Intel CEO Pat Gelsinger has described a bold plan to spend up to €80 billion on the European semiconductor ecosystem over the next decade but there was one key ingredient missing – the money. Or rather the source of the hefty subsidy that Intel needs to push ahead.
With particular reference to a plan to spend €17 billion building two fabs at Magdeburg in Saxony-Anhalt, Germany, Gelsinger said: “Still there’s still work ahead to obtain … the financial support to make the project competitive.”
The reference to financial support was buried in the back half of a sentence about construction and other permits, but it is key. It tells those that are prepared to listen that some mix of the European Union and the German national and regional governments must come up with their considerable end of the two-fab deal, or it could be scaled back, stretched out in time, or even fail.
Nonetheless, Magdeburg is a winner, for now, in the visionary statement from Intel’s CEO, which I give high marks to for its ambition.
Not the only winner
And Magdeburg is not the only winner. Intel plans to spend €12 billion at Leixlip near Dublin bringing up the Intel-4 manufacturing process. It also plans to create an R&D and design hub at Saclay in France and to spend €4.5 billion on a back-end facility somewhere in Italy. However, the Italian chiplet assembly and packaging facility is not due to start operations until between 2025 and 2027 (see Italy, Poland woo Intel for ‘chiplet’ packaging plant).
In the nearer term Intel’s planned acquisition of Tower Semiconductor (see Intel set to buy foundry Tower for $5.4 billion) would bring Intel into partnership with STMicroelectronics (see ST, Tower to share 300mm Agrate fab). And Intel and ST have a healthy history of cooperation.
Poland and Spain are also set to receive investment as part of a European-wide front that aims to support the European Union’s goal of grabbing 20 percent of global chip production capacity by 2030. How realistic that goal is, is a different discussion. But supporters would probably argue that it is a “stretch-target,” one that is less about the end-point and more about the direction of travel.
Moving the needle
Certainly, if Intel brings up these multiple capabilities it could nudge the needle on European production market share upwards – or at least stop it continuing to drop.
European Commission president Ursula von der Leyen said Intel’s welcome support would also help Europe break through the 3nm-node barrier. Maybe: the Intel-4 process was discussed as being part of an offering from Leixlip. But it remains unclear how successful Intel will be at progressing its road-map to 3nm and beyond. One thing is clear is that the European Union wants to build up both volume and leading-edge capability in its pursuit of strategic supply chain resilience.
But it is also notable that Intel has put a generous timescale on many of these plans. While planning for Magdeburg will start immediately, construction is not expected to begin before 2023 or 2024, with production planned to come online in 2027. Magdeburg is expected to deliver chips using Intel’s most advanced, Angstrom-era transistor technologies, serving the needs of both foundry customers and Intel itself. But Intel states that this plan is “pending European Commission approval.”
Well the European Commission clearly does approve of Intel’s plan in principle, so to me this missing approval can only be about the financial arrangements. I suggest there is still a gap to be closed between how much subsidy Intel expects (see Intel seeks $10 billion subsidy for ‘EuroFab’) and how much subsidy Europe thinks it will pay.
Time to find the cash
The lead time gives all parties the chance to get the money in place for, after all, where there is a will there is a way. As Gelsinger said: “We [Intel] are committed to playing an essential role in shaping Europe’s digital future for decades to come.”
And as von der Leyen said: “This is not just about a few big fabs.” She added: “This is a first major achievement under the European Chips Act covering R&D, manufacturing and packaging. It will pave the way for more companies to follow suit.”
But with pressures coming on all sides, not least from inflation and a potential recession, one should not let the protagonists’ enthusiasm for the plan obscure the issue of finance.
The long lead time to groundbreaking and big spending in Magdeburg would suggest that Intel and the European Commission are using the publication of the plan as a way to try and round up the money.
Intel plans to spend more than €33 billion on Magdeburg, Leixlip and in Italy but it expects a lot of that money to come from European taxpayers rather than from its shareholders. That means there is a lot of accounting yet to come.
Related links and articles: