by Zain Jaffer
Author’s note: I hold some of the stocks mentioned here, mostly in mutual funds. My previous background is in mobile advertising, where my previous startup was acquired by Blackstone in 2019. See https://www.blackstone.com/news/press/blackstone-closes-acquisition-of-vungle-a-leading-mobile-performance-marketing-platform/
Anyone who has worked in the Information Technology (IT) sector since the 80s has of course heard of Intel (NASD: INTC), the semiconductor giant that has dominated personal computer and server microprocessors these past few decades. Along with software giant Microsoft (NASD: MSFT), this “Wintel” duo has been at the forefront of computing.
Intel started after its founders Robert Noyce and Gordon Moore left Fairchild Semiconductor to form their own startup in 1968. Initially they started with memory chips, but saw a different future and pivoted to microprocessors, even when their memory business was still doing well. Eventually they caught the personal computing wave of the 80s with Microsoft when IBM (NYSE: IBM) picked these two companies as the core of their first PC [https://timeline.intel.com/].
Their late famous CEO Andrew S Grove popularized the concept of Strategic Inflection Points in the life of a company in his famous business book Only the Paranoid Survive [https://www.penguinrandomhouse.com/books/72469/only-the-paranoid-survive-by-andrew-grove/]. Grove argued, as Intel had done previously in its pivot to microprocessors from memory, that every CEO and company needed to be able to recognize when the tides of change had affected the fundamentals of a business, and act in time to keep the business alive to fight into the future.
Unfortunately it seems that Intel has not taken its own advice. Aside from its traditional rival AMD [NASD: AMD] its leadership in PCs at the lower end is being threatened now by mobile chip company ARM [NASD: ARM] [https://finance.yahoo.com/news/why-arm-chips-pose-a-threat-to-intel-and-amds-pc-dominance-180840226.html].
More importantly, the field of Artificial Intelligence (AI) requires Graphics Processing Units (GPUs) made primarily by Nvidia [NASD: NVDA]. GPUs are massively parallel processors that can handle huge training data sets in a fast manner. Although Intel (and AMD) processors can also handle parallel tasks as these have multiple cores, these are more optimized for office applications such as Word, Excel, Powerpoint, and not really for training huge AI data sets [https://blogs.nvidia.com/blog/why-gpus-are-great-for-ai/]. For AI inference on the user side, both AMD and Intel processors are fully capable, and many new PC models these days bill themselves as AI PCs.
However tech giants like Meta, Google, and Tesla are spending billions to dominate AI by buying Nvidia processors like their new Blackwell [https://www.nvidia.com/en-us/data-center/technologies/blackwell-architecture/] superchip. Or like Amazon, they are attempting to build their own [https://aws.amazon.com/machine-learning/inferentia/]. Although Intel has its own Gaudi chip [https://www.intel.com/content/www/us/en/products/details/processors/ai-accelerators/gaudi.html] clearly it has not gotten as much mindshare from buyers as Nvidia.
Intel has also gotten into the contract wafer fab (foundry) business that has long been dominated by Taiwan Semiconductor Manufacturing Corporation [NYSE: TSM] [https://www.intel.com/content/www/us/en/foundry/overview.html] as a hedge against threats in the South China Sea. It has also gotten a huge financial boost from the Biden administration to build new advanced wafer fabs in new places like Ohio, that will backup its Arizona and New Mexico fabs [https://www.intel.com/content/www/us/en/newsroom/news/us-chips-act-intel-direct-funding.html].
Unfortunately in early August 2024, Intel’s stock fell precipitously even after receiving the US government funding [https://www.foxbusiness.com/markets/intels-massive-job-cuts-after-receiving-taxpayer-money]. Recession fears as well as analyst fears that AI spending may not produce immediate returns contributed to this fall, along with other factors. To that end, Intel CEO Pat Gelsinger announced cutbacks in plans and spending, and will layoff 15,000 workers [https://www.washingtonpost.com/technology/2024/08/01/intel-layoffs-chips/] as it struggles to pivot in this new AI era.
Perhaps it is because they have been enjoying much success over these past few decades and have become comfortable with corporate and public spending on personal Wintel computers and servers that they totally missed the AI wave, and have been chasing too many businesses such as the foundry business.
If you think about it, their x86 microprocessor architecture has been around since the 80s [https://www.lenovo.com/ph/en/glossary/x86/]. They have kept it alive in all their processors even up to the present just to maintain backward compatibility for old documents and spreadsheets.
Perhaps it is time to leave their previous successes and glory days in the past and chart a new course. Maybe they should reread their late CEO Grove’s advice to be paranoid in order to survive.
###