It’s been a big year for silicon in Silicon Valley – but a brutal one for the company that owes the region its most nickname.

Intel, The 56-year-old chipmaker, co-founded by industry pioneers Gordon Moore and Robert Noyce and legendary investor Arthur Rock, suffered its worst year since its IPO in 1971, losing 61% of its value.

The opposite story happened at Broadcomthe chip company led by CEO Hock Tan is headquartered in Palo Alto, California, about 15 miles from Intel’s Santa Clara campus.

Broadcom’s stock price rose 111% in 2024 as of Monday’s close, its best performance ever. The current company is the product of an acquisition by Avago in 2015, which went public in 2009.

The driving force behind the divergent narratives was artificial intelligence. Broadcom took advantage of the AI ​​bandwagon while Intel largely missed it. The changing fortunes of the two chipmakers underscore how fleeting leadership is in the tech industry and how a few key decisions can result in market capitalization shifts worth hundreds of billions – or even trillions – of dollars.

Broadcom develops custom chips for Google and other major cloud companies. It also represents the essential networking equipment that large server clusters need to connect thousands of AI chips together. When it comes to AI, Broadcom has largely been in the shadows Nvidiawhose graphics processors or GPUs power most major language models developed at OpenAI, MicrosoftGoogle and Amazon and also enable the highest AI workloads.

Although Broadcom’s accelerator chips, which the company calls XPUs, are less well-known, they have become a key part of the AI ​​ecosystem.

“The reason it’s really skyrocketing is because they’re talking about AI, AI, AI, AI,” Eric Ross, chief investment strategist at Cascend, told CNBC’s “Squawk Box” earlier this month.

Intel, the dominant U.S. chipmaker for decades, has been largely excluded from AI. Its server chips lag far behind Nvidia’s, and the company has also lost market share to its long-time competitor Advanced micro devices while at the same time a lot of money is being spent on new factories.

Intel’s board ousted Pat Gelsinger as CEO on December 1 after a tumultuous four-year tenure.

“I think someone more innovative could have seen the AI ​​wave coming,” Paul Argenti, a professor of management at Dartmouth’s Tuck School of Business, said in an interview on “Squawk Box” after the announcement.

An Intel spokesman declined to comment.

Broadcom is now worth around $1.1 trillion and is the eighth U.S. technology company to surpass the trillion-dollar mark. It is the second most valuable chip company behind Nvidia, which has driven the AI ​​boom to a valuation of $3.4 trillion Apple among all public companies. Nvidia’s share price rose 178% this year, but actually performed better in 2023 when it gained 239%.

Until four years ago, Intel was the world’s most valuable chipmaker, approaching a market capitalization of $300 billion at the start of 2020. The company is now worth about $85 billion, was just ousted from the Dow Jones Industrial Average – replaced by Nvidia – and is in talks to sell key parts of its business. Intel is now ranked 15th in terms of market capitalization among semiconductor companies worldwide.

After Avago and Broadcom merged in 2015, chips for TV set-top boxes and broadband routers were the combined company’s largest business. Broadcom still makes Wi-Fi chips used in laptops as well as the iPhone and other smartphones.

After a failed purchase offer for the mobile chip giant Qualcomm In 2018, Broadcom turned its attention to software companies. The culmination of the buying spree was the announced takeover of server virtualization software provider VMware for $61 billion in 2022. Software accounted for 41% of Broadcom’s $14 billion in revenue last quarter, thanks in part to VMware.

What’s exciting on Wall Street is Broadcom’s role in working with cloud providers to develop custom chips for AI. The company’s XPUs are generally simpler and cheaper to operate than Nvidia’s GPUs and are designed to efficiently run certain AI programs.

Cloud providers and other major internet companies spend billions of dollars each year on Nvidia’s GPUs so they can build their own models and run AI workloads for customers. Broadcom’s success with custom chips is leading to an AI spending showdown with Nvidia as hyperscale cloud companies seek to differentiate their products and services from competitors.

Broadcom’s chips aren’t for everyone, as only a handful of companies can afford to design and build their own custom processors.

“You have to be a Google, you have to be a MetaYou must be a Microsoft or a oracle to take advantage of these chips,” Piper Sandler analyst Harsh Kumar told CNBC’s “Squawk on the Street” on Dec. 13, a day after Broadcom’s earnings release. “These chips are not intended for everyone.”

While 2024 was a breakout year for Broadcom – AI revenue rose 220% – the month of December pushed it to record levels. The stock was up 45% this month as of Monday’s close, 16 percentage points better than its best month yet.

At the company’s earnings call on December 12, Tan told investors that Broadcom had doubled shipments of its XPUs to its three hyperscale providers. Most famously, Google is relying on the technology of its Tensor Processing Units (TPUs), which are used to train Apple’s AI software released this year. According to analysts, the other two customers are TikTok parent ByteDance and Meta.

Tan said companies could spend between $60 billion and $90 billion on XPUs within about two years.

“We believe that in 2027 each of them plans to deploy 1 million XPU clusters in a single fabric,” Tan said of the three hyperscale customers.

In addition to AI chips, AI server clusters require powerful network components to train the most advanced models. Networking chips for AI accounted for 76% of Broadcom’s $4.5 billion in network revenue in the fourth quarter.

Broadcom said that overall, about 40% of its 2024 semiconductor sales of $30.1 billion would be related to AI and that AI sales would rise 65% to $3.8 billion in the first quarter.

“The level of success of the hyperscalers in their initiatives here is clearly an area that is up for debate,” wrote Cantor analyst CJ Muse, who recommends buying Broadcom shares, in a Dec. 18 report. The focus here continues be a significant boon to those who rely on custom silicon.”

The worst market year for Intel before 2024 was 1974, when the stock fell 57%.

The foundation for the company’s recent stumbles was laid years ago, when Intel failed in mobile chips to Qualcomm, ARM and Apple.

Thanks to its productive manufacturing relationships, rival AMD began to gain market share in the critical PC and server CPU markets Taiwan Semiconductor Manufacturing Company. Intel’s manufacturing process has lagged behind for years, resulting in slower and less energy-efficient central processing units, or CPUs.

But Intel’s most expensive touch is in AI – and that’s a big reason why Gelsinger was removed.

Nvidia’s GPUs, originally designed for video games, have become crucial hardware in the development of performance-hungry AI models. Intel’s CPU, once the most important and expensive part of a server, has become an afterthought in an AI server. The GPUs that Nvidia will ship in 2025 don’t even require an Intel CPU – many of them are paired with an ARM-based chip developed by Nvidia.

With Nvidia reporting revenue growth of at least 94% in the last six quarters, Intel was forced into downsizing mode. Sales have declined in nine of the last 11 periods. Intel announced in August that it would cut 15,000 jobs, or about 15% of its workforce.

“We are working to create a leaner, simpler and more agile Intel,” Chief Executive Frank Yeary said in a Dec. 2 press release announcing Gelsinger’s departure.

A big problem for Intel is the lack of a comprehensive AI strategy. The company has touted to investors the AI ​​capabilities of its laptop chips and released an Nvidia competitor called Gaudi 3. But neither the company’s AI PC initiative nor its Gaudi chips have gained much traction in the market. Intel’s Gaudi 3 sales missed the company’s $500 million target for this year.

Late next year, Intel will release a new AI chip codenamed Falcon Shores. It will not be based on the Gaudi 3 architecture but will be a GPU.

“Will it be wonderful? No, but it’s a good first step to finalize the platform,” said Michelle Holthaus, interim co-CEO of Intel, at a Barclays financial conference on December 12.

Holthaus and his interim co-CEO David Zinsner have vowed to focus on Intel’s products, leaving the fate of Intel’s costly foundry division unclear.

Before he left, Gelsinger championed a strategy in which Intel both entered the semiconductor market and made chips to compete with TSMC. In June, Gelsinger told CNBC at a conference in Taipei that once its factories are up and running, Intel wants to build “everyone’s AI chips” and offer companies like Nvidia and Broadcom an alternative to TSMC.

Intel announced in September that it plans to transform its foundry business into an independent entity with its own board and the ability to raise debt capital. But right now Intel is Intel’s main customer. The company said it does not expect significant sales from external customers until 2027.

At the Barclays event this month, Zinsner said the separate board for the foundry business would be “repealed today.” More broadly, he noted that the company wants to remove complexity and associated costs wherever possible.

“We will continually evaluate where we spend money to make sure we get the appropriate return,” Zinsner said.

REGARD: Intel plans to take its chip subsidiary Altera public