Intel’s long-awaited turnaround seems even further away now that the company reported really bad earnings for the first quarter. Investors were so disappointed that they pushed the company’s shares down by 10% on Friday, making them the lowest they’ve been all year.
Although Intel’s revenue isn’t getting smaller anymore and it’s still the biggest maker of the chips that power PCs and laptops, its sales in the first quarter were lower than expected. Intel also said that it doesn’t expect to do much better in the second quarter, which suggests that demand for its products isn’t strong.
This is tough for CEO Pat Gelsinger, who’s only in his fourth year running the company but Intel’s problems didn’t just pop up overnight. They’ve been building for decades.
Before Gelsinger came back to the company in 2021, Intel, which was once a huge name in Silicon Valley, had fallen behind other companies, especially those overseas like Taiwan Semiconductor Manufacturing Co., when it came to making computer chips. Now, Intel is spending a lot of money every quarter to try to catch up.
“Our top priority was to speed up our efforts to close the technology gap caused by more than a decade of not investing enough,” Gelsinger told investors on Thursday. He said the company is still planning to catch up by 2026.
Investors aren’t convinced. Intel’s stock is the worst-performing one in the S&P 500 this year, down by 37%. Meanwhile, the two best-performing stocks in the index are chipmaker Nvidia and Super Micro Computer, which makes servers powered by Nvidia’s chips.
Intel, which used to be the most valuable chipmaker in the U.S., is now just a fraction of the size of Nvidia. It’s also smaller than Qualcomm, Broadcom, Texas Instruments, and AMD. Intel was the biggest chip company in the world for a long time, but its revenue has been going down for seven quarters in a row, and Nvidia passed it last year.
Gelsinger is trying a risky new plan. Intel won’t just make its chips anymore, but it will also make chips for other companies that don’t have their factories — including Nvidia, Apple, and Qualcomm. But for this to work, Intel needs to be the best at making chips, which it calls “process leadership.”
Other chip companies want an alternative to Taiwan Semiconductor so they’re not depending on just one company. U.S. leaders, like President Biden, say Intel is a crucial American chipmaker and is important for the country’s chip supply.
“Intel is a big, well-known chip company that has been the leader for many years,” said Nicholas Brathwaite, a managing partner at Celesta Capital, which invests in chip companies. “I think it’s a company that’s worth trying to save, and they have to come back to being competitive.”
“I think everyone has been hearing them say things will be better next quarter for two, three years now,” said Counterpoint analyst Akshara Bassi.
Intel has been dropping the ball for years. It missed the boat on making chips for mobile phones when the iPhone came out in 2007. It’s also not been a part of the AI excitement, while companies like Meta, Microsoft, and Google have been buying as many Nvidia chips as they can.
How Intel Lost the iPhone and Fell Behind in the Mobile Revolution
The iPhone nearly used Intel chips. Steve Jobs and Paul Otellini discussed it, but Apple passed due to Intel’s slow pace and concerns about sharing technology. Instead, Samsung chips powered the iPhone initially. In 2008, Apple bought PA Semi and introduced its chip in 2010.
Smartphones, led by Arm-based chips, surpassed PCs in popularity. Arm chips, like those from Apple and Qualcomm, consume less power, making them ideal for small devices. Apple’s chip orders boosted TSMC’s manufacturing prowess, surpassing Intel.
Intel’s PC chip market share shrank, partly due to Apple’s shift to Arm-based Macs in 2020. Attempts to enter smartphones with the Atom chip failed. Now, Windows laptops with Arm-based chips are emerging. Intel’s mobile missteps set the stage for a challenging decade.
The Evolution of Transistors and Intel’s Struggle to Keep Pace
From the 1971 Intel 4004 with 2,000 transistors to today’s chips with billions, faster processors mean more transistors. Shrinking them allows more on a chip, measured by the “process node.” The smaller, the better.
Intel’s pride lies in regularly delivering smaller transistors. Engineers, especially at Intel, were celebrated. Moore’s Law, coined by Intel’s Gordon Moore, promised computing power doubling predictably, roughly every two years.
Intel’s “tick-tock” development meant a new chip every two years. But in 2015, CEO Brian Krzanich delayed Intel’s 10nm process. The next process, 7nm, faced delays too, leading to Gelsinger’s takeover.
Meanwhile, rival AMD, designing chips in California and manufacturing with TSMC, soared. AMD’s chips became competitive or better than Intel’s, especially in servers. By 2022, AMD overtook Intel in market cap and server CPU market share.
The GPU Revolution: Nvidia’s Lead, Intel’s Ambitions
Originally for gaming, GPUs found a new role in AI computations. Nvidia’s sales soared after the success of ChatGPT in 2022, with GPU-based servers becoming popular for AI tasks.
In these servers, Nvidia’s GPUs often overshadow the Intel CPU, a shift from older setups. Nvidia’s latest “Blackwell” GPU even excludes Intel, pairing two B100 GPUs with an Arm-based processor.
Most Nvidia GPUs for AI are made by TSMC in Taiwan, using advanced techniques.
Intel lacks a direct GPU competitor to Nvidia but offers the Gaudi 3 AI chip, developed after acquiring Habana Labs in 2018. Manufactured on a 5nm process, Intel relies on external foundries.
Intel aims for $500 million in Gaudi 3 sales this year, while AMD expects $2 billion from AI chip revenue. Analysts predict Nvidia’s AI GPU sales to hit $57 billion in the second half.
Intel is investing in a massive fab in Ohio, set to produce leading-edge AI chips, potentially even for Nvidia, with government support. CEO Gelsinger believes it will be operational in 2028.
Intel’s Uphill Battle: Gelsinger’s Pursuit of Semiconductor Dominance
Under Gelsinger’s leadership since 2021, Intel aims to catch up to TSMC with “four nodes in five years,” a challenging task he’s dubbed a “death march.” Progress is being made, with plans to match TSMC’s 2nm chips by 2026.
However, it’s expensive. Intel’s foundry division reported a $2.5 billion operating loss on $4.4 billion in sales, with a $7 billion loss in 2023.
While few companies have committed to using Intel’s fabs, Microsoft plans to manufacture server chips with them, already securing $15 billion in contracts.
Regaining leadership in transistor size is crucial for Intel. Gelsinger sees this as the key to Intel’s resurgence, believing it will rebuild customer trust.
