Considered moribund just a year ago, the American semiconductor giant is now at its all-time high on the stock market and has a string of strategic partnerships. However, the recovery remains fragile.
Just a year ago, Intel seemed on the brink of collapse. The CEO chosen to launch an ambitious transformation strategy, Pat Gelsinger, has just been ousted by the board of directors. Trailing behind Nvidia and AMD on AI chips, its new 18A technique for manufacturing wafers and its Ponte Vecchio GPU struggling to convince, the American semiconductor giant had a series of loss-making quarters and layoff plans, to the point that its very future seemed to be called into question.
We were starting to talk about a split sale of the company, with the foundry activities on one side and the design part on the other, to move on to a fabless model, like Nvidia and AMD, a solution notably advocated by former executives in a tribune co-signed in the American magazine Fortune. Appointed in a hurry to replace Pat Gelsinger, Lip-Bu Tan resembled the captain of a drifting skiff, navigating by sight in the middle of a storm.
Intel is back on track
A year later, Intel has transformed itself. The company returned to growth: its revenue rose more than 7% year over year in the first quarter, reaching $13.6 billion and beating Wall Street forecasts by more than $1 billion. It could approach 15 billion in the second quarter. If its turnover must still increase significantly to return to its historic peak of 2021 ($79 billion), the company’s action is currently at its highest in its history. It is trading at the time of writing for $130. It was barely $20 a year ago…
Growth driven in particular by the rise of inferencewhich, unlike training, is more demanding in CPUs (historical specialty of the company) than in GPUs, and leads the AI giants to increase their orders for chips from Intel. An inference market itself driven in particular by the rise in power of the inference market agentic AIas companies strive to deploy AI for concrete use cases.
Strong partners
Intel is also on the verge of closing its technological gap over its rivals Nvidia and AMD. Its future 14a engraving process seems to arouse the confidence of its customers, like Elon Musk, who chose Intel for its future Terafabannounced in March. Expected for 2027, the 14a process would be the first in the world to use High-NA EUV (High Numerical Aperture Extreme Ultraviolet) lithography, a technology provided by ASML, which allows finer engravings. Even the undisputed leader in AI chips, TSMC, isn’t using it yet. Thanks to this, Intel aims to pass the nanometer threshold in terms of transistor density (TSMC currently reaches two nanometers, which is already a prodigious performance).
Of course, all this is hypothetical for the moment, and Intel has already seen its technological ambitions facing difficulties scaling up to industrial scale in the past. However, Intel’s vision has convinced a new major partner: Apple. The two companies have just signed a preliminary agreement under which Intel will manufacture chips intended for “certain devices” of the Apple brand. A return to basics for Apple: Intel actually manufactured the chips for its Macs from 2006 to 2020, when the Apple group moved towards its in-house M1 chips, based on an ARM architecture and melted by TSMC.
This time, however, Apple intends to keep its in-house design, and simply buy production lines in Intel’s production factories, in order to no longer be solely dependent on TSMC. The apple group thus intends to secure its supplies while the Taiwanese giant overwhelmed by demand and struggles to produce enough. A bottleneck which, by the company’s own admission, weighs on Apple’s growth.
The rise in geopolitical tensions also poses a significant risk, as TSMC’s production is still largely concentrated in Taiwan, despite a diversification strategy implemented. in Japanin Europe And in the United States. If it still has to convince on its ability to compete with TSMC, Intel remains a powerful industrial company, the only one capable of promising similar volumes, and whose production is located in the United States and boosted by the CHIPS Act seems very attractive in a dangerous and uncertain international context.
Pat Gelsinger’s winning bet
Intel’s turnaround is partly the result of the strategy of its new CEO, Lip-Bu Tan. He operates as an excellent strategist and seller of the Intel brand, according to Ian Cutress, CEO of More than Moore, an analyst firm specializing in semiconductors. “Lip-Bu Tan talks the right talk and makes deals that reinforce Intel’s long-term vision. All of this reinforces the belief that the group’s foundry strategy could rival TSMC, both in manufacturing and packaging.”
He also knew how to maneuver intelligently with Donald Trump and obtain an investment of 10% of Intel’s capital from the American government. The Trump administration, which shares the objective set by the Biden administration to produce more chips in the United States, would also have maneuvered to facilitate the new agreement between Apple and Intel, according to information reported by the American media.
Donald Trump would have praised the merits of Intel to Tim Cook during an exchange between the two men at the White House, while Commerce Secretary Howard Lutnick would have met several times over the past year with senior Apple officials, including Tim Cook, as well as Elon Musk and Jensen Huang, of Nvidia, in order to try to convince them to do business with Intel, with success since the three players have since signed agreements with the American semiconductor group.
Lip-Bu Tan also trimmed headcount, empowered engineers over managers, and polished operations to increase the company’s efficiency and profitability.
But Intel’s recent successes are also the legacy of the strategy initiated by his predecessor, Pat Gelsinger, aimed at making Intel a major player in semiconductor production, serving states and companies wishing to relocate part of production. A strategy that is long-term, and therefore takes time to bear fruit, according to Mike Demler, an independent expert specializing in the chip industry. “Intel is a large company in a major transition phase, which takes years. Intel is also working to become the only American company capable of producing advanced semiconductors in the United States, which greatly increases the difficulty and cost of its task.”
A window of opportunity soon to close
Intel is currently benefiting from a combination of favorable elements to, if not catch up with TSMC, at least progress sufficiently to carve out a place for itself within the closed circle of AI chip founders. Subsidies from the CHIPS Act and money injected by the US government provide sufficient funds, while tensions between China and Taiwan and the memory of the semiconductor shortage during Covid are sufficiently vivid to determine AI players to diversify their supplies. Added to this is the bipartisan agreement around the fact that the AI value chain constitutes a major sovereignty issue, which guarantees support for the production of chips on American territory, regardless of the political party in power.
However, the fact that TSMC, AMD and Samsung are also striving to strengthen their production in the United States also means that the firing window is limited and that Intel will not be able to play the sovereignty card indefinitely. It must therefore quickly catch up technologically to establish its ecosystem around AI chips and remain competitive against its rivals.
“You need concrete evidence of purchased capacity – and even for Apple or Tesla, nothing like that has been confirmed yet. We hear a lot of things, and sometimes that’s enough to keep the market excited. The key factor for Intel will be whether its own products work and scale,” Ian Cutress said.
The success or otherwise of the 14a engraving process expected next year will constitute a life-size test for the company’s ambitions in this regard. “Everyone is looking at TSMC’s margins and waiting for Intel Foundry to reach similar levels. to TSMC’s SoIC and CoWoS) and progress on these technologies as well as their industrial rise over the coming years will be decisive.” The American giant is therefore not yet out of the woods, but its situation is undoubtedly much more enviable than a year ago.
