How Agentic AI Is Reshaping the Automotive Industry: Insights from the IAA Auto Show on Amazon Web Services' Intelligent Collaboration Approach

In September, Munich once again became the focal point of the global automotive industry. This year’s IAA MOBILITY 2025 is not only a feast of technology and design, but also clearly reveals a profound shift in the competitive landscape of the automotive sector: the battlefield has moved beyond the hardware race of “electrification” and fully entered a new era of software and ecosystem competition centered on “intelligence and user experience.”

In this transformation, automotive companies are caught in a common dilemma: despite having vast amounts of data, they are trapped in data silos and fragmented systems; despite understanding that AI is the core competitiveness of the future, they find it difficult to escape the "pilot trap," unable to scale AI from single-point applications to a global scale, and to convert technical demonstrations into real business value.

Against this backdrop, Amazon Web Services (AWS) presented a series of key insights and solutions at this year's IAA exhibition—not only focusing on the technology itself but also committed to helping automotive companies make the leap from "AI experimentation" to "AI operation." This showcases how AWS leverages cloud and AI to assist automotive enterprises in forging a solid and sustainable path for intelligent transformation.

IAA venue; Image source: Amazon Web Services

One, fromTools CollaboratorAI Paradigm Shift

Today, the automotive industry stands at a critical juncture in its transition to intelligent systems. The application of artificial intelligence in the automotive sector is no longer a novel topic—from visual inspection on production lines to predictive maintenance, from training of driver assistance systems to early anomaly detection, AI has long existed as a "tool" for enhancing efficiency. However, as Ozgur Tohumcu, General Manager of Amazon Web Services' Global Automotive and Manufacturing Division, pointed out in his keynote speech, we are currently experiencing a paradigm revolution comparable to the "internet boom of the 1990s": AI is evolving from an auxiliary technological tool to a "collaborator" with autonomous decision-making capabilities. This will redefine the operational models, innovation speed, and competitive boundaries of automotive enterprises.

However, this transformation did not happen overnight; its evolution can be roughly divided into three stages, each marking a fundamental enhancement in the depth of integration between AI and business.

Traditional AI stage (task automation):Focus on optimization in specific scenarios, such as visual quality inspection and fault prediction. Its core logic is responsive processing based on rules and historical data, which, although improving accuracy and efficiency, still remains within the realm of "executing preset tasks" as a tool.

Generative AI Stage (Interactive Intelligence):Represented by chatbots and content generation engines, it has achieved natural language understanding and content creation, reshaping customer interaction and service experience. It marks the beginning of AI "generating" rather than just "recognizing," but its operation still heavily depends on human questions and instructions.

Agentic AI System Stage (Autonomous Collaboration):The true paradigm shift lies in the maturity of Agentic AI technology. At this stage, systems no longer merely respond to commands but are able to autonomously plan, execute, and coordinate multiple subsystems to accomplish complex tasks based on objectives.

Mr. Ozgur uses battery manufacturing as an example: "Imagine an Agentic AI system where intelligent agents can respectively monitor key raw material price and supply fluctuations, track global shipping costs and logistics conditions, and analyze relevant policy and regulatory changes. Once these agents work collaboratively, by simply setting goals such as 'reduce battery operating costs by 10%', the system can autonomously plan and execute full-process optimization. This real-time processing of multi-source data and efficient collaboration among agents is something no human team can match."

Ozgur Tohumcu, General Manager of Global Automotive and Manufacturing at Amazon Web Services, delivered a keynote speech; Image source: Gasgoo.

The automotive industry is undergoing a profound paradigm reconstruction—especially with the application of AI, particularly Agentic AI technologies, which are not merely functional upgrades but are reshaping the underlying logic and value standards of the industry. In this context, the ability to successfully transition from "tool usage" to "intelligent collaboration" is likely to be a key variable for companies in shaping and even seizing future market discourse power.

2. Data Silos, System Lock-In, and the Gap of Scalability

Despite the unprecedented innovative landscape painted by the technological paradigm represented by Agentic AI for the automotive industry, companies still face multiple real challenges in promoting the large-scale implementation of AI, ranging from data infrastructure and system integration to organizational capabilities.

1. Data Dilemma: Concerns About "Fuel" Quality

Although the automotive industry possesses massive amounts of data, the data is often scattered across heterogeneous architectures, multiple clouds, and on-premises systems, resulting in isolated data silos that are difficult to integrate.

Data fragmentation and quality issues make it difficult for enterprises to build a unified and reliable data foundation, preventing them from providing high-quality, real-time consistent "nourishment" for AI training and inference, and slowing down the pace of AI iteration.

2. The Pain of Integration: How New and Old Systems Collaborate

In addition to data challenges, automotive companies are generally facing deep-seated obstacles in system integration. Traditional core business systems, such as ERP (Enterprise Resource Planning), PLM (Product Lifecycle Management), and CRM (Customer Relationship Management), often operate independently with different protocols, creating a significant technical debt.

Integrating innovative AI pilot projects seamlessly with traditional core business systems involves not only complex technical adaptation but also requires a profound transformation of existing business processes.

However, many companies' AI applications are still at the "isolated experimentation" stage and have not achieved cross-system integration, resulting in the so-called "pilot trap." The true business value needs to be realized through end-to-end automated processes, which requires AI capabilities to penetrate departmental barriers and achieve bidirectional interaction and functional collaboration with traditional systems.

3. The Difficulty of Scaling: The Huge Leap from PoC to Enterprise-Level Deployment

Advancing AI, especially Agentic AI systems, from one or two successful proofs of concept (PoCs) to enterprise-wide deployment is a systematic undertaking. This not only involves significant cost investment and infrastructure expansion but also includes the design of governance frameworks, ensuring security and compliance, as well as the profound challenges of talent and cultural transformation.

As illustrated by the case shared by Mr. Ozgur, even industry giants like Toyota—which has 1,800 dealerships, 63,000 employees, and dozens of factories in North America—have long been plagued by data integration issues. Data is scattered across various databases, data lakes, third-party data sources, and dealer-owned systems, making it difficult to form a unified view, let alone provide real-time, reliable data support for AI.

Moreover, as the scale of AI applications expands, companies must continually address organizational challenges such as ethical and compliance reviews, as well as upskilling their existing workforce. Especially in the European market, issues like data sovereignty and localization requirements have become unavoidable topics for many automotive companies when deploying AI on a global scale.

3. The Path to Breakthrough: Cornerstone, Engine, and Full-Chain Reach

In the face of data fragmentation, system disconnection, and scalability challenges, automotive companies need not just point solutions, but a systematic framework that spans underlying data, intelligent applications, and overall compliance. As demonstrated by Amazon Web Services at this year's IAA, its true value lies in providing enterprises with an end-to-end AI transformation path—from a unified data foundation to an Agentic AI-driven business engine, from full value chain scenario solutions to compliance assurance, fully supporting automotive companies in bridging the gap from "experiment" to "operation."

1. Cornerstone: Unified and Trusted Data Platform

If data is the "new oil" of the AI era, then a high-quality, easily accessible, and compliant data platform is the infrastructure for mining this treasure. To this end, Amazon Web Services is committed to helping enterprises integrate data scattered across multiple clouds, on-premise systems, and third-party sources to build a truly interconnected and real-time data platform. For example, one of the scenarios demonstrated by the company at the IAA exhibition: analyzing vehicle false braking incidents through AI, from customer feedback, fleet-level analysis to OTA repair release, is all completed on a unified data platform—breaking the data silos between engineering, product, and service in traditional enterprises.

In a market as sensitive to data sovereignty as Europe, Amazon Web Services has not shied away from challenges but has actively taken them on. By investing 7.8 billion euros in building the "Amazon Web Services European Sovereign Cloud," the company plans to launch its first region in Brandenburg, Germany, by the end of 2025. This is a completely independently operated cloud architecture with localized talent and technology, which maintains the high availability and security standards of Amazon Web Services' global infrastructure while fully complying with the EU's stringent data compliance requirements. For the automotive industry in Germany and even Europe, this undoubtedly provides a strong and credible foundation for data innovation.

2. Engines: Generative AI and Agentic AI: FromAssistance Autonomous”

When the data foundation is solid, generative AI and agentic AI become the core engines for unlocking data value and driving business automation.

The Volkswagen Group, for example, utilizes Amazon Web Services' Amazon Nova foundational model to automatically generate marketing images that meet brand standards, significantly saving labor and time costs. Audi has gone even further: it has collaborated with Amazon Web Services to develop and deploy four intelligent agents that focus respectively on cost control, security protection, infrastructure, and resource scheduling. These agents work together, reducing some critical operational processes from 30 minutes to just 15 seconds.

At the same time, the BMW Group and Amazon Web Services have collaborated on a decade-long journey of innovation, yielding fruitful results. Currently, more than 24.5 million connected vehicles globally from BMW are integrated with Amazon Web Services' cloud, processing over 16.6 billion service requests daily, with a system reliability of 99.95%. Additionally, leveraging Amazon Web Services' technology, BMW has developed a new electronic and electrical architecture for the Neue Klasse models. This collaboration is helping BMW accelerate software development and enhance user services.

Momenta empowers its autonomous driving data flywheel through Amazon Web Services' "AD Insight Hub" solution, leveraging multimodal large models to efficiently process massive scene data and accelerate algorithm optimization in extreme driving scenarios.

3. Tentacles: Scenario-based Solutions Penetrating the Entire Value Chain

True large-scale implementation requires integrating AI into every aspect of the business. Amazon Web Services, through five major exhibition areas and a joint exhibition area with partners, demonstrates the deep application of its solutions in four key fields: research and development, production, retail, and after-sales.

IAA; Image source: Amazon Web Services

R&D Domain:At this year's IAA exhibition, Amazon Web Services (AWS) introduced its "SDV Accelerator," aimed at addressing one of the core challenges in the automotive industry's transition to software-defined vehicles. By deeply integrating cloud development tools with third-party solutions validated by the AWS Marketplace, the accelerator helps automotive companies consolidate fragmented development environments and create rapidly deployable digital work platforms, thereby significantly shortening the development cycle of new models.

Production and Supply Chain DomainAgentic AI systems are redefining the boundaries of complex manufacturing and supply chain management. In the exhibition demonstration, AWS further showcased the practical application of Agentic AI in manufacturing operations—real-time identification of production bottlenecks, prediction of equipment failures, and automatic triggering of adjustment and optimization decisions through the system, upgrading the traditional manual response mechanism to a closed-loop automatic optimization, thereby enhancing factory operational efficiency.

In-vehicle Agentic AI Assistant Solution; Image Source: Amazon Web Services

Retail and Customer Experience Domain:The digital transformation of car-buying experiences has become a key driver of brand differentiation. Leveraging Amazon Connect, AWS deeply integrates generative AI into customer engagement processes: virtual assistants efficiently handle common inquiries, while human agents receive real-time, AI-generated customer profiles and behavioral insights during calls, enabling them to deliver highly personalized services. This seamless experience, spanning from initial consultation to after-sales support, is redefining the way consumers interact with automotive brands.

For example, Zeekr is building its Marketing Intelligence Platform based on Amazon Web Services' AI services (such as Amazon Bedrock). This platform aims to integrate data from the entire sales, marketing, and after-sales service chain, empowering terminal stores and customer service systems, with the goal of providing users with a smarter and more personalized service experience.

After-sales and service domain:Intelligent after-sales service is a crucial battleground for enhancing user loyalty and operational efficiency. Amazon Web Services and KGDATA have jointly developed the "Automotive Services GraphRAG" solution, which leverages knowledge graph and large language model technologies to build an intelligent assistant system that integrates maintenance history, fault databases, and diagnostic processes. This enables technicians to quickly identify issues and recommend solutions, significantly reducing vehicle downtime.

Four, Journeying towards an Intelligent Future with Chinese Innovators

As the intelligent transformation of the automotive industry enters deeper waters, Chinese power is accelerating its global deployment. From electric vehicles to intelligent driving assistance, from vehicle networking services to digitalized supply chains, Chinese car manufacturers and technology companies are no longer confined to the domestic market, but are now turning their attention to Europe, North America, and even the entire globe.

However, the road to going global is long and challenging: technical architecture must be globally scalable, data compliance has become a key entry barrier for major markets, and cultural differences and ecosystem integration are long-term and complex challenges.

Amazon Web Services (AWS), with its global infrastructure, extensive compliance practices, and vast partner network, is becoming a trusted "companion" in the overseas expansion of China's automotive ecosystem. It offers a comprehensive empowerment system that includes an underlying data foundation, Agentic AI-driven engines, scenario-based solutions, and a global compliance ecosystem. This not only reduces the cost of technology reuse for enterprises but also provides end-to-end support for Chinese companies to participate in global competition, from architecture to compliance, and from development to operations.

In conclusion:

At this IAA Auto Show, Amazon Web Services presents far more than just a collection of technical tools; it offers a comprehensive transformation framework encompassing strategy, architecture, and practical implementation. It is no longer merely a cloud service provider but a strategic partner exploring the boundaries of intelligence together with automakers.

Looking ahead, we expect to see more automakers and technology partners working closely together to build a new generation of business models that are not only smarter and more personalized, but also safer, more efficient, and compliant. In this industry transformation marked by “Agentic AI,” those enterprises that take the lead in breaking data silos and achieving large-scale AI operations will no longer be mere users of technology, but will become pioneers redefining the value of automobiles.