Intel's (INTC) solid third-quarter earnings report briefly extended its share price's upward momentum over the past week. But multiple Wall Street analysts say the company's vital manufacturing business is far from a turnaround and could pose a long-term risk to its stock price.
The venerable chipmaker, whose central processing units (CPUs) are widely used in data centers and consumer electronics such as laptops, reported September quarter profit and revenue that beat analysts' expectations on Thursday. The financial report shows that the company's financial situation has improved after the U.S. government, SoftBank (9984.T) and Nvidia (NVDA) made large investments in Intel and injected large amounts of capital.
Intel shares rose as much as 8% in after-hours trading following the earnings release, but ended only slightly higher on Friday.
"We understand the desire to declare victory for this troubled company, but this battle is far from over," Bernstein analyst Stacy Rasgon wrote in a note to investors following the earnings release.
At the heart of the problem is Intel's cash-burning manufacturing business — which the company opened to outside customers in 2021 with the launch of Intel Foundry Services (IFS) — is still far from profitable, and its long-term success is unclear.
In the three months ended September 27, Intel's foundry business had revenue of $4.2 billion and its loss narrowed to $2.3 billion, an improvement from the $5.8 billion loss in the same period last year. But analysts expect IFS's fourth-quarter loss to widen to $2.5 billion, while revenue shrinks to $4.1 billion, according to Bloomberg data.
The key issue facing the business is the difficulty in obtaining large order commitments from customers. Rasgon said he calculated that only $8 million of the foundry business's $4.2 billion in revenue came from external customers.
This problem was evident last quarter: Intel had made it clear that its latest 18A manufacturing process would be mainly used for internal product production due to its failure to attract external chip design companies such as Nvidia and Broadcom (AVGO). In the future, IFS will try to attract external customers with the next-generation 14A manufacturing process, so the process will be critical to the success or failure of the foundry business. But CEO Lip-Bu Tan reiterated in a conference call with analysts that the company will only add 14A manufacturing capacity after seeing customer demand.
Chen Liwu said on Intel's earnings call on Thursday that the company was "engaging with potential external customers" and was "encouraged by the initial feedback."
But Rasgon noted that “the 14A process is still a long way off.”
To make matters worse, just as Intel's manufacturing business is in trouble, its product business continues to lose market share to Advanced Micro Devices (AMD). Analysts said these two factors mean that Intel's stock price may face the risk of lack of momentum once the short-term market enthusiasm subsides.
Deutsche Bank analyst Ross Seymore wrote in a note to clients on Friday: "While Intel may remain an event-driven stock in the short term - potential news such as foundry partners, artificial intelligence cooperation, new product launches and other potential news may boost market optimism, we believe that its stock price will likely face headwinds when the market finally returns to focusing on fundamentals."
Intel said its own manufacturing unit is critical to the U.S. supply chain, given that most of the world's computing chips are produced in Taiwan by rival Taiwan Semiconductor Manufacturing Company (TSM). That view has paid off: In August, the U.S. government announced it would acquire a 9.9% stake in the chipmaker.
However, there are also views that TSMC has committed to investing US$165 billion in new factories in the United States, which will help reduce supply chain risks, making Intel's above argument untenable.
"We think investors may believe that Intel's third-party foundry business can be profitable, but we don't think so - as we believe Intel's foundry business is several years behind TSMC," Citi analyst Chris Danely said in a research note. Danely has previously suggested that Intel should divest its third-party manufacturing business.
Intel has been hampered for years in launching new manufacturing processes. Under former CEO Pat Gelsinger, the company had an ambitious plan to catch up with TSMC and attract outside customers by rolling out "five generations of processes in four years." Each planned “process node” generation represents a new generation of chip manufacturing technology.
However, the results of this plan were far less than expected, and Intel's stock price fell sharply in 2024. New CEO Chen Liwu and his management team are more cautious and have begun formulating Intel's catch-up strategy in the fields of manufacturing and artificial intelligence.
In a conference call Thursday afternoon, management acknowledged that yields on the 18A process—the proportion of good chips that can be produced from a silicon wafer—have not yet reached expected levels.
Intel Chief Financial Officer David Zinsner said that while yield rates were "acceptable," they were "not yet at the levels needed to achieve appropriate margins."
Zinsner also revealed that the 18A process will not reach peak production capacity until "the end of this decade."
Analysts also pointed out that another complicating factor is that the level of market demand for the latest chips manufactured using the 18A process (codenamed "Panther Lake" and "Clearwater Forest" respectively) is still unclear.
Bank of America analyst Vivek Arya reiterated an "underperform" rating on Intel stock on Friday, writing in a report: "Importantly, given the slow internal adoption of the 18A node (peak capacity will not be reached until 2030) and domestic foundry competition in the United States, we do not expect a material improvement in the current unfavorable cost structure of Intel's foundry business."