Cars in the AI Era: We May Be Underestimating What's Coming

An Underestimated Revolution Is Underway

When most people talk about "AI + automobiles," the first image that comes to mind is usually autonomous driving. But in reality, AI's reshaping of the automotive industry is far more profound and comprehensive than we imagine. The automotive industry in 2026 stands at a historic crossroads — much like when smartphones disrupted feature phones, AI is redefining cars from mere "transportation tools" into "intelligent terminals."

What we may be underestimating is precisely the breadth and speed of this transformation.

Beyond Autonomous Driving: AI Is Permeating Every Pore of the Automobile

Over the past few years, industry attention to AI in the automotive sector has been heavily concentrated on autonomous driving. The L2, L3, and L4 classification system became the core narrative of competition among automakers. But entering 2026, an increasingly clear reality has emerged: autonomous driving is just the tip of the iceberg.

First is the leap in smart cockpits. The maturation of large language models has evolved in-car voice assistants from "following commands" to "understanding intent." The new generation of cockpit AI is no longer a simple voice controller but a "digital co-pilot" capable of understanding context, sensing driver emotions, and proactively recommending services. Imagine sitting in your car exhausted, and the AI automatically dims the ambient lighting, plays soothing music, and recommends the nearest coffee shop — this isn't science fiction but a feature already gradually rolling out in production vehicles.

Second is the revolution in R&D processes. Traditional automotive R&D cycles typically take three to five years, and AI is dramatically compressing this timeline. From wind tunnel simulations to crash testing, from material selection to styling design, generative AI and reinforcement learning are replacing vast amounts of repetitive engineering work. Some automakers have revealed that with AI tools, the design cycle for certain components has been shortened from months to weeks.

Third is the intelligent transformation of supply chains. AI-driven demand forecasting, inventory management, and logistics scheduling are helping automakers tackle increasingly complex global supply chain challenges. Especially against the backdrop of heightened geopolitical uncertainty, AI's predictive capabilities have become a critical pillar of supply chain resilience for automakers.

Large Models in Vehicles: A Quiet Arms Race

Since 2025, a competition around "putting large models in vehicles" has been quietly heating up. Mainstream automakers and tech companies, both domestic and international, have been positioning themselves in the vehicle-mounted large model space, vying to seize the next high ground of intelligent mobility.

Tesla continues to iterate the end-to-end neural network behind its FSD system. On the Chinese front, new-energy vehicle makers such as NIO, XPeng, and Li Auto are integrating or developing their own in-vehicle large models. Meanwhile, tech companies like Huawei, Baidu, and SenseTime are entering as Tier 1 suppliers, providing AI infrastructure for automakers.

Notably, the dimensions of this competition are shifting. The early focus was on "whose model has more parameters," but the core questions have now become:

• On-device inference capability: How can real-time large model inference be achieved within the limited computing power of in-vehicle chips?
• Data closed-loop efficiency: How can training value be efficiently extracted from massive volumes of driving data?
• Depth of multimodal fusion: How can vision, voice, touch, and sensor data be truly integrated into a unified whole?

The answers to these questions will determine the true ceiling of automotive intelligence in the years ahead.

The Underestimated "Second Curve": AI Reshaping Business Models

If the technological transformation is already exciting enough, AI's reshaping of automotive business models may be the part that is truly underestimated.

First, software-defined revenue. Traditional automakers' revenue is essentially locked in at the moment of vehicle delivery. But in the AI era, OTA upgrades, subscription services, and on-demand paid features are creating continuous revenue streams. Tesla's FSD subscription has already validated the viability of this model, and more automakers are following suit. In the future, a car's "lifetime value" may far exceed its purchase price.

Second, data asset monetization. Each smart vehicle generates up to several terabytes of data per day. This data is not only fuel for training autonomous driving models but also a valuable asset for insurance, urban planning, retail, and other sectors. How to extract data value within compliance frameworks is becoming a new challenge for automakers.

Third, the acceleration of Mobility-as-a-Service. The commercialization of Robotaxis noticeably accelerated in 2025. Baidu's Apollo Go scaling operations in Wuhan and Tesla's continued investment in Robotaxi both point to a clear trend: AI may fundamentally change the way people "own" cars. When autonomous driving becomes sufficiently safe and affordable, "on-demand mobility" will pose a substantial challenge to traditional car purchasing models.

Challenges and Concerns: The Other Side That Cannot Be Ignored

Of course, beyond the optimistic narrative, challenges cannot be avoided.

The boundaries of safety and liability remain blurred. When AI participates in or even takes over driving decisions, how should accident liability be determined? Globally, relevant laws and regulations are still struggling to keep pace with technology.

The tension between computing power and energy consumption is growing. Deploying large models in vehicles means higher chip computing demands and greater energy consumption. For electric vehicles, how to strike a balance between intelligence and driving range is a real-world problem.

The red line of data privacy must not be crossed. In-cabin cameras, microphones, and sensors collect data around the clock. Ensuring that user privacy is not abused is a trust issue that AI-powered vehicles must address.

The brutality of industry shakeouts. The rising technical barriers brought by AI mean that small and mid-sized automakers with insufficient resources may be forced out at an accelerated pace. In 2026, consolidation and elimination in the automotive industry will intensify further.

Outlook: The Automotive Industry's "iPhone Moment" May Be Arriving

Looking back at history, every paradigm shift in technology has been accompanied by dramatic reshuffling of industry landscapes. Nokia, the dominant force of the feature phone era, faded into obscurity during the smartphone wave, while Apple — not even a phone manufacturer at the time — became the king of the new era.

The automotive industry in 2026 stands at the same crossroads. AI is not only changing the product form of automobiles but also redefining the fundamental question of "who qualifies as an automotive company." Traditional automakers, new-energy vehicle startups, tech giants, and AI startups — all players are at the same table, and the rules of the game are being rewritten by AI.

What we may be underestimating is not the progress of any single specific technology, but the systemic reshaping of the automotive industry by AI as a "general-purpose capability." From R&D to manufacturing, from products to services, from business models to competitive landscapes — no segment can remain untouched.

The endgame of this transformation remains unknown, but one thing is certain: In the AI era, the greatest risk for the automotive industry is not that change is happening too fast, but that we underestimate the speed and depth of that change.