Intel totals automotive group Intel is shuttering its automotive efforts and laying off the bulk of the team responsible. The chip giant has been building hardware for the car sector for nearly 50 years, but some staff are falling prey to the new CEO's cost-cutting program. The news, previously reported by The Oregonian, fits into new CEO Lip-Bu Tan's …
This is stupid. The automotive sales are perhaps not huge, but they are always there, even during the pandemic, when the automakers slashed their orders then went "ohshit, we needed that"
This'll give "the next Intel" a foothold to grow.
This won't be apparent in the next 5 years, but I'm sure in the next 10 it'll be obvious. And I'll bet you a dealer lot full of cars we'll have forgotten about it and be going "however did Intel get in this situation?" after it does another dozen or so similar stupid decisions.
No no, see, a huge and ever-growing number of people will want incrementally-faster x86 architecture processors for decades to come, just as long as Sales and Marketing Group Accenture AI can come up with new and exciting variants on the logo and the brand naming schema. Core Ultra? Core Extreme? Core Plus Ultra Pro i11?
Apple? ARM? RISC-V? Pfffft. Who'd want them when they could have Intel®????
No-one ever got fired for buying IBM Intel, right?
---> pint for all the ex-Intel folks - who really, really deserved to be treated better than this.
If Intel - and Foundry ….
- can’t make good money on the Automotive sector which is becoming more and more tech oriented and still just catching up on post-Covid supply chain issues
- are so bereft of marketing nous and ability they need to outsource it to Accenture (of all people) …
… they may as well just liquidate the business now whilst it still has some value and be done with it…. before they join the ranks of Polaroid, Kodak, Woolworths and Pan-Am.
Perhaps someone with some insight into chip manufacturing can help me out here?
Is it the case that automotive chips (or at least those of the kind Intel is currently making) can be made adequately with simpler processes? If so, I can imagine that manufacturers who are not investing heavily in cutting-edge technology might well be able to produce them more cheaply. Though, as the parent poster points out, the trend for automotive technology seems to be towards significantly greater complexity, perhaps the view is that the cost pressures mean it will always lag behind in terms of RoI.
The automotive world is different. You can split the landscape into several sections : sensors , engine / drive train control , safety , body control and infotainment.
Intel doesn;t play in the sensor market ( they got rid of mobileye). Engine / drive train control is the domain of microcontrollers. TI, NXP and ST-Micro rule that world. Intel doesn't do microcontrollers anymore. They had a lot of offerings in the 8051, 8096 families in the 80's and 90's but systems no are invariably Arm Cortex based.
Safety (stability, antilock brakes etc) requires ASIL level D certified stuff. Those are highly specialized. The ARC processor (for ASICs) , NXP S32 (ARM cortex) family , TI Jacinto (ARM Cortex) and Freescale Qorivva (PowerPC architecture). They feature multiple cores that run in lock-step. The cores are sometimes different architectures (one 64 bit, one 16 bit) to prevent software duplication and error path duplication.
Body control requires processors with lots of IO (analog and digital, PWM, timers and BLDC bridge drivers) , connectivity (CAN, LIN , Flexray, Automotive ethernet ) and lots of on-board ram and flash. Intel is not a player there. Again, the domain of ST, NXP, TI, Infineon. Intel had an offering but i don't even know if anyone uses it.
The biggest problem for all of the above is the automotive temperature range. These device need to operate from -55 to +125 (AEC-Q) or even +155c. This requires additional design effort but also additional testing effort during manufacturing. It's a cut-throat market where devices cost 1 to 25 $ .. not a pond Intel and AMD want to fish in.
So, what's left ? infotainment. No highest-level AEC-Q requirement, 0 to 85c suffices, not safety. Low power desirable, embedded graphics desirable since we are all moving to touch-screens, and the ability to run commodity OS (typically a linux kernel based system. This is contrary to the other domains where they either run bare-metal or use the likes of VxWorks , QNX or other RTOS).
Intel has done well using their mobile processors + ARc graphics , but they are being overtaken by AMD. AMD now owns Xilinx FPGA technology and can offer highly customizable platforms like the Zync family. The Zync processors are even used in aerospace as flight, engine and FTS controllers. AMD is branching out into ADAS and AI. the Ryzen for infotainment and Versal for AI
It's a weird shift.
Yes Intel flogged - for a pittance - their StrongARM/XScale business in 2006 …. the year before the iPhone launched.
Their ARM business was sector leaders powering Compaq iPaq’s, Dell Axims, Creative Xen, many microcontrollers, Kindle etc ….
They threw it away on the bollocks that was Intel Atom.i think Intel still hold an ARM licence…..
Yeah, Total Crack Craziness.
My understanding of car tech is that is it about the level of the pre-internet 70's / 80s - 1 "mainframe" cpu, with some subsidary processors and sensors, wired together with a metal string to send elctricity between some so that they are "networked".
Little or no security, no integration architecture within the car, no data management, no service platform, no architecture for connecting to the wider world of external networks.
"Intel inside" should have been a gateway to creating and selling more advanced tech.
OTOH Intel's history with Technology stacks is pretty poor - HDMI, Thunderbolt, Storage, Compilers.
And how is it Tesla was the first to figure out OTA updates ?
There's this idea -- created by marketers? -- that the marketing department does not need to understand the product to successfully market it.
"The Intel i786™ will make your breath fresher, your teeth whiter, your hair fuller, and your penis larger!" (Cut to slow L/R pan across the Budwiser Bikini Babes Intel Insiders wearing nothing but strategically-placed blue body paint in the shape of the circle-Intel logo. SFX: Do-do-do-DO!)
- Invent something everyone wants.
- Grow big on the proceeds.
- Grow even bigger on the fame and proceeds.
- Start ignoring the opposition as they are "not important".
- Start ignoring what your customer wants (like exponentially better products at a cheaper price).
- Ignore all the minnows of competition running under your feet & past your legs.
- Begin to realise all these minnows have produced better products at cheaper prices.
- Start cutting back, so that it becomes impossible to compete.
"As we have said previously, we are refocusing on our core client and data center portfolio to strengthen our product offerings and meet the needs of our customers," Intel said in a statement to The Register on Wednesday.
This is very close to what I and others suggested that should be doing more than a year ago.
It was pretty plain the see they had expanded far beyond processors and were getting into all maner of other markets while the money was flowing with no real aim for what any of it would do beyond expand portfolio.
It was a whatever then, sure expand. But the moment Ryzen showed and ate their lunch, they shrugged instead of doing anything meaningful. Now they are in trouble and suddenly doing the things hat were obvious to the outsider.
I am sure the person in the position to make these decisions has a lot more to deal woth too, but if this is what the Clevels get headlines for, I could have told them their business plan from bed, half asleep.
All this to say yeah it's silly I am annoyed with a multi billion dollar company that does not even know I exist for having the foresight of a toddler.
Accenture has 3.5 lahk employees in India out of 8 lakh total, and they will add another 0.5 lahk this year, in India [Accenture to promote 43,000 employees in India in FY25, Times of India, May 22, 2025]. (1 lahk = 100K) I expect Accenture has got wads of money to subsidize the AI-zation of their India workforce, so they are currently offering very cheap prices to Intel, compared to what Intel would have pay for marketing employees in the US where both the employee and Intel would be paying into the social security system.
However, - guess what? - just because it is cheaper doesn't mean that it is better value. I predict that outsourcing to Accenture's new AI marketing model will be a complete waste of time and money for Intel as the left and right hands will have poor communication. Worse, if Intels start trusting Accenture with critical market strategy and tech data via easy network access for AI efficiency, it will be a security disaster and conversely very expensive. Intel's products are not junk food or something that can marketed with AI videos. Marketing and tech need a close relationship and a tight feedback loop passing high value information.
Nevertheless, maniacal focus on hyper-competitive product development is a good and necessary strategy. Without such products, there will be nothing to market.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
