Intel's latest high volume EUV chip production could help it retake the lead in the semiconductor...

Jimmy2x

Posts: 239   +29
Staff
Forward-looking: The explosive growth of artificial intelligence (AI) has shifted the semiconductor technology race into overdrive. Yesterday, Intel announced the arrival of its extreme ultraviolet (EUV) Intel 4 technology at Fab 34 in Ireland. The highly anticipated milestone leaves Intel positioned to deliver on its promise of delivering five new process nodes over the next four years.

The heavy computing requirements of AI push evolving manufacturing processes that shrink overall chip sizes while allowing a higher concentration of compute elements on each chip. As a result, companies like Intel, TSMC, and Samsung are constantly investing in facilities, equipment, and resources to stay one step ahead of the competition in the semiconductor market. On Friday, Intel announced the arrival of its Intel 4 manufacturing process, a vital component of the company's upcoming Intel Core Ultra processor lineup, at its Fab 34 manufacturing facility in Leixlip, Ireland.

Intel 4, previously known as the 7nm process node, will reportedly provide upwards of a 20-percent increase in clock speed while significantly reducing power consumption. It brings an overall increase in yield due to the decreased die size. The new manufacturing process improves Intel's current Intel 7 (10nm++) process node currently powering the 12th gen Alder Lake family of CPUs and the 13th and upcoming 14th gen Raptor Lake processors.

The new manufacturing process relies on the latest extreme ultraviolet (EUV) lithography systems from ASML. The EUV systems are roughly the size of a small shipping container and support high-volume production of nodes as small as 7nm to 5nm. Its latest TWINSCAN NXE:3400C can produce upwards of 170 wafers per hour.

Intel's transition to the Intel 4 manufacturing process is a huge step forward on the company's four-year roadmap. If it can stay on pace, the company will move to the Intel 3, 20A, and 18A processes in the coming years.

The new capability will undoubtedly help Intel gain ground in the already competitive, sometimes contentious, semiconductor race. Earlier this year, Intel argued that US companies should be eligible for higher CHIPS Act subsidies than those owned by foreign companies such as TSMC. The Taiwanese chip giant asserted that allocating grants based on a company's registered headquarters location was an inefficient and inaccurate distribution method. Both manufacturers were later hushed by Commerce Secretary Gina Raimondo, stating that the program's focus was strengthening national security and not to boost struggling chipmakers.

Permalink to story.

 
I stopped believing Intel in 2017. Haven't restarted. Sad, for a flagship US enterprise that had always had my highest respect and is at the core of my interests.

Give us an email notification box for "When Intel can again be trusted".
 
How will this technology survive a Carrington Event?

I have a maybe 5 to 7 TWINSCAN NXE:3400C- can't remember have many stored in my shed - since I'm on the other side of the world - leasts some of us can take some protection

National Oceanic and Atmospheric Administration - has a 3 day protection project

Anyway of greater concern would be our grid - and mission critical devices

Think Military make devices that are a bit more harden

Maybe China would like to buy those Twinscans I have laying around - gotta be worth something
 
Here's a crazy idea. How about we wait and see what kind of products they churn out with this new "breakthrough" they're claiming? I have a feeling it's going to be another 5-15% improvement over current chips and no doubt Intel will pay these sites handsomely to celebrate it as some kind of miracle breakthrough that puts them years and years ahead of the competition.
 
Logically, incompetent companies should receive fewer subsidies.

Intel 4 is still inferior to TSMC 5nm.
 
Transistor density in MTr/mm2
TSMC 5nm: 138
TSMC 4nm: 143
Intel 4: 123

SRAM bit-cell size in μm2
TSMC 5nm: 0.021
Intel 4: 0.024
there are other parameters to performance of a given node than quoted density (leakage, frequency, current). For example the ccd of the 5nm ryzen cpu achieves "only" ~100Mtr/mm2 of the quoted 138Mtr/mm2. If intel optimizes their node for CPUs maybe the cpu on the intel 4 will have better density and performance, but we need to see a cpu on the new node in order to make some real world comparisons. on wikichip they say that "Based on most of the recent publically available foundry data, Intel 4 HP cells appear denser than TSMC N5 HP and are likely closer to or better than TSMC N3 HP cells and denser than Samsung’s 3GAE"
 
Last edited:
Back