Can Intel still hope for a fresh start? Or is he already condemned? Monday evening, the semiconductor giant suddenly pushed its leader, Patrick Gelsinger, into retirement. He leaves behind a group struggling in a thriving industry, driven by the unprecedented success of Nvidia.
Now, the tech giant faces a dilemma. Separate from his foundry, responsible for his financial difficulties and the dissatisfaction of his investors. Or persevere in the high-risk strategy of his former boss, betting on his longer-term success and the support of theANDtat.
Patrick Gelsinger’s crazy bet
Arriving in February 2021 at the head of Intel in a context of crisis, Gelsinger focused on manufacturing activity – that is to say the semiconductor production and assembly chains – to restore the image of the historic player in the sector. There was urgency: its main competitor, AMD, was already starting to eat away at its historic business of selling processors – the famous CPUs, the performance heart of computers and servers – in addition to spilling over into other types of chips.
The emblematic boss of Intel therefore launched an ambitious, and very expensive, plan to make Intel’s foundries the best in the world, and thus seriously compete with the Taiwanese TSMC and the Korean Samsung. But he will not see the outcome of this strategy… now on hold.
If Patrick Gelsinger jumps, it is above all because of market distrust. The company’s stock is down 53% since the start of the year. The manager therefore suffered the same fate as his predecessor Bob Swan, also abruptly dismissed at the start of 2021, under pressure from activist investors. The new retiree’s task was not easy. Some would even say lost in advance. And the equation will not be any easier for his successor.
« All of Intel’s competitors in the processor market, whether Nvidia, AMD, Qualcomm and Broadcom have abandoned chip manufacturing to concentrate on research into new architectures. Intel therefore finds itself isolated: it is the only American company that has retained all vertical integration, that is to say that it creates both chip architectures and their manufacturing. Except that she fell behind on both sides », analyse Xiadong Bao, cmanager of the Big Data fund at Edmond de Rothschild Asset Management.
A point of view shared by Alvin Nguyen, principal analyst at Forrester. “ Intel dominated the market for many years, because they had the best architectures and the best foundries. But they no longer have either », he adds.
How Nvidia became the essential barometer of the artificial intelligence boom
Patrick Gelsinger could have tried to follow the path opened by AMD, which sold its foundry activity in 2008. But he preferred to take the opposite approach to his competitors. According to him, it was necessary to open new factories and strengthen the foundry activity. Because today, AMD, Nvidia and even Apple entrust the chip manufacturing work to TSMC and Samsung. Rather than trying to catch up in chip architectures, Gelsinger figured he could make Intel the preferred manufacturer for these companies.
« Intel purchased a lot of equipment and machines from specialists like ASML, to the point of blocking orders from its competitors TSMC and Samsung for a year. The company is trying to build the world’s best foundry in Oregon », says Alvin Nguyen.
A gamble too far away for investors
Over the last three years, Intel has announced no less than $100 billion in manufacturing investments, according to the media’s count The Register. In the United States, but also in Europe and the Middle East. The new factories would feature new manufacturing processes, named Intel 18A and 20A. Problem: this bet takes time, with a delivery objective no earlier than the end of 2026. That is, five years, an eternity on the scale of a tech industry which is benefiting from a historic craze linked to generative artificial intelligence. .
« This investment is based on a much longer-term return on investment schedule than investments in chip architecture. It is only from 2027, or even 2028, that Intel can hope to achieve the potential competitive advantage it is aiming for. Above all, there is no guarantee that this bet will work. Samsung had made a similar attempt, without success. The investment then becomes a loss », Develops the Forrester analyst. The expert also says he is convinced by the strategy, and deplores the lack of patience of investors. “SIf Intel focused on the short term, it would have no chance of catching up. But the market is no longer accustomed to such risky approaches. »
Colossal failures and losses
The problem for Gelsinger is that while his vision slowly comes together, the company seems to be collapsing in on itself. In its historic market, the sale of CPUs, AMD continues to gain ground, whether in desktop computing, servers, or laptops.
And even when Intel tries, it stumbles. At the start of 2023, the Sapphire Rapids processors, which were already almost two years late for their launch, encountered such significant bugs that the company was forced to briefly stop deliveries… Commercial failure also for the Gaudi3 processors, supposed to compete with those of AMD and Nvidia, to run artificial intelligence and thus allow Intel to benefit from the technological boom.
Worse, while waiting for the inauguration of the factories of the future, Intel’s current foundries are so far behind in technology that the company subcontracts the manufacturing of its latest processors… to TSMC, the same competitor it is trying to overturn. In other words, Intel is investing massively in an activity whose order book it is emptying.
As a result, at the beginning of 2024, Gelsinger made the decision to officially separate the “Foundry” activity from the rest of the group’s financial results. A way to put aside the division’s 7 billion losses in 2023 – mainly linked to colossal investments – and to show that the rest of the activities are not doing so badly. In fact, the group still posted a profit of $1.7 billion (however in free fall) over the year. But since this decision, Intel Foundry has announced no less than $11 billion in additional losses throughout 2024.
Stampede
Obviously, the share price collapsed, and to make the situation worse, investors tried to take the group to court. They accuse executives of having underrepresented the financial difficulties of the Foundry division. Even if the situation is ultimately hardly surprising, with the investment policy.
Semiconductors: how Intel buried the European “Chips Act”
To compensate for the losses, Gelsinger announced in August a gigantic layoff plan: 15,000 employees, or 15% of the company’s total workforce, will be laid off by the end of the year. As for capital expenditures, they will be cut by 20%.
The leader’s grand plan is then slashed from all sides. The Magdeburg factory project in Germany, costing 30 billion euros – including 10 billion in public funding – is put on hold, before possible abandonment. Likewise for the $4.6 billion assembly plant in Wroclaw, southern Poland. It remains to be seen how the European Union, supposed to financially support these projects within the framework of the Chips Act, will react.
Towards a split of activities?
With the upcoming change of direction, Intel could opt for the big leap: letting go of its foundry, and Gelsinger’s big bet, like AMD before him. The company would thus get rid of its bottomless pit of expenses, and reassure its investors.
But the decision is not entirely in the hands of the group. First, it is difficult to imagine the American government letting this happen without taking a stand. Intel remains the only company capable of producing cutting-edge processors on its territory. Its survival therefore becomes a national security issue, since it guarantees strategic independence in an industry that is essential to all technology.
Furthermore, as part of the Chips Act put in place by the Biden administration – the money for which is urgently distributed before the arrival of the Trump government – Intel was allocated $7.86 billion, as well as an additional $3 billion for a project dedicated to the Department of Defense.
But to touch the envelope, Intel must maintain more than 50% control over its manufacturing division. Even if he decides to make it a separate business. “ It’s almost impossible to project Intel’s future today. There are a lot of questions about his strategy, and itis a very politically sensitive company, subject in particular to the evolution of relations between the United States and China », conclut Xiadong Bao.
