You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
There is an interesting comment on LinkedIn by a former Intel employee suggesting that Nvidia should buy Intel. I remember when Intel was looking for a CEO (three CEO's ago) I suggested that Intel should buy Nvidia for the technology and for the CEO (Jensen). What a different world this would be if that would have happened?!?!?!?
Why NVDA should buy INTC...
Intel has been struggling a lot lately with it's stock price down to a low of $18.84 from over $50 this past year. I remember hearing old timers at Intel refer to the company as a manufacturing company, not a microprocessor company. The true strength of Intel is its production and packaging. The company even famously pivoted from producing DRAM to processors in the 1980's. Microprocessors served them well for decades as computation and Internet scaled. However, the centrality of the microprocessor in computation today is diminishing, displaced by AI. Intel is trying to develop it's own AI hardware with very limited success, and the company enterprise valuation has shrunk to $110B
Nvidia meanwhile, with its $2.6T enterprise valuation has the world's most lucrative ASIC product but can't produce enough of them. It seems to me like a perfect opportunity for Nvidia to buy Intel and complete their pivot to AI from microprocessors. I think it could do this on a all stock basis even. Nvidia would acquire the ability to manufacture it's own chips which could unleash new capabilities and profits. There are numerous other benefits like preventing Intel from failing and reshoring manufacturing of critical technology to America
Intel is too important to fail and I am bullish on it's long term prospects.
I think mixing chip fabrication and chip design in the same company has always been a recipe for poor decision-making. Finding a CEO competent to lead a chip foundry and a hardware-software design company is very unlikely as far as I can tell. It takes a computer scientist to tell when the design proposals are compelling, and a materials scientist or a chemical engineer, perhaps an electrical engineer, to understand when the fab proposals are compelling or not. The two knowledge bases appear to be orthogonal. And I don't think Hock Tan is an exception to my generalization. Broadcom already has a specialized fab for indium-phosphide components used in fiber optics, so he knows first-hand how different the fields are. I think Tan is too smart to try to manage Intel.
Let me clarify. I think Hock Tan is the right person to buy Intel, break it up, sell off the fabs and underperforming businesses, and purge the company of useless MBAs.
Leaving aside the likely antitrust angle, why would nVidia want to take on all Intel's problems right now ? Or anyone else for that matter ?
The chip companies that are doing well right now are highly focused on specific product areas and markets.
Intel, by comparison, is starting to look a little like one of those 1970s conglomerates which plays in lots of different areas, but dominates in none. I'm not that convinced by the "too big to fail" argument. Eventually that runs out of road - is GM still too big too fail ? It's the threat of actual failure that keeps companies honest and competitive. [added] Indeed it may well be that the absence of the threat of failure for over 20 years is why Intel is where it is today.
Let me clarify. I think Hock Tan is the right person to buy Intel, break it up, sell off the fabs and underperforming businesses, and purge the company of useless MBAs.
That makes a lot more sense, but what Tan specializes in is buying a smaller company with a captive customer base, shrinking various "overhead" groups (like finance, HR, sales, etc.), raising prices or complicating software licenses, and thereby raising gross margins. It is private equity companies that specialize in buying companies and selling off pieces. It take a very brave PE company to swallow the risks with Intel.
That makes a lot more sense, but what Tan specializes in is buying a smaller company with a captive customer base, shrinking various "overhead" groups (like finance, HR, sales, etc.), raising prices or complicating software licenses, and thereby raising gross margins. It is private equity companies that specialize in buying companies and selling off pieces. It take a very brave PE company to swallow the risks with Intel.
Intel is a smaller company with a captive customer base and bloated overhead costs.
AVGO market cap = $735b, INTC = $85.6b
It's not that big of a risk, they just need to get rid of the foundry side of the business, which they could probably sell to private equity. They can also finish spinning of MobileEye, spin out or sell Intel Capital, sell off the GPU business ect. In the end AVGO would just retain the x86 business which is a cash cow, and some software businesses.
I don't think so. As I pointed out in another post recently, Intel has been on the ropes several times. I also don't think it's stupidity, well, mostly not. (Itanium may one exception.) Sometimes the financial costs or various risks of taking the path you really want to take causes you to make decisions that are compromised, and the results of those compromises leave you wishing you had made a different decision. I'm pretty sure Paul Otellini felt that way with his decision not to take on Apple. For example, I know that part of the Apple discussion in Intel was whether to build one or more fabs to accommodate Apple's volumes. Even back then, fabs were a multi-billion dollar effort, Apple was trying to drive a hard bargain, and Intel's fab question back then was to determine how many fabs you thought you might need in 3-5 years, and then round down to ensure high utilization. iPhone volume in 2011 was more than 70 million units per year. iPad was another 32 million. Intel would have had to add a lot of manufacturing capacity, and learn to be a foundry at the same time (when they weren't), to bet on Apple. I'm honestly not sure what I would have done in PSO's place.
Intel may have a lower market capitalization, but more than twice the number of employees, and Broadcom CAPEX is less than $500M. I think swallowing Intel is tougher than you think.
Intel is a smaller company with a captive customer base and bloated overhead costs.
AVGO market cap = $735b, INTC = $85.6b
It's not that big of a risk, they just need to get rid of the foundry side of the business, which they could probably sell to private equity. They can also finish spinning of MobileEye, spin out or sell Intel Capital, sell off the GPU business ect. In the end AVGO would just retain the x86 business which is a cash cow, and some software businesses.
And whiywould buy it no private equity will own a Capital Hog Nvidia deal won't go through that is 100% guaranteed also the Billions of investment for which they have been given grant and money has already been invested unless Spinned off as USMC
And whiywould buy it no private equity will own a Capital Hog Nvidia deal won't go through that is 100% guaranteed also the Billions of investment for which they have been given grant and money has already been invested unless Spinned off as USMC
Private equity would buy given the right terms. Look at GloFlo, it traded hands through private equity a number of times. The thing about private equity, they know how to make money off failing businesses.
Intel may have a lower market capitalization, but more than twice the number of employees, and Broadcom CAPEX is less than $500M. I think swallowing Intel is tougher than you think.
If the choice is Elon+government vs the fabs going bankrupt, the cleanrooms getting demolished, the tools getting sold off to China and all the Intel fab employees out on the street facing almost perpetual unemployment, I think I would go with Elon running US owned fabs. Elon is a brilliant engineer leader for what it’s worth.
If the choice is Elon+government vs the fabs going bankrupt, the cleanrooms getting demolished, the tools getting sold off to China and all the Intel fab employees out on the street facing almost perpetual unemployment, I think I would go with Elon running US owned fabs. Elon is a brilliant engineer leader for what it’s worth.
Give fabs to TSMC or pay them to take them (or to GF and TSMC). Intel gets capacity in exchange
nvda buys product team. Intel product team starts to resign 6 months in after finding out ALL NVDA Development timelines are half as long as Intel development timelines.
Intel Engineers complain "we normally spend 2 years randomly changing the product targets and markets every week and then telling engineering to start over.... how come that isnt in nvda timeline?"
Jensen says "not even I can fix this!" ..... then he sells Intel product group. Everyone is happy again.
If the choice is Elon+government vs the fabs going bankrupt, the cleanrooms getting demolished, the tools getting sold off to China and all the Intel fab employees out on the street facing almost perpetual unemployment, I think I would go with Elon running US owned fabs. Elon is a brilliant engineer leader for what it’s worth.
You are under the assumption no other company will come in and buy out Intel if Intel financials becomes extremely dire. You are also under the assumption that China will be getting the fab tools which under current political environment most likely will not happen. It is more likely TSMC will be getting the fab as they want the extra fab capacity. As for Elon being an engineer, he is more of a sales person than an engineer in my opinion and one with no experience in semi-con at all. There are more capable people taking over the Intel fab business than Elon if you want the US Government to take over Intel.
The probability of the US government nationalizing Intel or even part of it is nil. There are examples of the government nationalizing businesses, but they are from long ago, companies like railroads in Alaska. Seizing the fabs, as in using eminent domain powers, seems dumb.
