siliconbruh999
Well-known member
Pat is nowhere related to Intels decision now
Sources Say Intel Is An Acquisition Target
- Thread starter XYang2023
- Start date
Post in thread 'Intel Shortlists Suitors for Programmable Chip Arm Altera' https://semiwiki.com/forum/index.ph...programmable-chip-arm-altera.21712/post-80208
Mentioned here multiple times. The current valuation is off by a large margin. The market is an averaging machine.
Valuation (finance) - Wikipedia
The majority shareholders of Intel are long-term institutional investors with a high cost basis. For a simple comparison, it doesn’t make sense for Marvel to be valued higher than Intel. Nevertheless, Marvel can still serve as a benchmark for valuation.
Intel can also opt to implement further cost-cutting measures and limit the scope of its foundry operations to engineer a financial turnaround. It doesn’t need mergers and acquisitions to prop up its share value.
Artificer60
Well-known member
I don't know that I agree with Mr. Rasgon that there are no good answers here. I think that their plan is a good answer. The company won't survive without it. The bottom line is that Intel does not generate enough demand on their own to support long term cutting edge R&D with the market for their products shrinking. Intel needs the foundry volume to support the R&D costs. I don't think they need huge volumes in addition to Intel products to do that, but they need some additional volume. I believe that the companies looking to second source from TSMC would provide enough volume to fill in the gap that Intel can't fill on their own.
What they do need to do differently and what Gelsinger didn't seem to be willing to do, was to plan their fab expansion at a rate commensurate with their foundry customer growth. Yes, Intel will need more fabs, but they don't need all that capacity to be cutting edge. In fact most of their fabs will be relegated to trailing foundry products in a decade or so. Only a fraction of the capacity needs to be cutting edge and that is a very different environment for Intel.
Intel's real problems boil down to two things. First, near term cash flow, which is what I think got Gelsinger fired and the temporary CEO's have been given the mandate to address. Second is the impatience of the market. Gelsinger wasn't even given the 4 years he said it would take to get Intel's technology back to being competitive. This turn around was never going to be quick and while many people have given that fact lip service, when the rubber hits the road the market hasn't been willing to give them the time.
Your milage may vary.
David Zinsner previously mentioned that they only need a small external volume (compared to Intel's products) to make the foundry model work. Currently, in each conference call, they have mentioned that there are potential customers inquiring about 18A.
Artificer60
Well-known member
I would be far more concerned about the hit to morale that Intel would take if Musk acquired them. From what I've seen at X he has not won the hearts and minds of the employees. Intel's employee morale has got to be at an all time low and another significant blow could cause the whole house of cards to collapse. Whoever they bring in as the new CEO is going to have to walk pretty carefully for the first couple of years to win employee trust and convince employees that the company has a future. Rather than another massive bloodletting I would thing a more cautious approach of only allowing hiring for limited roles over the next couple of years would produce better results.
I'm also not sure that the bloat is the problem so much as it is a symptom of the real problem. Intel has had several major head count reductions over the last couple of decades and each time the bloat comes back. This leads me to believe that the issue is the bureaucratic culture that resulting in excessive hiring of middle managers. Unless the cultural issues are addressed, then cutting more heads isn't going to solve anything in the long term. Intel will just rehire when they finally get the ship righted and will start the whole cycle again.
CEOs tend to be egotistical. I've only worked for one over the decades who wasn't and yet was still very competent.
I think middle managers are probably a problem at Intel, but there are also bloated program management groups, bloated finance and legal groups, far too many very senior engineers and fellows for the results the company is achieving, a CTO office with an organization, when there never used to be one years ago, and bloated product management and marketing groups. I'm of a mind that Intel needs to completely rethink what their definition is of an effective organization.
As I have said, several companies will try 18A but not on big runners since the risk is high. TSMC is building out N3 and N2 and packaging as a result. Intel needs to fill fabs and the semiconductor industry needs fab alternatives. It may be a shotgun wedding but it really could be a strong marriage. There still are no signs of Samsung 2nm. There are three semiconductor ecosystem conferences (Chiplet Summit, DesignCon, and DVCon) here in Silicon Valley in the next three weeks. Let's see what I can find out. The rest of the media will be mailing it in.......
First: Selling assets could help near term cash flow. Altera and MobilEye and others. AMD did the same thing in their turn around. They sold fabs, packaging facilities, and real estate for near term cash if I remember correctly. Including their iconic HQ in Sunnyvale which is now high density housing.
Second: This is true but Pat was overly optimistic (5N4Y) and did not deliver on it. Intel did not need a cheerleader CEO and that is a big part of Pat's personality. I was also told that Pat's religious beliefs got in the way of management discussions. Praying on the outcome of Intel did not help.
That's a new data point. Did you hear about examples?
I felt the same way about religion.First: Selling assets could help near term cash flow. Altera and MobilEye and others. AMD did the same thing in their turn around. They sold fabs, packaging facilities, and real estate for near term cash if I remember correctly. Including their iconic HQ in Sunnyvale which is now high density housing.
Second: This is true but Pat was overly optimistic (5N4Y) and did not deliver on it. Intel did not need a cheerleader CEO and that is a big part of Pat's personality. I was also told that Pat's religious beliefs got in the way of management discussions. Praying on the outcome of Intel did not help.
PG mentioned something along the lines of, "God chose him for this job."
The job is very simple: maximize shareholder value while managing acceptable risks.
He blended company management, the priorities of the U.S. government, and religion into one, which was clearly the wrong approach. How can you compete with others when operating that way? It felt more like political leadership than the leadership expected from a publicly traded company.
I thought that approach reflected a serious misalignment.
As a side note, my mother once told me about PG. A priest from the U.S. who knew PG very well shared his story in one of their talks. I didn’t even know what to say when I heard that.
Yes, but ChatGPT proves me wrong yet again:
Pat Gelsinger, who served as Intel's CEO from February 2021 to December 2024, is known for his devout Christian faith. During his tenure, he advocated for creating inclusive environments that respect and accommodate various religious beliefs within the workplace. For instance, Intel has supported multiple faith-based employee resource groups, including ones for atheists and agnostics, to foster a culture of inclusivity.
While Gelsinger's leadership saw Intel's revenue decline from $72 billion in 2019 to a forecasted $52 billion in 2024, there is no evidence to suggest that his religious beliefs negatively impacted Intel's performance. The challenges faced by Intel during this period were more likely due to strategic business decisions and market dynamics rather than Gelsinger's personal faith.
Yes, I understand this but I am also talking about focus. Intel needs to focus and distractions such as MobilEye and Altera are not helping, my opinion. I did not mean to say that Intel should sell fabs, packaging, and real estate. That was what AMD needed to do. Intel needs to build competitive products and process technologies, fill fabs, and service foundry customers. Amongst other things of course but being a fabless semiconductor company is not one of those things.
That is so dumb. A faith-based resource group for atheists? I don't believe it. ChatGPT must be hallucinating. If it isn't, Intel needs a major corporate enema.
I always suspected that his religious beliefs influenced his hiring decisions. He seemed very reluctant to let people go. The restructuring implemented after the August earnings was overly generous, and it felt like he didn’t care about costs.Yes, but ChatGPT proves me wrong yet again:
Pat Gelsinger, who served as Intel's CEO from February 2021 to December 2024, is known for his devout Christian faith. During his tenure, he advocated for creating inclusive environments that respect and accommodate various religious beliefs within the workplace. For instance, Intel has supported multiple faith-based employee resource groups, including ones for atheists and agnostics, to foster a culture of inclusivity.
While Gelsinger's leadership saw Intel's revenue decline from $72 billion in 2019 to a forecasted $52 billion in 2024, there is no evidence to suggest that his religious beliefs negatively impacted Intel's performance. The challenges faced by Intel during this period were more likely due to strategic business decisions and market dynamics rather than Gelsinger's personal faith.
After the August earnings, he said, "We (I) have waited far too long to make some difficult decisions." However, Lipu had already suggested making the company more efficient. Why did he wait until the last minute to act?
I believe they are already independent companies, so it should be fine. As for real estate, it’s reasonable to divest those assets.
I do think keeping Mobileye and Altera under the same group is beneficial. Intel has a software-defined vehicle initiative, and there is synergy among them:
US to ban smart cars containing Chinese tech
BYD electric cars for export wait to be loaded onto a ship at a port in Yantai, in eastern China's Shandong province (-) (-/AFP/AFP) The United States finalized a rule Tuesday effectively barring Chinese technology from cars in the American market, taking aim at software and hardware from the...
semiwiki.com
That was not the problem. Even under Pat they said by end of decade that external foundry a minority of total wafer volumes. The demand drop of 5 18A fabs to 3, and also writing down 10nm tooling that Intel anticipated never again filling is incompatible with intel foundry service demand forecasts being too rosy. More evidence of this claim is that the Malaysia advanced packaging fab was started under BS's tenure as CEO. Yet the start up was recently delayed to align with market conditions. Combine this information with the collapse in Intel product revenue and the sheer absurdity of CCG telling the world that the PC TAM in 2022 and later was over 300 something million units per year (like it was during COVID) when modern data shows the 2023(act)-2025(est) TAM is in that 150-180M range. How is it possible for Intel foundry to not have gotten ahead of their skis if their customers CCG said their volume would be double and DCAI's volume is presumably less than a third (based on revenue and Intel saying mix shifted to higher asp products).
Only partially agree. The plurality of the Intel IDM will need to be on leading tech forever. If one looks at ARL/MTL like 40% of the content is on leading process (by the common N and N+1 definition), but 100% is on leading edge by TSMC'S definition (N7 and below). LNL and PNL are even higher leading edge content where only the PCH is on a trailing process. You are definitely right about disaggregation flattening out and extending the tail on internal wafer requirements for a given technology. Over time maybe you are right and the majority of Intel capacity will be on trailing processes but that will take a lot of time. Especially because Intel had talked about trying to be able to more easily convert between 3 18A and 14A fabs. To me this sounds like Intel preparing for not being able to fully fill their existing fabs once the IDM moves their capacity tail products onto the next node. But yeah what you suggest is definitely the long term goal. TSMC makes stupid amounts of money at comical margins on that trailing edge stuff all with minimal investment. I just think it will take a while before trailing edge foundry demand exceeds Intel IDM+leading edge foundry customer demand.
Artificer60
Well-known member
Completely agree he was overly optimistic and that resulted in some of the problems we are seeing now. However, IIRC correctly he announced 5N4Y in the summer of 2021. That gives Intel until summer of 2025 to complete this. It is also my understanding that Panther Lake is going to be available by the end of this year which means 18A will need to be in production no later than the middle of this year. So I'm not sure how you can say this has failed yet? If Panther Lake launches this year, I think it is fair to say that Intel hit this goal. Though, realistically I don't think we will know until the end of this year, unless Intel pushes the Panther Lake launch date out before then.
Artificer60
Well-known member
That is fair. I guess I was looking more at the willingness to swallow things like CCG's projected TAM in the first place and building projections on that. Every rumor I hear seems to point to Gelsinger being ousted because he wanted to invest in growth faster than the board felt was prudent.
You are correct, my comment was directed at the long term. I was thinking 10 years because at one node every two years that will make even 18A and TSMC 2nm N-5 nodes. But you raise a good point that in 10 years most of Intel's production will still be Intel products, which are inherently leading edge heavy. So the transition to being able to milk that tail for serious revenue will take quiet a bit longer than I was estimating.