The First Xbox Handheld. Can Intel afford to lose the console/handheld market?

[hist78]

The other problem Intel faces in console bidding is that most of GPU designs are already designed for TSMC. AMD have been designing their own CPUs and GPUs in TSMC, while only GPUs Intel foundry manufactured were their in-house GPUs. That's part of reason why Samsung 8nm won Nintendo console bidding. New Nintendo console uses NVIDIA Ampere microarchitecture, which was designed for SF8 thanks to RTX 3000 series and NVIDIA Odin SoCs in early 2020s. IP ecosystem matters a lot.
So Intel has to either win business using their in-house designed console SoC (18A + ARC + Intel CPU, in this case Intel "Product" team has to fight against AMD products) or get some design win first(AMD GPU working better in Intel foundry?) in leading edge nodes and wait for process to mature...

Intel's Arc series of GPUs are also manufactured by TSMC. If AMD can succeed using TSMC, why can't Intel, with TSMC, achieve the same?

 

[siliconbruh999]

Intel's Arc series of GPUs are also manufactured by TSMC. If AMD can succeed using TSMC, why can't Intel, with TSMC, achieve the same?

cause Intel has a foundry to run they are planning to get GPUs back to Intel nodes

 

[hist78]

cause Intel has a foundry to run they are planning to get GPUs back to Intel nodes

If Intel moves more chip production back in-house to boost its foundry scale and cut costs, but ends up losing deals and missing revenue opportunities (such as Xbox Handheld), while AMD and Nvidia succeed without owning a fab, then how can Intel remain competitive in the long run?

 

[siliconbruh999]

If Intel moves more chip production back in-house to boost its foundry scale and cut costs, but ends up losing deals and missing revenue opportunities (such as Xbox Handheld), while AMD and Nvidia succeed without owning a fab, then how can Intel remain competitive in the long run?

i think you are wrong here switching to Intel foundry doesn't make them miss opportunity. 18A/AP may not be good for external due to the EDA but they are very good nodes for Internal use

 

[hist78]

i think you are wrong here switching to Intel foundry doesn't make them miss opportunity. 18A/AP may not be good for external due to the EDA but they are very good nodes for Internal use

In the meantime, while waiting for the 18A or 14A, can Intel compete and stop its revenue from spiraling downward?

Intel's competitors and customers won't wait and don't need to wait for Intel.

 

[siliconbruh999]

In the meantime, while waiting for the 18A or 14A, can Intel compete and stop its revenue from spiraling downward?

Intel's competitors and customers won't wait and don't need to wait for Intel.

That's something only time can tell .

 

[semiman]

Intel's Arc series of GPUs are also manufactured by TSMC. If AMD can succeed using TSMC, why can't Intel, with TSMC, achieve the same?

Actually, that's really interesting point. Intel can either bid for SoC(customer: Sony and MS) or foundry(customer: AMD or Intel Products Group) to milk some money from console markets. Either has pros and cons.

1. SoC bidding allows intel to utilize various foundries and IPs
2. SoC bidding has lower products margin compared to CPUs and APs.
3. PS and XBOX software is already optimized to AMD SoCs, so software backward compatibility could be an issue.
4. Using Lunar lake + ARC variants(Intel's best choice for low power products) does not benefit Intel foundry division

1. Foundry bidding has relatively better profitability since it's priced 'per wafer' basis (SoC bidding still gives better margin though)
2. Intel foundry have never manufactured HPC chips(GPU) from external customers yet.
3. AMD(who are behind the mirror) might not want to work with IFS

Either choice has at least one roadblocks for Intel. It's surprising to see how the bussiness has become so complicated...Maybe they have focus on Panther lake first? Diffcult times for Intel.

 

[TICin852]

AMD will never fab on an Intel process. Let's be real and honest here. That's why Intel products has to win over the Playstation/Xbox console teams with their architecture. That is why their GPU offerings at Intel actually will matter - when Sony and MS are looking at the generation later, everyone will look at what architectures are available and go with that. For all handhelds and gaming products right now, Intel products would have to work to make the bid - which makes MJ's statement of "we will only pursue products with margins >50%" very frightening.

foundry(customer: AMD or Intel Products Group

 

[MKWVentures]

This market is not worth it: Not a leader today, No competitive advantage, no strategic advantage... Can't win, don't try.

Regain DC CPU Share and margin, Manage Client CPU share and margins, get 10% of the DC AI GPU market. That will double to stock price.