In Intel's case, share buybacks would make sense, especially at current levels.
Are you kidding me? Buybacks shares now? Their client business and data center CPU market is under the the dumps now. Even without the external foundry initiative, just from the fact there was downturn in PC market, DC CPU disruption form AI and the fact they are outsourcing wafers from TSMC, they would have negative net income now and cash flow would be pressured. Buybacks are for when your companies fundamentals are doing well but the stock price is depressed irrationally from its fair value. Intel is not in that position and would have been here even without the external foundry initiative.
He chose a high capital intensity, long-duration, high risk, and low-return strategy for Intel. Such decisions resemble those made by governments. With that amount of capital, Intel could have competed with Nvidia—and probably with more visibility.
The IDM 2.0 strategy is formulated when supply chain issues related to COVID are full blown in 2021/2022. It was the right strategy at that time. You are looking backward with what has transpired now. Hindsight is 20/20. Anyone looking at what happened in 2023-2024 can easily suggest in 2025 that Intel should have focused on AI and went fabless even instead of doing 5N4Y and IDM 2.0.
Given the capital Intel used to have, I believe it should have been able to compete with Nvidia. It just needed the right person to lead that effort. Once again, this reflects PG's management style—particularly in his hiring decisions.
What is this capital you are talking about? Intel have roughly the same cash equivalent to what they had pre PG. Do you really think Intel is not competing in AI because of just lack of cash or capital? I honestly think you have very surface level understanding of how these markets work especially semiconductors and are thinking about software companies. You can't spin up an AI semi business in matter of 2 years. What nVidia has achieved is culmination of 10+ years of investment in capital and r&d. AMD at least had a working DC GPU (MI250) already focused on HPC & AI which they can capitalize on. Intel had what Max GPUs and PVC, both are failures even in HPC & AI!! Only thing they had going was Gaudi 2 and they marketed it and couldn't even make $500 mil. Gaudi 3 we will see how it goes this year.
You are talking as if Pat Gelsinger did not effort to capture AI, he spun up the AI PC effort and Gaudi. Both did not make a difference. Falcon shores - I have no idea why it was cancelled probably not competitive and realized training focused market is gone and should focus on rack scale GPGPU to compete. I can only guess.
Lets say Intel used their cash to an M&A in 2021 to acquire an AI winner. What can they buy in 2021/2022 for $40B (50% cash and 50% debt/equity) that could have given the upper hand in AI GPGPU for Intel? Broadcom - they were a $200B company in 2021! If pivoting focus in AI is so easy, all the AI semi startups now would be raking in billions of dollars of revenue.
On a side note - look at the cash burn under Bob swan and BK!
Is that just 10nm and buybacks?
I left a message on his Twitter feed early on, asking him to focus on AI and GPUs like Nvidia—possibly before or around the time ChatGPT was introduced. I'm sure many others made similar suggestions. But he chose to go his own way.
That is a good idea. I should have left a message on Texas Instruments CEO twitter page as well asking them to focus on building AI GPUs instead of building microcontroller/ fab capacity in US. That would have helped them well in these troubled times of industrial & automotive market slump . You can ask a peacock to sing but it is not going to happen anytime soon but it can dance well. Intel did not have anything at that time to meaningfully compete in AI and emergence of AI resulted in ~$5Billion loss of revenue in their core DC CPU markets. This is classic case of disruption. Pat Gelsinger had 1.5 years to see this disruption coming but 1.5 years unfortunately not long enough to make a difference even if he foresaw it all the while his focus were on something else (IDM 2.0 & 5N4Y) and fixing a shitshow roadmap that previous CEOs left behind.
If AMD can win some shares with their ROCm, Intel certainly could do it when it had enormous capitals than all its peers.
I want you to honestly answer why Qualcomm is finding it difficult to sell their AI notebook CPUs with full backing of Microsoft? Why Intel's ARC alchemist launched in 2022 around the time ChatGPT launched failed miserably? Why Intel's ARC graphics cards are only <1% of total market share? B580 for all its goodness still not definitely outsell comparable AMD Radeons nevermind the Geforce RTX! And Intel is mostly selling these at or near cost to gain a entry foothold. If Intel can't meaningfully compete in client graphics cards with GPUs, why should we expect they would compete meaningfully in DC GPGPUs? It is wishful thinking that Intel could pivot to AI GPUs and gained any market share.
Hence, in my opinion, focusing on AI would be easier than focusing on the foundry business. As CEO, he should prioritize maximizing shareholder value. The AI path is not only easier but also offers greater potential rewards.
Hmm!! lets see! Intel was a successful CPU/ leading edge semi-manufacturer (although with some recent missteps) looking at all the supply chain issues with COVID and geopolitical implications of all the leading edge capacity located in a tiny island located near a nuclear superpower that hell bent on invading that tiny island, IDM 2.0 is an easy strategy that allows them to develop leading edge node for Intel and share a portion of the R&D costs with the external foundry customers. That foundry business was and is a lucrative market as TSMC has shown.
On the other hand, in 2021 the AI GPU market was nearly 0 and only after Nov 2022 increased to multibillion $$ business on the back of cash rich balance sheets of 3-4 hyperscalers. And once a strategy is selected in semi manufacturing business it is not easy to pivot that ship in matter of months.
Intel's trouble is mainly from downturn in PC business (& not planning for that), factory underutilization of fab due to outsourcing of products to TSMC and loss of revenue in DC CPU business. These caused them to lose revenue and profit also cash while they were making increased investment in trying to become a Foundry. If you say Foundry initiative was poorly executed by overexpansion and lack of investment in IP & EDA, then I agree to some extent (I believe the IP & EDA part is the biggest mistake) but IDM 2.0 was the best strategy Intel could have selected in 2021 given the prevailing conditions.
Also on the other front, under PG Intel's products have improved a ton, product road map has improved a ton, Intel's node process have improved a ton. That has to mean something for Intel. With PC and DC CPU business showing life again, they will carry Intel P&L for a while. Exactly why I don't think Pat Gelsinger was not a failed CEO. He has done good things and made mistakes as well.
Lip Bu Tan has inherited a solid business with lots of tailwinds but with a bad stock price. Intel 18A has couple of design wins to share the R&D costs, 18A-P IP Ecosystem is being expanded and 14A was already pivoted to be fully industry standard EDA compatible under PG (early 14A PDKs were already released to customers in Q1'25 itself for feedbacks, this is not something you can do in 4 weeks in the job).
