As artificial intelligence redefines vehicles, the big question is no longer what AI can do, but who will build the foundations on which it can thrive—the data, infrastructure, validation frameworks, and digital ecosystems that make intelligent mobility possible. Organized by NASSCOM Center of Excellence, in collaboration with the Centre of Excellence in Advanced Automotive Research (CAAR), and IIT Madras, the event brought together leaders from across the automotive, technology, startups, and research ecosystems. AI dominated many conversations, but the essence of the event ultimately pointed to the unglamorous work of building the shared infrastructure, validation processes, and ecosystem structures that will determine whether automotive AI scales responsibly — or stalls! As the event’s exclusive media partner, AEM brings you this report.

By Sarada Vishnubhatla_sarada@autoelectronics.co.kr
한글로보기

Related Story: Insights from Vector Informatik′s Global and India Leadership







Sanjeev Malhotra CEO, MeitY Nasscom CoE


 
Opening the event, Sanjeev Malhotra, CEO, MeitY NASSCOM Centre of Excellence (CoE), set a collaborative tone. As emerging technologies such as artificial intelligence, autonomous driving, and software-defined vehicles continue to transform the automotive landscape, there is a need for platforms that bring together industry, academia, startups, government bodies, and research institutions to exchange ideas, challenge conventional thinking, and accelerate innovation.
His reflected on a premise that threaded through nearly every session. 
“It is clear that the next chapter of automotive innovation will depend as much on ecosystem architecture as on any individual technology. With the central government’s strong backing and encouragement, I see good adoption of this taking place in the coming months,” he said.
For global automotive electronics suppliers and technology companies — including those operating across Korea, Japan, and Europe — India's trajectory here carries direct relevance. The country is not only strengthening its position as an engineering and R&D hub but is increasingly shaping the platforms, standards, and innovation frameworks that will feed into global SDV supply chains.



Building an Automotive Innovation Ecosystem

There is now a compelling need to create technologies tailored to India's unique requirements, and the event resonated with it too. 
Dr. Anand Lakshmanan, Senior Project Advisor, CAAR (IIT Madras), pointed out the gap between importing solutions and building them indigenously. "CAAR is incubating startups to support innovative technology development. On one hand, we support such niche technologies, and on the other, we work towards resolving industry challenges," he said.
Agreeing with Sanjeev, he said that stronger collaboration between academia, industry, and government is not optional. In fact, it is the precondition for translating research into deployable solutions. 
He proposed a new perspective, "In India, we are still at an early stage of adoption of SDVs. So why not bring in AI into it at this stage itself?"



From Software-Defined Vehicles to AI-Defined Mobility

The path the automotive industry is taking globally is now established. Several speakers acknowledged and described a world already moving beyond the SDV paradigm toward what some called ‘AI-defined mobility’ — a stage in which artificial intelligence is embedded not only within vehicles but across the entire engineering lifecycle, from planning and design through to deployment and iteration.
The implications extend well beyond technology. AI-driven engineering workflows are expected to reshape organizational structures and redefine the skills required across the OEM–Tier 1–technology ecosystem. India is definitely catching up albeit in a small way.
Dr. Guru Prasad A S, Global Head – Mobility Practice, L&T Technology Services, described the emergence of hybrid delivery models where engineers and AI agents collaborate throughout the development process. Rather than replacing human expertise, AI becomes an accelerator — improving productivity and compressing development cycles and enabling what LTTS describes as Engineering Intelligence. He identified three structural obstacles to large-scale AI deployment in India's automotive sector.
The first is data access. "While customers understand the value of AI, their data confidentiality concerns often limit the availability of the datasets required to build and train models." The second is the absence of a structured AI adoption roadmap — a particular challenge in an industry where safety-critical systems demand rigorous, and phased validation. The third is architectural. "Today, many organizations are trying to embed AI within individual applications. The real game changer will be to introduce a dedicated AI layer within the software stack — one that sits above the middleware layer, processes vehicle data centrally, and serves intelligence to multiple applications."
Anuraag Bhardwaj, Vice President & Head – Automotive Industry Platform, Capgemini, pointed to the convergence of more powerful computing platforms, expanding real-world datasets, and growing edge-case libraries as factors that are already shrinking vehicle development cycles. He was confident that, "India is evolving technologically. Our chips are becoming more powerful, and we are increasingly reducing cloud-related latency. The technology foundation is now in place."




Top row (from left): Sanjeev Malhotra, CEO, MeitY NASSCOM Centre of Excellence; Dr. Anand Lakshmanan, Senior Project Advisor, CAAR, IIT Madras; Dr. Guru Prasad AS, Global Head – Mobility Practice, LTTS; Anuraag Bhardwaj, Vice President & Head – Automotive Industry Platform, Capgemini; Rajagopalan Rajappa, CTO, Tata Elxsi

Bottom row (from left): Naveen Krishnappa, Technical Lead & Product Expert, Vector Informatik; Thiru Srinivasan, CEO, CAAR; Vikrant Kumar, AVP & Delivery Business Leader, KPIT; Srinivasa Prasad M N, Associate Director, Accenture; Pooja Arun, Vice President, Mercedes-Benz Research & Development India




The Hidden Infrastructure Behind Automotive AI

If AI is the visible surface, infrastructure is the substrate — and several speakers discussed the possibility that it is the substrate that may as well determine the scale.
Rajagopalan Rajappa, CTO, Tata Elxsi, laid out what that substrate requires: high-definition maps, large-scale driving datasets, long-tail scenario databases, digital twins, simulation environments, and open AI frameworks capable of serving as common foundations for innovation. Without these shared assets, he felt, sovereign automotive AI models — capable of addressing India's specific road conditions and mobility patterns — cannot be built.
"With generative AI and transformer architectures coming into play, it is no longer just about Software-Defined Vehicles," he said. "The direction is clear. We are moving towards AI-Defined Vehicles, where AI increasingly shapes feature definition, user experience, software architectures, and even the evolution of autonomous driving systems. The most important consideration, however, remains the development of sustainable and responsible AI."
The infrastructure challenge extends to software deployment itself. Naveen Krishnappa, Technical Lead & Product Expert, Vector Informatik, described OTA updates as a foundational SDV capability — but one that carries risks absent in consumer electronics. Unlike smartphone updates, vehicle software changes often affect safety-critical functions. Managing deployment across millions of vehicles requires secure cloud infrastructure, rollback mechanisms, and safety architectures that guarantee vehicle operability even when updates fail.
Cybersecurity was a related concern. Thiru Srinivasan, CEO, CAAR, highlighted the need to build resilience across hardware, software, and testing environments as vehicles become more connected — a challenge that grows in proportion to the intelligence being added.



Trust, Validation, and Safety: The Real Challenge

Building intelligent systems is becoming achievable. Demonstrating that those systems are safe, reliable, and trustworthy is a different problem — and by most accounts at this event, the harder one.
Vikrant Kumar, AVP & Business Leader – Delivery, KPIT, was precise about where validation breaks down. Large datasets may be available for model training, but ensuring that validation datasets are sufficiently diverse and representative of real-world conditions is a separate challenge. "The gap between controlled development environments and unpredictable road conditions adds further complexity. Synthetic data and simulations can accelerate development, but real-world validation continues to be indispensable."
Regulatory readiness compounds the difficulty. AI technologies are advancing faster than the frameworks governing their deployment, validation, and accountability. Trust, Kumar suggested, will be built incrementally — through demonstrated safety outcomes, effective regulation, and responsible adoption, not through declarations.
Srinivasa Prasad M N, Associate Director, Accenture, examined the accountability dimension. As vehicles become more software-driven, who actually owns the responsibility across a wide ecosystem of suppliers, technology providers, and engineering partners?
He noted, “The OEMs remain the customer-facing entity and hence, ultimately accountable. Having said that, it's a collateral damage and the rest of the ecosystem – vendors, tier-1s, the engineering service providers and others – will feel the impact too. Though it is a fact that SDVs are still an evolving domain and concepts are yet to mature, I believe that strict regulatory and safety frameworks will enhance both the credibility and accountability.” 
Dr. Guru Prasad reinforced the point that safety-critical systems such as braking, steering, airbags, and ADAS require a phased deployment strategy, progressing deliberately from low-risk applications toward higher-stakes use cases. 
And that is exactly why adoption of responsible AI becomes imperative leaving little room for shortcuts.



Beyond Vehicles: The Rise of Intelligent Energy

Several sessions extended the conversation beyond the vehicle itself to the energy systems it increasingly depends on — and may eventually contribute to.
Pooja Arun, Vice President, Mercedes-Benz Research & Development India, described the evolution of battery management systems from being reactive monitoring tools into becoming predictive intelligence platforms that are capable of anticipating battery behaviour and optimizing performance in real time. 
Her framing was instructive, "Battery technology could be like a processor and software intelligence could be like an operating system. If we don't have the intelligence or a powerful processor, we cannot make it efficient."
The implication is significant. Electrification is not only a chemistry problem. The software intelligence managing and optimizing battery performance may matter as much as the cells themselves.
On infrastructure, deploying chargers alone is insufficient. She explained, "The underlying energy infrastructure must also be capable of supporting growing demand. This is where technologies such as Vehicle-to-Grid, smart charging, and AI-enabled energy optimization become increasingly important."
EVs are not merely consumers of electricity. They are distributed energy assets — capable of storing, managing, and supplying energy when required. The discussion showed that the boundary between mobility and energy infrastructure may already be crumbling. 



Building a Connected Mobility Future

The event's final sessions widened the frame further — from vehicles and energy systems to the broader architecture of connected mobility.
Thiru Srinivasan pointed to emerging opportunities in vehicle-to-vehicle communication, centralized computing architectures, cybersecurity, and advanced HMI systems designed to reduce driver cognitive load. Dr. Ramakrishna Srinivasan, CEO, Mobility and Intelligent Transportation Collaborative (MInT), IIT Madras, pushed the boundary further still — to seamless data exchange across vehicles, infrastructure, and transportation networks.
"As data becomes increasingly available and computing capabilities continue to advance, we are better positioned to build models that help us understand and optimize mobility systems as a whole," he said. "Improving system-level efficiency is critical to creating a more sustainable transportation ecosystem."
The ambition is real. So are the gaps. What the Automotive Innovation & AI Meet 2026 made clear is that the industry understands where it needs to go. But the harder work — building the data foundations, simulation environments, validation frameworks, and collaborative structures needed to get there — is still largely ahead of it. India's role in that aspect is growing. It waits to be seen if the ambition is achieved sooner or later. 

AEM(오토모티브일렉트로닉스매거진)