Intel's Lost Decades: The Fall Of The Chipmaker

A post-mortem on how Intel ended up as the token laggard in the semiconductor industry

When the news about Pat Gelsinger stepping down from Intel dropped Monday morning, I was ready to crown it as the hottest news in business this week. Little did I know that the CEO of the $500 billion UnitedHealth Group would be assassinated in the middle of Manhattan. Anyways, Intel’s board ousted Gelsinger after a tanking stock price of over 50% just this year.

Q2 losses totaling $16.64 billion and massive layoffs of more than 15,000 employees worldwide have stained his tenure. After such a traumatic year, it's hard for a board that only cares about stock price and public perception to find a thick enough silver lining to keep Gelsinger around.

Pat Gelsinger took the reins after Brian Krzanicih’s horrendous tenure, which was mired in a lack of flexibility and open-mindedness. Brian’s team fell complacent with revenue growth, while net income was flat around $10 billion. Additionally, Intel's failure to double down on mobile and GPU computing left It grossly behind other giants (spoiler: it does not bode well when AI becomes a big thing).

Pat was applauded as the right choice for Intel. Previously the CTO, his deep engineering background was perfect for taking risky bets in R&D to set up Intel to recover their gap in competitiveness. However, 3 years have passed, Gelsinger is out, and Intel is again stuck in a rut. So, how did we get here?

Losing Their Edge

Ask any engineer who worked in California in the 80s and 90s and they’ll tell you that Intel was a Silicon Valley darling. Intel was the largest semiconductor manufacturer by revenue, had a majority market share, great financial prowess, and was recognized as a household driven by the “Wintel” boom.

Intel’s glory days were fueled by three consecutive CEOs - Robert Noyce, Gordon Moore, and Andy Grove. The mistakes started racking up when Intel brought in Paul Otellini, the first non-engineer as CEO. Otellini wasn’t too shabby while he was around. In 2012 under Otellini, Intel’s profits exceeded the total of key competitors Broadcom, Nvidia, Qualcomm, and Texas Instruments.

Where Otellini fell short was failing to bet on the right long-term investments. 2007 was a pivotal time for Intel. The iPhone came out and the company was met with a decision. Either they could reinforce their dominance in the computer chip market or bet on the burgeoning mobile chip market.

What if I told you they chose against mobile chip production simply due to a financial estimation error? Otellini, the supposedly “business-oriented” CEO, decided not to work with Apple to produce mobile chips because Apple’s ideal price was less than Intel’s forecasted cost. In hindsight, Otellini notes that the forecasted cost was wrong, and the volume driven by the mobile market was 100 times anyone's expectations.

Missing the mobile market came with a massive opportunity cost. Today, mobile accounts for over 33% of TSMC’s product line - even after the boom in high-performance computing chips. Part of missing out on mobile involved differences between Intel’s x86 architecture and the ARM-based architecture that were optimized for memory and power usage, and well suited for mobile devices. Rather than recognizing x86’s shortfalls when it came to smaller devices, Intel doubled down and believed their identity as a chipmaker and designer (IDM) would allow them to excel in mobile computing.

While Otellini dropped the ball on mobile, he did try to salvage it by presenting a wild idea to the board in 2005: acquiring Nvidia for $20 billion. The board hesitated and obviously, it did not go through. The indecisiveness marked the start of another failed opportunity. Intel would double down on CPU instead of GPU and ditch attempts to enter the graphics space. Unsurprisingly, Nvidia has emerged as a clear leader in both the consumer (gaming) and enterprise space with a commanding 88% market share and growing.

Source: Techspot

Splitting Attention

“I do not expect to join any company which is simply a manufacturer of semiconductors. I would rather try to find some small company which is trying to develop some product or technology which no one has yet done.”

In his resignation letter to Fairchild Semiconductor Company, Robert Noyce noted that he would rather work on an interesting product at a small company over a simple but large manufacturer. However, Intel is unique compared to other semiconductor companies because it is both a chip factory (fab) and a chip designer. This would be like if Nvidia and TSMC were rolled up into one company.

This worked for the first 30 years but has now led to concerns about their ability to move nimbly with changes in the market (ie producing mobile or cutting-edge graphic computing chips). It’s why Krzanich fell complacent and why AMD, Nvidia, and TSMC have overtaken Intel. As an Integrated Device Manufacturer (IDM), Intel splits its budget between two business lines, limiting the individual growth of each line. Hence, Intel moves slower than foundries like Samsung and TSMC and chip designers like Nvidia and AMD.

But Intel’s situation is not new. It turns out that operating a foundry costs billions to maintain and grow - AMD found this out the hard way. In 2008, AMD was in a similar boat. The company was bleeding cash and decided to spin off its Foundry division for $700 million. The results speak for themselves.

Source: Tim Culpan Substack

AMD’s spinoff went on to become GlobalFoundries - a $25 billion chip manufacturer. Maybe Intel can turn this around as an IDM, but the company seems better off divesting and doubling down on one or the other. Given Intel is close to production on their groundbreaking new 18A Panther Lake process, it’s more than likely their brand as a chip manufacturer will continue for the time being. So, what do you think Intel should do?

That’s all for today folks! If you enjoyed this piece covering Pat Gelsinger’s surprising retirement and Intel’s astronomical fall, hit the like button or drop a comment with your thoughts. Contact us by replying to this email if you have any insider takes or leave a comment on our Substack page!

Thumbnail creds: https://www.hatchwise.com/resources/the-history-of-the-intel-logo