Intel discloses CPU roadmap, confirms Nova Lake architecture for 2026 rollout

zohaibahd

Posts: 934   +19
Staff
In a nutshell: Intel has pulled back the curtain on its plans for its next-generation PC processors, codenamed "Nova Lake," set to launch in 2026. The revelations came during the chipmaker's quarterly earnings call, where it delivered a roadmap that will guide Intel's client computing business over the next couple of years.

Taking center stage was the confirmation of "Nova Lake" – Intel's next major client CPU architecture set to arrive in 2026. This will succeed the company's recently launched Arrow Lake desktop processors and could include laptop variants.

The news was delivered by Michelle Johnston Holthaus and David Zinsner, the chipmaker's newly appointed co-CEOs, who replaced Pat Gelsinger following his departure last year.

Before Nova Lake makes its debut, Intel has another big release on the horizon: Panther Lake, the company's next processor family, scheduled to launch in the latter half of 2025. Holthaus revealed that while Panther Lake will be largely manufactured in Intel's own fabs, Nova Lake will utilize a mix of internal and external production sources, possibly including TSMC.

"2026 is even more exciting from a client perspective as Panther Lake achieves meaningful volumes and we introduce our next generation client family, code name Nova Lake," Holthaus stated. "Both will provide strong performance across the entire PC stack with significantly better costs and margins for us."

Intel hopes the upcoming chips will bolster its competitiveness against rivals like AMD and Qualcomm amid rising competition. Holthaus acknowledged that profit margins on the company's current Lunar Lake parts took a hit due to integrated memory costs.

Details around Nova Lake remain scarce, with the co-CEOs not revealing too much. It is rumored that the processor will be fabricated on Intel's 14A node or TSMC's 2nm process. TSMC's 2nm technology employs gate-all-around transistor architecture like Intel's 18A and promises significant improvements in performance and power efficiency. Mass production is slated for late 2025.

On the manufacturing end, Intel recently spun off its foundry operations into an independent subsidiary with its own board. As reported by PCWorld, Holthaus made it clear that the new entity will be treated like any other supplier, having to earn Intel's business.

"A stronger Intel Product combined with a more competitive Intel Foundry is a recipe for success," she added.

However, Holthaus tempered expectations, noting that gaining customer trust would be a "slow ramp" initially. "We won't win every deal out of the gates. We'll be selective and focused on areas where we are confident that we can be a meaningful contributor."

Funding from the US CHIPS Act has provided a boost, with Intel receiving $1.1 billion in Q4 2024 and another $1.1 billion in January as part of a total $7.86 billion grant allocation. Company executives also confirmed an ongoing engagement with the administration regarding the initiative.

Permalink to story:

 
However, Holthaus tempered expectations, noting that gaining customer trust would be a "slow ramp" initially. "We won't win every deal out of the gates. We'll be selective and focused on areas where we are confident that we can be a meaningful contributor."

Who doesn't like transparency? I like the company vibes already.
 
With Zen 6 rumoured to have 32 cores per CCD, new interconnect tech and upgraded I/O die, all on the same socket, Intel needs a miracle.
 
With Zen 6 rumoured to have 32 cores per CCD, new interconnect tech and upgraded I/O die, all on the same socket, Intel needs a miracle.

It's so ironic how Intel used to be king just 5 years ago.

On a less serious note, still stuck on the lake naming schemes...wow...I like lakes but can't they mix it up just a little bit more...how many lakes doth this make now?
 
It's so ironic how Intel used to be king just 5 years ago.

On a less serious note, still stuck on the lake naming schemes...wow...I like lakes but can't they mix it up just a little bit more...how many lakes doth this make now?
You could say that they reached the bottom of the lake in terms of naming :)
 
It's so ironic how Intel used to be king just 5 years ago.

AMD was king in 2005... then Intel launched Core 2 Duo in 2006 and AMD didn't respond with anything competitive until Ryzen launched.

We need Intel to be competitive - AMD are already taking advantage of that with increased pricing, which to be fair is what I'd do in their position. But we need both to be relatively competitive to help keep pricing from getting too high.

Intel needs a miracle.

See above, AMD needed one for 11 years before getting one with Ryzen. Don't rule out Intel coming back.
 
AMD was king in 2005... then Intel launched Core 2 Duo in 2006 and AMD didn't respond with anything competitive until Ryzen launched.

We need Intel to be competitive - AMD are already taking advantage of that with increased pricing, which to be fair is what I'd do in their position. But we need both to be relatively competitive to help keep pricing from getting too high.



See above, AMD needed one for 11 years before getting one with Ryzen. Don't rule out Intel coming back.
Problem is: AMD ditched several architectures between 1999 and 2015. After K8 design complete (1999), AMDs next architecture was Bulldozer (design complete 2009), so at least two or probably three major architectures were tried and ditched because they were too hot, slow or both. Even Bulldozer was rushed out because AMD needed Something new. Between was K10 that was only tweaked K8 so that doesn't count. AMD also was supposed to do "fixed Bulldozer" but as we know, AMD went all in with Zen. In other words AMD had 16 year gap between major architectures so Intel had easy time.

After Zen launch, AMD has released major improvements steadily. While Zen was indeed clean sheet design, Intel keeps adding more on Pentium Pro design year after year. No, Sandy Bridge is NOT totally new architecture despite what some say. It was tweaked and reconstructed Nehalem but shares way too many similarities to be considered totally new architecture. That is reason why Intel chips run so hot. Adding more and more on existing architecture means more and more bloat that causes big power consumption on high clock speeds.

Intel needs miracle because only way Intel can keep up with AMD is totally new architecture. And as we have seen from AMD, totally new and someway revolutionary architectures are big risk. Jim Keller shared some details from one before he left Intel. Haven't heard anything about that since (around 5 years ago) so that architecture was probably ditched.
 
AMD doesn’t really need to do much to keep the throne. Shrink the die and up the cache - that would result in higher frequencies and better gaming performance - which I guess is what the plan is when they move to 2nm
 
AMD was king in 2005... then Intel launched Core 2 Duo in 2006 and AMD didn't respond with anything competitive until Ryzen launched.

We need Intel to be competitive - AMD are already taking advantage of that with increased pricing, which to be fair is what I'd do in their position. But we need both to be relatively competitive to help keep pricing from getting too high.



See above, AMD needed one for 11 years before getting one with Ryzen. Don't rule out Intel coming back.
Indeed, Intel still has enough money to pay OEMs to stay in the game even if they have bad products. Like they did multiple times in the past.
 
Back