Original publication: AutoSens, September 2025 – AI in Automotive: Today’s Breakthroughs and Tomorrow’s Roadmap.

Artificial intelligence is rapidly becoming the backbone of everything in automotive, from advanced driver-assistance systems to next-generation EV platforms and smart factories. To understand how AI is shaping the industry today and what’s coming next, we sat down with Jason Kusmanoff from CoreWeave, the world’s #1 AI cloud provider, to discuss the trends, challenges, and opportunities redefining automotive innovation.‍

1. From driver-assistance systems to next-gen EV platforms, AI is quickly rising to the top as a highly functional, efficient way to easily accelerate innovation and create better driver experiences. What emerging trends are you seeing related to AI and automotive? How do you imagine those trends evolving over the next 1 - 3 years?

We’re seeing AI shift from being a popular option to becoming an essential layer across the automotive stack. In the near term, three areas stand out: advanced driver-assistance systems (ADAS), predictive maintenance, and intelligent energy management for EVs. Each of these relies on AI models that can process massive amounts of real-time data to make decisions that directly impact driver safety and efficiency.

Over the next few years, I expect to see AI evolve into a differentiator for in-car experiences — from personalised infotainment to adaptive safety systems — while simultaneously driving operational efficiency in manufacturing and supply chains. The winners will be those who can develop and deploy these models rapidly, without being bottlenecked by infrastructure. That’s exactly where CoreWeave is helping OEMs and suppliers today.

2. CoreWeave works with a broad range of customers across the full AI development lifecycle, from initial pilot to production. What are some of the more common hurdles you see organisations running into, and what unique capabilities does the CoreWeave AI Cloud offer that helps simplify or accelerate that process?

Most organisations face hurdles in three phases: scaling training workloads, managing costs during experimentation, and operationalising models for real-world deployment. The infrastructure needed for training large AI models is often completely different from what’s required for low-latency inference in a vehicle or factory. That mismatch can create complexity and slow progress.

CoreWeave’s AI Cloud is built specifically to address this. Unlike generalised hyperscalers, we’ve optimised for AI workloads end-to-end — from training to inference — with specialised infrastructure and orchestration that reduces complexity. Our clients don’t need to worry about whether their GPU resources can scale up or down; we provide a frictionless environment so they can focus on developing better AI, faster.

3. AI has a wide range of uses in automotive manufacturing organisations, ranging from augmenting in-car tech to facilitating efficiencies and automations internally within the manufacturer’s organisation. How do you advise potential clients seeking to balance those two areas of AI investment, and where do you see the most significant potential in maximising the value of this transformative AI technology?

We encourage clients to think of AI investment as a portfolio strategy. Enhancing in-car experiences captures consumer attention and creates differentiation in the market, but operational efficiencies in manufacturing often deliver faster, measurable ROI. For example, using AI for predictive maintenance or optimising supply chains can reduce downtime and costs almost immediately.

The most significant value will come when these two domains reinforce one another. Imagine insights gathered from connected vehicles informing smarter production runs, or efficiencies in the factory accelerating time-to-market for new EV features. The organisations that view AI as an integrated capability across the entire lifecycle — not just a siloed feature — will unlock the greatest potential.

4. What role does infrastructure complexity and variance play in helping either facilitate or throttle innovation in automotive manufacturing companies? What’s the value of leveraging an AI cloud versus either working with a hyperscaler or simply going DIY and building a private cloud?

Infrastructure can either be an enabler or a roadblock. Many automotive companies underestimate the variance between AI workloads: training requires massive compute bursts, inference needs low latency at scale, and simulation workloads demand flexible, high-throughput systems. Trying to manage that complexity internally often leads to underutilised infrastructure or long bottlenecks.

Hyperscalers offer breadth, but not the specialisation. DIY private clouds may give you control, but at the cost of agility and scale. The value of an AI cloud like CoreWeave’s lies in delivering specialised, elastic infrastructure purpose-built for AI — enabling teams to innovate without waiting weeks for resources to spin up or wrestling with arcane GPU management. It’s about letting R&D teams move at the speed of their ideas.

5. When it comes to running AI applications, what are the biggest challenges manufacturers are running into today? How does the CoreWeave AI Cloud help them clear those hurdles so they can focus on accelerating AI innovation rather than struggling with AI infrastructure? How do you see that dynamic evolving as the need for AI expands and accelerates?

The biggest challenge we see is that most clouds weren’t designed with AI in mind. Manufacturers need to run simulations with thousands of variables, train increasingly large neural networks, and deliver sub-second inference in connected cars — all while keeping sustainability and compliance front of mind. That combination is daunting.

CoreWeave helps by abstracting away the infrastructure burden. Our clients get access to low-latency, highly scalable compute that’s already optimised for AI, whether they’re training large language models to interpret driver behavior or deploying vision systems for ADAS. Looking ahead, as AI becomes table stakes, the difference will be between those who can run AI reliably and those who can run it at scale, sustainably, and continuously evolving. That’s the role we’re preparing to play.

6. What’s the future of AI in automotive look like? What advice do you have regarding how today’s automotive manufacturers can prepare for what’s coming next so that they can stay on the forefront of what’s possible tomorrow?

The future of AI in automotive is a fusion of intelligence across the vehicle, the factory, and the ecosystem. In three years, I expect AI will not just enhance driving but actively shape vehicle design, manufacturing workflows, and after-sales services. Cars will be smarter, factories more adaptive, and supply chains more predictive.

My advice to manufacturers is twofold: start now, and start integrated. Don’t treat AI as a side experiment — embed it across design, production, and operations. And critically, build your infrastructure strategy around flexibility and scale. The automotive leaders of tomorrow will be those who can adapt rapidly, leverage data responsibly, and continuously push AI innovation into the core of their business.

At CoreWeave, we see our role as enabling that future — not by dictating how manufacturers should innovate, but by giving them the infrastructure canvas to make their most ambitious ideas possible.

‍