and buying PCs to pass the time - that market has more or less dried up now and Intel is going back to it's more "stable" operating mode. Personally I prefer long term stability, not that I play the stock market directly, as I'd prefer to see returns over time versus taking a huge risk and losing my shirt hoping for that "lottery" win that's more luck than skill.
Not gonna talk about their chip issues this year, still rocking a 2009 iMac with a Windows 10 install as my daily driver and not really looking for a screaming new machine at the moment, but whatever problems they're having are also cyclical - AMD kicks their ass for a while then Intel catches up and kicks their ass, then AMD catches up to kick their ass again.
Re: (Score:2)
History is not indicative of future performance (Score:2)
Re: (Score:2)
Intel hasn't had a particularly good products since... Pentium M? Maybe P3. Around that time.
Their fortunes really depend on how good AMD's products are, and how much Intel can escape its own screw ups. Right now they have massive liabilities because they CPU they made in the last couple of years is going to die a premature death, and it's only really a question of how much of it they can weasel out of.
Re: (Score:2)
The Core 2 was good, although you could argue it was a continuation of the Pentium III/Pentium M family. The *Bridge and *well Xeons were excellent. Skylake was when it all fell apart.
Their latest flagship CPU (Score:2)
is still having the same problem as the previous iteration.
Risc-V (Score:2)
Intel seem to be losing the battle with Arm (and its licensees Apple Silicon, Qualcomm, Nvidia...) using x86-64.
What of their Horse Creek collaboration with SiFive? To me that seems a natural evolution, Intel develop the motherboard with fully open source Linux IP for WiFi, Ethernet, GPU etc and let SiFive design the CPU.
Lol ok (Score:2)
Re: Lol ok (Score:1)