CES 2025 – Why this processor will (perhaps) save Intel
This is a processor that has a lot of power. About its number of cores or its frequencies, we know nothing – or not much. But this chip that Intel's new co-CEO held in her hand during the company's press conference at CES in Las Vegas is anything but anecdotal. The little piece of silicon that Michelle Johnston Holthaus showed to the press had the final shape of a commercial processor. And materializes the future of the company.
Similar in appearance to the Core Ultra 200HX chips, this processor is much more than a new generation for Intel: it is perhaps its lifeline.
In a context where its CEO Pat Gelsinger has been pushed out and where Intel is dominated (desktop chips), heckled (laptop PC chips) or absent (high-end GPUs and for data centers), the American company is seen by many like a sick animal. Even already buried by some.
The upcoming arrival of Panther Lake is therefore, more than the Core Ultra 200H/HX presented at this CES, a potential turning point for the company and in particular for its IFS division which engraves chips.
An ultra strategic node 18A
On the chip design side, Intel has not yet validated the various rumors: some leaks speak of new high-power cores (P-Core) Cougar Cove and even a 3rd generation Xe graphics part (GPU). . What is certain, however, is that the technological argument of PC Panther Lake is its engraving technology called Intel 18A – the A here referring to the outdated unit of angstroms (1010m). A key process for Intel, with some analysts believing that former CEO Pat Gelsinger literally “bet the future of the company” on its success.
Theoretically equivalent to a 1.8nm from TSMC – but in fact closer to the future 2nm node – the Intel 18A node should allow a drastic reduction in consumption and a notable improvement in performance. Thanks be given to a (theoretical) return of leadership in fine engraving at Intel, made possible by the massive use of EUV engraving.
This engraving finesse is not Intel's only lever. In addition to the pure reduction in the size of the transistors, we must also take into account the design and organization of these transistors. And Intel is taking advantage of 18A to introduce two new features that will benefit Panther Lake: RibbonFET and PowerVia.
Two key technologies
The first is called RibbonFET and concerns the design of transistors. Transistors whose overall organization is turned upside down since the channels are now completely surrounded by the gate. A structure which should make it possible to improve performance through greater component density and a reduction in voltage.
The second technology, called PowerVia, affects the way in which electrical current flows in the chip. Also called “backside power delivery”, this complete reorganization of the circuits – and of the manufacturing method – should allow a natural frequency gain of 6%, a reduction in resistance limiting the voltage drop in circuits by 30% and an improvement in cell utilization of up to 90%.
Between engraving and these two technologies, the 18A manufacturing node of the future Panther Lake chip promises a lot on paper. But the question of the quality of execution – that is to say the implementation of large volume production – remains unresolved.
The success of this execution will have an impact not only on the overall performance of the Panther Lake chip, but also on Intel's wallet. Indeed, if Lunar Lake is a good chip, it is also thanks to the pieces of chips that Intel manufactures – and therefore buys – from TSMC. A purchase that costs Intel dearly and greatly harms the profitability of its chips.
The future of factories linked to the success of 18A
At the instigation of Pat Gelsinger, Intel spent tens of billions of dollars to bring the industrial tool up to standard. A tool made up of factories that he grouped into an entity capable of producing for others: Intel Foundry Services. A tool whose objective until the departure of P. Gelsinger last December 2024, was to regain technological leadership over TSMC as part of the 5N4Y plan. Launched in 2021, this plan provides for the successive launch of five manufacturing nodes in four years. The focal point of this strategy is, you guessed it, the 18A with which Panther Lake is produced.
However, Panther Lake will be the ambassador of 18A: if the performance of Intel's chip is good, this will encourage chip designers to turn, at least in part, to Intel factories in order to diversify their supplies. This could in return allow Intel to increase its volumes, its know-how and its turnover.
In short: if the Intel 18A engraving is effective, the Panther Lake chips will be competitive, which should drive factory business and stabilize or even revive Intel. The importance of the Panther Lake generation therefore goes well beyond the success of the chip itself: it will weigh heavily on the future of the giant.