Right now, I'm imagining AMD as Danny DeVito's character, Frank Reynolds, in that episode of Always Sunny where the gang wrestles for the troops. He's the Trashman, which is an awful wrestling persona, but at the end of the match, everyone else is seriously injured on the floor of the ring. With a bewildered look on his face, he slowly raised his arms, and the crowd went wild.
Sometimes you just gotta stick to the core competencies and stick to them well.
> Sometimes you just gotta stick to the core competencies and stick to them well.
Part of me thinks this is what Radio Shack could've made work. They almost made it to the point in time where hobbyist electronics / maker culture was popular again, but by that time they were just the place people used to buy cell phones from.
Radio Shack got bad. Even when they still had components, it's hard to justify driving to the store to spend $30 for an aux cord when Amazon will have it here tomorrow for $10.
Notably, the whole sector seems to have collapsed in favor of online shops like Sparkfun and Adafruit. Most small cities used to have a shop where you could go pick through trays of random electronic components. Now, I don't even know where you'd go to do that in major cosmopolitan centers that certainly have the populations to support it. It's just too easy to make a shopping list of components at your desk and hit buy.
Unfortunately, I suspect there are also lots of projects that are on hold while that one component you just realized you needed makes its way through the mail.
If I had millions to buy the Radio Shack IP, my hail mary play would have been to convert Radio Shacks to mini maker spaces. Stock "jellybean" components and entry level tools, and then put in work benches with various tools available for learning and "one off" needs. Take a page out of Walmart's playbook and hire a bunch of retired EE's and developers that just want something to do, and take a page out of Apple's playbook and sell a "once a week come chat with an expert one on one" membership. Pair one-on-one members with your retired EEs and market the hell out of it to highschool kids and hobbyists.
The goal is to get people in the store and building together. Of course no one is going to drive to the store and spend $30 on an aux cable when Amazon will ship it to you tomorrow. But they might just spend $20 on a pack of resistors and a USB breakout board while they're already at the shop using the oscilloscope to probe out a signal on their in flight project and ordering from amazon means stopping progress for the day.
I'd probably also aim to have on-prem competitions, think ant-weight battle bots stuff, micro-mouse or even pinewood derby stuff.
Heck you could probably even lure in some additional markup sales on components by having a "grab-bag" bench, in which you can't bring your own stuff, but you only pay for what you take out. Want to build a battle bot but don't know in advance what pieces you will or won't need? Build it on the grab-bag bench and we'll total up your costs after the fact. Sure you probably could get everything cheaper if you ordered it all online, but how convenient to just have a massive wall of parts already and only get the things you need and not need to store the rest of the 500 resistor pack you had to order to get the 20 you needed.
It might not have worked, the whole thing still might have ended in liquidation, but it feels like it would have had a better chance than batteries and cell phones.
Broadcom getting involved with VMware did wonders for all VMware customers - I'm sure all current users of Intel products and services are also delighted in a similar manner! ;-)
Unfortunately, it could only be an improvement after the 13th generation problems. I strongly believe that Intel owes me $150 after releasing a firmware update that turned my processor into one that cost $150 less.
You lack imagination. Broadcom brings a whole slew of problems you may not have thought possible. We couldn't bet them to actually give us a quote to purchase VMware licenses for a new cluster (to replace the current one, as we're a current customer) for the first five months of 2024. Multiple month long freezes in sales while they restructured pricing, multiple times.
Their goal isn't necessarily to run a business that's profitable long term, it may be to segment it out and squeeze various segments to wring the last bits of juice out of them until leaving a dried husk while providing middling quality to the most lucrative segments to string them along.
You don't want to be in the segment destined to be juiced.
Intel has "software defined, license activated" features baked in the silicon now. It's a wet dream for Broadcom.
> You need AVX only? It's only bundled with all AVX flavors and vector operations add-on pack, just for $500/socket/year* (*:for processors up to 10 cores. For higher core counts and SMT enabled options please reach out to your nearest Broadcom sales representative).
Speculation: Broadcom is now really two companies in one, the original Broadcom and CA Technologies. The CA Technologies side of Broadcom bought VMWare and are now applying all the classic CA Technologies moves to it. The original side of Broadcom is the one considering buying Intel
holy shit. it really is the end, isn't it? cut the company into pieces and sell to the highest bidder. three letter agencies would be interested... a couple months ago.
To be fair, they are elected by shareholders, so it does make sense to prioritize shareholders.
But with a company like Intel, one does really need long term vision to maximize value in the long run. We don't know if he's looking for quick fixes (short term) or long term fixes.
He's been on the board that hired the last 3 CEOs, so I have little faith in their getting things right.
Intel started having issues before pat. Now, I sold my amd shares a few years ago and stopped following the semi sector, but as long as I have memory Intel was starting to show cracks after a few ryzen iteration(s). I mean, they were still faring well, but it was clear - at least to me - they were stuck with less competitive products
They missed two epochal changes in the semiconductor industry - mobile and AI/GPU. They don't have a contract foundry business that could compete with TSMC. They appear to be failing with the thing the originally should have been capable of, their next gen chips.
When they missed out on mobile it appeared to be a big blunder, especially when one considers the long term impact of electricity usage in datacenters. When Apple released the M1 chips that could run x86 faster than native, it was clear what was going to happen to Intel. Intel is now done -- dead, finished. This is a legacy business where the owners can try to get more money than they paid for it, but if they have to put up new capital (that isn't free government money), they will not be getting their money back.
AMD was at a very similar spot a decade ago. It took time, pain and faith, but they turned it around. Intel could very well do the same, but its Lisa Su has just got fired a few months before the plan started sprouting.
What a nightmare. Imagine Broadcom merging with IBM. They'd have the legacy IT market cornered whether virtualized x86 or mainframe, customers be damned. Perhaps they already do. Mainframe users often rely on CA Tech's software, which Broadcom bought years ago.
Since reading a Gibson novel that mentioned the dystopian mega tech conglomerate (DatAmerica maybe?), I've always enjoyed the mental exercise of M&A to create a monster corporation.
Intel + Broadcom + IBM + HP + Oracle + Apple + Google
Fab-wise, presumably they'll just gobble up the sites and equipment. Maybe some of the next-gen tech is interesting. But then we'll be left with less competition in the fab industry. This is very unlikely to be good news for anyone other than TSMC shareholders.
That might be true, no idea, but I was actually making a point about the rapidly diminishingly fab competion and what that's going to mean for customers (and consumers).
Yes, there won't be an Intel when they're done. Previous customers will be using AMD x86 or ARM after suffering being screwed by Broadcom for a few years.
Oregon exports more than $30 billion worth of electronics annually, most of which are semiconductors manufactured in Intel’s Hillsboro factories.
----------
The company employs more people than any other business in the state.
----------
The state’s chip industry pays an average yearly wage of more than $150,000, according to government data, well over double the state average. Hefty pay from Intel is an especially big deal in Oregon, which relies on personal income taxes for the bulk of state revenue.
----------
...Intel might consider selling its business off in pieces.
That could be calamitous in Oregon, turning Intel’s huge local campuses into satellite operations of other companies that don’t need the full range of corporate and technological functions that have been the foundation of Intel’s Hillsboro sites for decades.
“The people who work at Intel have to be really panicked,” Hutcheson said.
Intel’s research factories in Hillsboro support high-volume production sites in Arizona, Ireland, Israel and -- soon -- in Ohio. If that manufacturing network were someday to break up, it’s not clear what role would be left for the company’s top engineers and scientists in Oregon.
I wonder if this couldn't be the chance for Apple to make the definitive move towards vertical integration. If you think about it, it's the only missing part in their vertical integration story.
It would really be a sad day and the begin of truly dark era for the entire computing industry, though.
As an alternative... Amazon would probably be another interesting buyer. Amazon is another company dead serious about having their own silicon. And owning the main supplier of x86 chips as well in-house capabilities to manufacture their own Graviton chips would give them an incredible edge over other cloud computing providers.
Broadcom is a very hostile and abusive company. That would be the end of Intel chips.
Things Broadcom would do: sell you a chip, nerf if after a year and charge you a subscription to unnerf.
Look at what they did to VMWare. If you were a VMWare customer you were and are in deep shit right now. The fact this conversation between Intel and the Grim Reaper is taking place means everybody, from consumers to OEMs to clouds need to divest away from Intel today.
Sell what you have while you still can get some cash for it and remove it from consideration for any future plans.
Lots of people still use older processor nodes. For example Groq[1] uses 14nm chips for their inference chips. Older fab tech is cheaper to use (less demand and higher yields). Most chips made are on older processes, it's only the bleeding-edge CPU/GPU chips that use bleeding-edge fab tech.
I'm assuming TSMC has new and old lines. A 555 chip isn't done on the same line as an AMD Zen 5 chip. Permits, buildout, infrastructure, suppliers, you already get a big warehouse that makes chips and you can bring in your own people to optimize it on day one not year 4.
Intel has a couple of euv machines that tsmc could use and they could take Intel's current nodes and sell them as US made chips without threatening their own top nodes.
I wonder how much this is costing Intel in terms of enterprise sales? I'm guessing that headlines like this will accelerate the switch to AMD and ARM. Who wants to design servers around chips from a company that is not likely to exist in its present form? That opens up all kinds of supply chain and support concerns.
I'm not seeing what Broadcom brings to the table here. They don't have chip mfg experience. Sure, they're a customer, but so what?
TSMC is not going to be motivated to build up a competitor who weakens the economic wall around Taiwan.
What I could see happening based on recent events is that Trump tells Taiwan they have a couple years of US security left. TSMC is then forced to consider putting more effort into mfg in the US(and other countries). What that would mean for the rest of the elec mfg supply chain is unknown. Taiwan is quite critical beyond just chip mfg.
Right now, I'm imagining AMD as Danny DeVito's character, Frank Reynolds, in that episode of Always Sunny where the gang wrestles for the troops. He's the Trashman, which is an awful wrestling persona, but at the end of the match, everyone else is seriously injured on the floor of the ring. With a bewildered look on his face, he slowly raised his arms, and the crowd went wild.
Sometimes you just gotta stick to the core competencies and stick to them well.
> Sometimes you just gotta stick to the core competencies and stick to them well.
Part of me thinks this is what Radio Shack could've made work. They almost made it to the point in time where hobbyist electronics / maker culture was popular again, but by that time they were just the place people used to buy cell phones from.
Radio Shack got bad. Even when they still had components, it's hard to justify driving to the store to spend $30 for an aux cord when Amazon will have it here tomorrow for $10.
Notably, the whole sector seems to have collapsed in favor of online shops like Sparkfun and Adafruit. Most small cities used to have a shop where you could go pick through trays of random electronic components. Now, I don't even know where you'd go to do that in major cosmopolitan centers that certainly have the populations to support it. It's just too easy to make a shopping list of components at your desk and hit buy.
Unfortunately, I suspect there are also lots of projects that are on hold while that one component you just realized you needed makes its way through the mail.
> Now, I don't even know where you'd go to do that in major cosmopolitan centers
Micro Center. They are what Radio Shack might have become if they'd been well-funded and stayed in business.
Unfortunately there are not very many of them.
https://www.microcenter.com/site/stores/default.aspx
If I had millions to buy the Radio Shack IP, my hail mary play would have been to convert Radio Shacks to mini maker spaces. Stock "jellybean" components and entry level tools, and then put in work benches with various tools available for learning and "one off" needs. Take a page out of Walmart's playbook and hire a bunch of retired EE's and developers that just want something to do, and take a page out of Apple's playbook and sell a "once a week come chat with an expert one on one" membership. Pair one-on-one members with your retired EEs and market the hell out of it to highschool kids and hobbyists.
The goal is to get people in the store and building together. Of course no one is going to drive to the store and spend $30 on an aux cable when Amazon will ship it to you tomorrow. But they might just spend $20 on a pack of resistors and a USB breakout board while they're already at the shop using the oscilloscope to probe out a signal on their in flight project and ordering from amazon means stopping progress for the day.
I'd probably also aim to have on-prem competitions, think ant-weight battle bots stuff, micro-mouse or even pinewood derby stuff.
Heck you could probably even lure in some additional markup sales on components by having a "grab-bag" bench, in which you can't bring your own stuff, but you only pay for what you take out. Want to build a battle bot but don't know in advance what pieces you will or won't need? Build it on the grab-bag bench and we'll total up your costs after the fact. Sure you probably could get everything cheaper if you ordered it all online, but how convenient to just have a massive wall of parts already and only get the things you need and not need to store the rest of the 500 resistor pack you had to order to get the 20 you needed.
It might not have worked, the whole thing still might have ended in liquidation, but it feels like it would have had a better chance than batteries and cell phones.
In my fan fiction universe Radio Shake buys Amiga and with their sales channel it establishes itself as a superior competitor to PC and Mac.
https://youtu.be/L1EFKZs3Bvk?si=dundU4QVL_U0F0z7 (NSFW probably)
Broadcom getting involved with VMware did wonders for all VMware customers - I'm sure all current users of Intel products and services are also delighted in a similar manner! ;-)
Unfortunately, it could only be an improvement after the 13th generation problems. I strongly believe that Intel owes me $150 after releasing a firmware update that turned my processor into one that cost $150 less.
You lack imagination. Broadcom brings a whole slew of problems you may not have thought possible. We couldn't bet them to actually give us a quote to purchase VMware licenses for a new cluster (to replace the current one, as we're a current customer) for the first five months of 2024. Multiple month long freezes in sales while they restructured pricing, multiple times.
Their goal isn't necessarily to run a business that's profitable long term, it may be to segment it out and squeeze various segments to wring the last bits of juice out of them until leaving a dried husk while providing middling quality to the most lucrative segments to string them along.
You don't want to be in the segment destined to be juiced.
Intel has "software defined, license activated" features baked in the silicon now. It's a wet dream for Broadcom.
> You need AVX only? It's only bundled with all AVX flavors and vector operations add-on pack, just for $500/socket/year* (*:for processors up to 10 cores. For higher core counts and SMT enabled options please reach out to your nearest Broadcom sales representative).
ref: Intel on Demand (https://www.intel.com/content/www/us/en/products/docs/ondema...)
I'm sure Intel's open source contributions will quadruple, seeing how BCM is such an avant-garde.
Curious. What are some examples of BCM being avant-garde?
He’s being sarcastic
I think Intel is a different kind of acquisition for Broadcom than VMWare.
VMWare had/has a strong moat which can be exploited by jacking up prices. Intel doesn't have that.
Speculation: Broadcom is now really two companies in one, the original Broadcom and CA Technologies. The CA Technologies side of Broadcom bought VMWare and are now applying all the classic CA Technologies moves to it. The original side of Broadcom is the one considering buying Intel
God please no, we need competition. Whoever fired Pat Gelsinger should be fried.
This is the guy who fired Pat Gelsinger:
> Yeary has been telling individuals close to him that he is most focused on maximizing value for Intel shareholders, the report added.
RIP Intel.
holy shit. it really is the end, isn't it? cut the company into pieces and sell to the highest bidder. three letter agencies would be interested... a couple months ago.
To be fair, they are elected by shareholders, so it does make sense to prioritize shareholders.
But with a company like Intel, one does really need long term vision to maximize value in the long run. We don't know if he's looking for quick fixes (short term) or long term fixes.
He's been on the board that hired the last 3 CEOs, so I have little faith in their getting things right.
Whoever hired Pat Gelsinger should be fired … from a cannon.
Intel started having issues before pat. Now, I sold my amd shares a few years ago and stopped following the semi sector, but as long as I have memory Intel was starting to show cracks after a few ryzen iteration(s). I mean, they were still faring well, but it was clear - at least to me - they were stuck with less competitive products
They missed two epochal changes in the semiconductor industry - mobile and AI/GPU. They don't have a contract foundry business that could compete with TSMC. They appear to be failing with the thing the originally should have been capable of, their next gen chips.
When they missed out on mobile it appeared to be a big blunder, especially when one considers the long term impact of electricity usage in datacenters. When Apple released the M1 chips that could run x86 faster than native, it was clear what was going to happen to Intel. Intel is now done -- dead, finished. This is a legacy business where the owners can try to get more money than they paid for it, but if they have to put up new capital (that isn't free government money), they will not be getting their money back.
AMD was at a very similar spot a decade ago. It took time, pain and faith, but they turned it around. Intel could very well do the same, but its Lisa Su has just got fired a few months before the plan started sprouting.
[dead]
What a nightmare. Imagine Broadcom merging with IBM. They'd have the legacy IT market cornered whether virtualized x86 or mainframe, customers be damned. Perhaps they already do. Mainframe users often rely on CA Tech's software, which Broadcom bought years ago.
Sir there are children frequenting this site.
Since reading a Gibson novel that mentioned the dystopian mega tech conglomerate (DatAmerica maybe?), I've always enjoyed the mental exercise of M&A to create a monster corporation.
Intel + Broadcom + IBM + HP + Oracle + Apple + Google
They'd from Chips to Cloud.
What would HP bring to this hypothetical merger?
Talent and IP (they swallowed competition like SGI) but I was thinking about having a company that (still) made business machines and servers.
Shitty printers.
Barbarians at the Logic Gate.
Broadcom getting involved with Intel will be excellent business ... for all of Intel's competitors.
They’d hike prices 10x for the whales: hyperscalers and PC manufacturers and ignore the needs of anyone who does not buy CPUs by the tray.
The problem is... the whales have easily obtainable alternatives, which are better to boot. VMWare game plan won't work.
Tsmc might be good for the fab business. Broadcom very very bad for everything else. Broadcom historically doesn’t acquire, it kills.
Fab-wise, presumably they'll just gobble up the sites and equipment. Maybe some of the next-gen tech is interesting. But then we'll be left with less competition in the fab industry. This is very unlikely to be good news for anyone other than TSMC shareholders.
Historically, Broadcom has done even better than TSMC for its shareholders.
5 year annual return is 56% for Broadcom and 33% for TSMC
10 year annual return is 40% to 27%
15 year annual return is 41% to 26%
https://dqydj.com/stock-return-calculator/
That might be true, no idea, but I was actually making a point about the rapidly diminishingly fab competion and what that's going to mean for customers (and consumers).
Yes, there won't be an Intel when they're done. Previous customers will be using AMD x86 or ARM after suffering being screwed by Broadcom for a few years.
Hm, bad news for Oregon. From this article: https://www.oregonlive.com/silicon-forest/2024/12/intel-sput...
Oregon exports more than $30 billion worth of electronics annually, most of which are semiconductors manufactured in Intel’s Hillsboro factories.
----------
The company employs more people than any other business in the state. ----------
The state’s chip industry pays an average yearly wage of more than $150,000, according to government data, well over double the state average. Hefty pay from Intel is an especially big deal in Oregon, which relies on personal income taxes for the bulk of state revenue.
---------- ...Intel might consider selling its business off in pieces.
That could be calamitous in Oregon, turning Intel’s huge local campuses into satellite operations of other companies that don’t need the full range of corporate and technological functions that have been the foundation of Intel’s Hillsboro sites for decades.
“The people who work at Intel have to be really panicked,” Hutcheson said.
Intel’s research factories in Hillsboro support high-volume production sites in Arizona, Ireland, Israel and -- soon -- in Ohio. If that manufacturing network were someday to break up, it’s not clear what role would be left for the company’s top engineers and scientists in Oregon.
I wonder if this couldn't be the chance for Apple to make the definitive move towards vertical integration. If you think about it, it's the only missing part in their vertical integration story.
It would really be a sad day and the begin of truly dark era for the entire computing industry, though.
As an alternative... Amazon would probably be another interesting buyer. Amazon is another company dead serious about having their own silicon. And owning the main supplier of x86 chips as well in-house capabilities to manufacture their own Graviton chips would give them an incredible edge over other cloud computing providers.
Broadcom is a very hostile and abusive company. That would be the end of Intel chips.
Things Broadcom would do: sell you a chip, nerf if after a year and charge you a subscription to unnerf.
Look at what they did to VMWare. If you were a VMWare customer you were and are in deep shit right now. The fact this conversation between Intel and the Grim Reaper is taking place means everybody, from consumers to OEMs to clouds need to divest away from Intel today.
Sell what you have while you still can get some cash for it and remove it from consideration for any future plans.
Why would TSMC want to buy factories from Intel?
Intels plants are far from start of the art. Given their inability to compete and catch up with TSMC
Lots of people still use older processor nodes. For example Groq[1] uses 14nm chips for their inference chips. Older fab tech is cheaper to use (less demand and higher yields). Most chips made are on older processes, it's only the bleeding-edge CPU/GPU chips that use bleeding-edge fab tech.
[1] https://groq.com/wp-content/uploads/2024/08/GroqChip%E2%84%A...
I'm assuming TSMC has new and old lines. A 555 chip isn't done on the same line as an AMD Zen 5 chip. Permits, buildout, infrastructure, suppliers, you already get a big warehouse that makes chips and you can bring in your own people to optimize it on day one not year 4.
Intel has a couple of euv machines that tsmc could use and they could take Intel's current nodes and sell them as US made chips without threatening their own top nodes.
Ignoring their newer fabs for a minute (which do exist), because there's still a load of money to be made on reliable older nodes.
This data is a couple years out of date now but it still shows how relevant older nodes are: https://stockdividendscreener.com/technology/semiconductor/t...
> Why would TSMC want to buy factories from Intel?
I think the reason is obvious...it has no choice
because Trump said TSMC is stealing US chip biz, and threaten 100% tariff [1]. I think this is TSMC's attempt to avoid tariff
[1]https://www.pcmag.com/news/trump-intel-was-great-until-taiwa...
I wonder how much this is costing Intel in terms of enterprise sales? I'm guessing that headlines like this will accelerate the switch to AMD and ARM. Who wants to design servers around chips from a company that is not likely to exist in its present form? That opens up all kinds of supply chain and support concerns.
Revenge of the MBAs.
Anyone who buys Intel would probably be making the same mistake Boeing did with McDonnell Douglas.
My hope for China reaching parity on node size or at least get close to it grows each day.
Why are you hopeful of that?
price crash and the whole world getting access to the latest tech
Gift link: https://www.wsj.com/tech/broadcom-tsmc-eye-possible-intel-de...
Ironical that two companies led by the same CEO, the first one already acquired by BC and now the second one too!
Genuine question. Are there any examples of big companies getting bigger that ended well for the consumer?
I'm not seeing what Broadcom brings to the table here. They don't have chip mfg experience. Sure, they're a customer, but so what?
TSMC is not going to be motivated to build up a competitor who weakens the economic wall around Taiwan.
What I could see happening based on recent events is that Trump tells Taiwan they have a couple years of US security left. TSMC is then forced to consider putting more effort into mfg in the US(and other countries). What that would mean for the rest of the elec mfg supply chain is unknown. Taiwan is quite critical beyond just chip mfg.
Perhaps they're buying the IP part of Intel?
[dead]
Since the WSJ article seems hardwalled, we changed the URL from https://www.wsj.com/tech/broadcom-tsmc-eye-possible-intel-de... to a story that reports on it.
[flagged]