Xeon Sierra Forrest and Granite Rapids = 1,119,000 units
NEX = 11,151,377 the dice bank is filled
Gaudi 3 = 32,000 courtesy of TPM
Core Arrow desktop = 4,739,608 and another 14,327,127 q4 credit
Core Raptor desktop warranty = 20,301,167
Core Meteor and Raptor embedded = 5,031,377
Core Lunar mobile = 31,206,224 good until Panther
Total = 78,653,881 units

Summary

Intel Stated CPU Division Revenue = $12.1 B / 78,653,881 units = $153.84
Intel Stated CPU Operating Cost = $8.8 B / units = $111.88
Intel States CPU Operating Income $6 B / units = $76.28 that is Price ‘A1’

CPU Divisions $12,100,000,000 revenue / units = $153.845 per unit
CPU described as Operating Cost = $111.88 per unit
CPU Gross is actually fixed (marginal) cost of production = $41.96
CPU R&D contribution = $51.48
CPU MGA contribution = $17.48
CPU NET revenue less costs = $1 same as q4 2023

CPU PE&C contribution = $165.67
CPU Restructure contribution = $71.47
CPU Tax accrual contribution = $100
CPU NET LOSS = ($337) per unit that is $5,622,000,000 restructuring

Fixed cost of production $41.96
R&D as fixed cost $51.48
MG&A as variable cost = $17.58
Total $111.02

Less restructuring $71.47 is Price ‘A2’
Less fixed cost of Production $41.96
Equals variable cost $39.54

The full report including q3 x86 production and component channel share is here top of comment string here,

https://seekingalpha.com/article/4731858-intel-corporation-intc-q3-2024-earnings-call-transcript

Mike Bruzzone, Camp Marketing

Good questions. I think antitrust bodies have the power to prevent mergers and acquisitions, but I don’t think any government has any power to do any busting up except inside its sovereign domain. I am counting Europe as a domain in this regard, given the EU.

This is the problem, right? TSMC is essentially a monopoly at this point. It’s also in a politically dangerous location, but even without that it’s a monopoly. If Intel fails, does the department of justice try to split up tsmc? It’s not a US corporation. How do antitrust laws work when none of the remaining vendors are US based?

Agreed. We need a second, good foundry. This is dangerous. Of course, having a second good foundry is ALSO dangerous.

Which begs the question: Is the world always dangerous? HA! It seems to be built into the fabric of spacetime. . . . probably on purpose.

It’s probably okay (in my mind, not unusual) as the DJIA is meant to depict overall market performance in the US based on just 30 stocks, and so they add and remove stocks from companies in similar activity categories every year (to keep just 30). Replacing Intel (underperforming) with Nvidia (overperforming) is going to give DJIA a bit of an “optimistic boost” though, that may not make it a more accurate representation of real industrial activity (imho).

What Intel is doing with (essentially) “reshoring” of advanced chip production fab capacity is very expensive, but was deemed strategically essential in view of the supply chain disruptions that arose from COVID. I expect that, through this investment, short-term profit loss will be replaced by steady gains in the medium and long-term, once fabs are operating at capacity (it just takes a bit longer than what some financial analysts are used to).

While Intel is trying to regain market share, my opinion is they sold off too many innovative assets and discontinued others. I think recovery will be difficult because the biggest problem is not bad CPUs but disruption by the AI industry’s need for faster storage, faster networking and faster GPU accelerators.

Now that they have the HPCG bullriding suplex champion Granite Rapids in the field (and Diamond as a presumably straightforward PCIe6.0/CXL 3.0 successor), it’s high noon indeed for a focus on Falcon Shores, and related media PR, tech info, details, specs, size of saddlebags, etc … The last plot in Wednesday’s AMD Instinct story says it all: a 30.4x revenue increase in 1 year! That’s the kind of silver dollar that could really help pay for foundry bringup at advanced nodes IMHO.