Intel’s expanded SDV automotive platform will bring these 3 much needed benefits for the auto industry

Intel is here to help the automakers, but how the industry responds will determine if we might see the "Intel Inside" slogan for your future car.
#intel #ces2025 #automotive #ev #ai #acu #gpu #arcgraphics #car

By Vijay Anand - 11 Jan 2025

Jack Weast, Intel Fellow, VP and GM of Intel Automotive, had much to share about their all-new expanded approach for this year, including launching new partnerships and a new ACU and GPU for their platform.
(Image: Intel)

The auto industry is facing an uphill battle as it struggles to transform and find balance in offering suitable ICE (internal combustion engine), Hybrid or full electric vehicle (EV) models for their respective consumer markets, battling against high costs for navigating this industry change, as well as trying to stay competitive against a deluge of Chinese EV models. You might have heard that Volkswagen nearly had to consider closing some of its factories in its home ground, which it narrowly averted, amongst other harsh realities of the auto industry.

Of all the companies, you would think Intel has the least relevance in helping the auto industry. However, it is precisely the opposite, where its heritage in advanced processors and managing workloads will help make a difference.

Last year, Intel’s automotive division debuted its first-generation software-defined vehicle (SDV) system-on-chip (SoC) family to power a multitude of functions in a modern car and bring the AI PC experiences to the car to redefine driving experiences. This year, at CES 2025, Intel debuted an expanded automotive product portfolio and new partnerships designed to accelerate and help traditional automakers transition to EVs and SDVs. Most crucial is its whole-vehicle platform approach, where its system-level view focuses on three elements to reduce cost and advance the vehicle architecture:-

High-performance software-defined central compute solutions
ECU consolidation with software-defined zonal controllers
EV power-train solutions

Intel Automotive is bringing innovative solutions that reduce cost in the SDV revolution. Our whole-vehicle approach, combined with cloud integration, delivers a complete solution that drives down the total cost of development and deployment while empowering automakers to build the future of mobility faster, more efficiently and more profitably. -- Jack Weast, Intel Fellow, vice president and general manager of Intel Automotive

A focus on reducing energy costs and increasing energy efficiency of EVs

Intel debuted a new kind of processor - the Adaptive Control Unit. (Image: Intel)

By the 2030s, Intel estimates there will be a 50% higher demand for lithium carbonate than current supply estimates. Weast simplified the state of the matter by pointing out that EVs today are just like how early mobile computers were, where battery life was short, and batteries were bulky.

Intel has identified that more advanced modulation techniques can derive a lot of savings to help save on energy conversion losses all through the electric vehicle’s power train. However, the reason why the industry isn’t yet utilising such optimised pulse pattern modulation is because traditional automotive microcontrollers and DSPs do not have enough horsepower due to the sequential processing flow in most traditional automotive microcontroller units (MCUs). This means they are unable to process another signal if it’s already occupied with the current request. Simply adding more processing cores doesn’t solve this issue as it introduces latency complications, which isn’t adequate to tackle safety-critical workloads.

Typical power loss across each step in an EV powertrain. (Image: Intel)

This is why Intel is today launching a completely new processor – the Adaptive Control Unit (ACU). Weast says that the Intel ACU U310 is a microcontroller for a software-defined age that is capable of tackling new kinds of control algorithms with up to 1,000 times faster data processing than conventional approaches. The result of deploying this new ACU into an EV drivetrain? Intel says there’s energy efficiency at every level, reducing the size, weight and cost of its cooling solution, reducing motor size down by 25% as it can now run more efficiently, thus reducing costs, and is able to reduce the battery voltage by up to 20% while still getting the same range.

Rising to the challenge

To prove the reliability and dependability of Intel’s ACU, Stellantis Motorsports is developing an electric power train in partnership with Intel to enter the FIA Formula E motorsports. For those unfamiliar, Stellantis is one of the top automakers and has 14 automotive brands under its fold, such as Citroen, Fiat, Maserati, Peugeot, Jeep, Chrysler and many others. They are partnering with Intel to rely on the best next-gen inverter control to help them fight for the top spot at the podium.

With the inverter playing a crucial role in controlling the electric motor by using power from the battery as well as recovering energy from the braking phase, any gains in efficiency will translate into an important winning edge in these races where even small savings can go a long way. To that extent, Stellantis is banking on Intel’s semiconductor technology and software expertise to deliver optimised power delivery and enhance overall vehicle performance.

When one development gives rise to solving another aspect of efficiency

Through the process and development of the ACU for EV power train applications, Intel also realised that it can also be used for a wide variety of zone-oriented applications as well.

The fundamental difference between the ACU (and its flexible logic control) and a traditional microcontroller is its ability to consolidate multiple microcontroller workloads (currently handled by multiple ECUs) onto a single domain (or zonal) controller. Automakers and tier-one suppliers are looking for a single-chip solution that's also adaptive and re-configurable in real time on the fly. This helps lower both development and deployment costs.

Embracing the whole-vehicle platform

With the expansion of Intel’s automotive portfolio, Intel can now address industry challenges through a whole vehicle architecture. This fully embraces a software-defined mindset across the entire vehicle, and Weast mentions that this is when significant cost savings can be achieved.

The vehicle architecture can finally take a big leap forward with the power controller, zonal controller and high-performance software-defined vehicle SoC working together. By embracing a fully software-defined concept, workloads can be shifted from one socket to another, similar to how a data centre manages workloads across the server farm to spin up or down resources as the workloads are anticipated.

For example, E-mirrors operated by an array of cameras can use a zonal controller for sentry monitoring to save on power. (Image: Intel)

For example, an EV car in sentry mode can utilise a 4W edge/zonal controller for activity monitoring and wake up the main compute engine when activity is detected to take the next string of actions instead of running the central compute engine all the time and thus burning up to 50W of power continuously. Yet another form of savings is to have the energy management system talk to the power train subsystem to derive even more savings via known user/router behaviour to capture more synergy and efficiencies. This is part of Intel’s expertise in managing compute power based on actual workload needs through dynamic variable voltage scaling, which has been their processors' mainstay for the last two decades.

In a whole vehicle platform, there are many more opportunities to optimise energy usage by intelligently turning off certain ECUs predictably and deterministically waking them up when required. For example, you wouldn’t need the seat heater ECU powered during summertime, right? Similarly, many other non-critical functions can be calibrated for greater efficiency and power savings in a more intelligent manner.

Reducing R&D and Deployment costs

Rob Boetticher, director of Automotive Solutions Architecture at Amazon Web Services (AWS), joins Jack Weast, VP of Intel and GM of Intel Automotive, on stage at CES to announce the Intel Virtual Development Environment (VDE) on AWS. (Image: Intel)

What about R&D costs to develop these new vehicles? Remember we mentioned that automakers are up against high development costs to offer new platforms to take on the growing interests of hybrid and EV options? While Weast pointed out such developments are already moving to the cloud, it won’t do anyone any good if nobody is aware of how the system will actually run on the in-vehicle silicon. This is why Intel Automotive and Amazon AWS are teaming up to offer the exact same in-vehicle silicon that automakers will deploy in their cars mirrored in the AWS environment. This will allow developers to check both functional aspects and performance accurately while developing on the cloud. It’s estimated to offer a 15 to 20% reduction in overall platform development time, which is key to market competitiveness.

OEMs currently do a great job at hardware validation, but software validation is a difficult and time-consuming hurdle. According to SBD Automotive, a global automotive research and consulting firm helping develop smart, safe and sustainable mobility, up to 45% reduction in validation time is attainable through Intel’s approach.

Intel Automotive VD environment on AWS is now available. Furthermore, Intel and Red Hat are now working together to bring Red Hat’s in-vehicle operating system to help bring true software-defined architectures to the automotive industry.

Entering the AI Cockpit era

With AI PCs out there, consumers’ expectations are changing to demand that their cars feature the same kind of AI capabilities that one is enjoying on other devices. Be it the growth of LLMs, demand for AI PC-like gaming in-vehicle or the desire for automakers to differentiate their in-vehicle experiences, Intel is hearing that their customers want more computing, more AI and more graphics to accomplish more inside the vehicle.

While Intel launched Arc graphics for automotive in last year’s platform, Intel is now riding on the back of the recent B-series discrete graphics cards (check out the performance here) to launch the Intel Arc B-series for Automotive at CES 2025. Boasting 200 TOPS and up to 12MB of onboard memory, Intel expects automakers to be able to run an entire LLM inside the vehicle and eliminate cloud latency issues when asking a question and waiting for a response – more so when you’re underground, inside a tunnel and many other scenarios. With this, you’re essentially getting an Intel Battlemage class GPU in your car to enjoy AAA-title PC gaming. Enjoy next-level 3D visual experience with SR-IOV (more on this later) built into the silicon. In fact, Intel Automotive is teaming up with Epic’s Unreal Engine to deliver richer in-car experiences like faster graphics and smoother displays!

Here's a walkthrough of the whole platform coming together in this Intel Automotive demo area at CES 2025:-

A note on SR-IOV: In a nutshell, it allows the silicon to support virtualised functions and allocate work to virtual machines to enforce the separation of workloads. This is essential when running mixed criticality workloads – safety and non-safety applications – which can run concurrently on the same operating system. This brings large improvements in software system architecture, resource utilisation, design, testing and deployment without compromising on safety concerns.

Intel Arc B-series Graphics for Automotive is set for production by the end of 2025.

This is a reference of Intel's first-gen SDV automotive platform and first-gen Arc graphics performance, which will only get better with the new Arc B-series graphics engine for Automotive in time to come. (Image: Intel)

A start of more partners to come?

Marques McCammon (right), president of Karma Automotive, joins Jack Weast, VP of Intel and GM of Intel Automotive, to endorse Intel’s whole-vehicle systems approach to re-architecting the automotive. (Image: Intel)

While a lot depends on how the industry will react and respond to Intel’s SDV architecture, at CES 2025, USA’s only dedicated full-line luxury vehicle manufacturer, Karma Automotive, announced on stage that it would be implementing Intel’s whole SDV architecture on its upcoming 1,000 horsepower Karma Kaveya. This super coupe is currently slated for production in late 2026. Hopefully, this announcement will give other automakers more confidence and interest to pursue feasibility studies with Intel that could hopefully lower their overall development costs, total car value and increase power efficiency while also maintaining their desired level of performance and expectations to their customers.

In this difficult time for the automotive industry, Intel’s whole vehicle platform solution could be the catalyst some automakers need for cost reduction and a leg up on innovation.

Our articles may contain affiliate links. If you buy through these links, we may earn a small commission.