The Sentence of Life, the Grammar of Control: Rereading SDV at CES

CES talks about “the future” every year. But what stood out most on this year’s stage wasn’t the shape of the future - it was the language used to describe it. LG began with life. In an age where technology keeps pushing us forward, LG did the opposite: it questioned technology itself. A good life, it suggested, isn’t about faster upgrades - it’s about moments when experience returns to us. That worldview ran through TV, home appliances, the CLOiD home robot, and finally mobility, arriving at one conclusion: SDV doesn’t start inside the car. It’s a way of redesigning the flow of life through software.
Bosch, on the next stage, opened with the opposite language: the ability to bridge the gap between the physical and the digital. And to ensure the idea didn’t remain abstract, Bosch grilled a steak live on stage. It wasn’t a playful show - it was a physical proof that a closed loop of sensing - control - AI - quality outcome can converge on a target result. SDV, too, is not merely about how well AI can talk; it’s about whether software can safely bring physical systems - brakes, steering, suspension, and the cabin - into a desired state.
At CES 2026, SDV didn’t exist only as an “automotive” keyword. It appeared as something completed first in the living room, proven again on a frying pan in the kitchen, and then extended into every space. Following LG and Bosch, I found myself rethinking what the SDV race is really about - and what we should be watching.

By Sang Min Han, AEM han@autoelectronics.co.kr
한글로보기




CES 2026 opened with LG Electronics. The next stage belonged to Bosch. What was striking was that both companies placed human beings at the center - yet the language they used to explain SDV was entirely different. LG framed SDV as a continuous experience across spaces and life. Bosch brought SDV back as the capability to control the physical world. Same CES opening, but a very different temperature.







LG: SDV Completed First in Life

CES 2026 did not begin with product explanations. The opening video - launched with the question, “Hey LG, what does the future look like?” - felt less like a declaration and more like a challenge to technology. Technology has begun to speak in our place, urging us to move faster, upgrade sooner, and follow more quietly. In the race for a “better life,” the essence of life itself has been fading. So the future LG drew wasn’t a flashy demo. It was a life that begins with music that makes you smile, where 9-to-5 doesn’t feel like labor, and where the car exists as a space that can open the heart - a future about us, about humans. Innovation should not run ahead of life or dominate it; only when it touches real experience does a good life become complete.
On stage, CEO Jaecheol Lyu drew a clear line in how LG defines AI:
“When everyone talks about AI, we asked one question: What kind of AI do people need? Our answer was ‘Affectionate Intelligence.’”
His next question was even more symbolic:
“What if AI could step out of the screen and work in real life?”
For LG, AI cannot remain a conversational interface or a cloud service. It must operate inside the realities of life - habits, emotions, and cultures that differ from person to person. And LG chose the home as its starting point. As a brand already deeply embedded in everyday living spaces, LG sees understanding the rhythm and context of real households as its competitive edge in the AI era. LG’s AI home vision is ultimately about giving people their time back.


TV: An AI Hub Designed to Disappear
In the TV segment, Aaron Westbrook brought out both numbers and architecture. The Wallpaper TV - existing “invisibly” in an ultra-slim 9mm form factor - was not a design gesture. It was an attempt to change the connection structure itself. True Wireless removes cables and clusters; LG spoke of 4K 165Hz wireless transmission (with ultra-low latency), improved brightness (3.9x), reflection suppression, and the Alpha 11 AI Processor Gen3. They cited NPU performance at 5.6x versus the previous generation, with CPU up 50% and GPU up 70%, framing it as design as performance.
But the core wasn’t hardware - it was the definition of the hub. Westbrook elevated the TV to the center of the AI home. A multi-AI structure on webOS integrating Google Gemini and Microsoft Copilot, Voice ID personalization, and security via LG Shield - these move the TV from “display” to a gateway through which life data and agents flow. For automotive, the vocabulary feels familiar: personalization, security, OTA, multi-agent - already operating as real-world language.







Home Appliances: Not “Added Features,” but a “Changed Role”
In the home appliance segment, Angela Gozenput described the shift not as feature additions but as a change in role. LG appliances are no longer machines waiting for commands; they are evolving into agent appliances that execute complex goals on their own. LLM-powered LG SIGNATURE refrigerators and ovens handle storage and cooking context through natural language, while built-in camera recognition evaluates cooking progress and intervenes proactively. The point isn’t an AI that “talks well,” but devices moving ahead of you toward life goals - freshness, cooking outcomes, routines.







Robot: A Physical Proof of Sense - Think - Act
At the center of LG’s worldview stood the robot. The home robot CLOiD, introduced by Brandt Varner, appeared as a physical embodiment of LG’s Sense - Think - Act structure. It recognizes the weather, reads the user’s condition and adjusts exercise plans, suggests dinner menus, and controls lighting and temperature - showing not conversational AI, but AI that acts.
LG defined the robot as a home-dedicated agent, connecting vision information and language to physical action through a Vision - Language - Action (VLA) model. And the motor and actuation know-how accumulated through appliances expands naturally into robotics. Here, the robot doesn’t look like an experiment; it looks like a natural evolution where appliances meet AI.
For the automotive world, this structure is not unfamiliar. Sense - Think - Act is exactly the core loop that autonomy and SDV seek inside vehicles: sense via sensors, decide via software, respond via physical systems. What becomes braking, steering, suspension, thermal management, and cabin control in a car becomes lighting, temperature, meals, and movement at home. SDV begins to look less like a “vehicle architecture problem” and more like a universal way of controlling the physical world through software.







Mobility: SDV Is Not a Car, but a Space
Finally, Randon Fuller brought every prior session into mobility. LG positioned itself not as an automotive parts supplier but as an Experience Architect. The declaration - bringing AI experiences that start at home onto the road - redefines mobility as orchestration of space experience rather than a bundle of parts.
LG’s scenarios focused less on braking or steering and more on spatial experience: gaze tracking and gesture recognition, the Mobility Display Solution extending the windshield into a display, continuity where content seen at home continues inside the car and across side windows, and real-time translation for sign language from outside the vehicle. The SDV LG described is the capability to compose a personalized space in real time, beyond the vehicle architecture itself.
LG’s SDV does not begin inside the vehicle. It is already completed in life and at home; the car is simply the next extension. SDV is not a car technology - it is the process of redefining the flow of life through software.

 

---

Bosch: The Grammar of SDV That Controls the Physical World

On the stage after LG, Bosch began with the opposite language. Tanja Rueckert, CDO, and Paul Thomas, President of Bosch in the Americas, emphasized not “portfolio diversity” itself, but the capability to bridge the physical - digital divide. While most companies are strong on one side, Bosch defined itself as a company that can handle both worlds simultaneously. Hardware is strong because it can be continuously leveled up through in-house software capabilities; software is strong because Bosch understands deeply the physical systems on which it runs. The differentiation is not a handful of features, but the grammar of control and systems.
Bosch reinforced this with numbers. They stated a goal of over €6 billion in software and services sales in the early 2030s, and over €2.5 billion in AI investment by the end of 2027. They also cited more than 100,000 employees trained in AI - about a quarter of the workforce. This wasn’t “we do AI,” but a scale that signals a transformed operating DNA.







The Steak Metaphor: Sensor - Control - AI - Outcome Quality
The highlight of the day exploded not in the car segment, but in the kitchen. Chef Marcel Vigneron walked on stage and began grilling steak - right there. The applause wasn’t only for the humor of “steak for breakfast.” It was because Bosch executed, in the shortest possible experiment, the physical - digital integration it had been describing for 30 minutes.
At the core was the sensor-based temperature control (Autochef) built into the 800 Series induction cooktop: continuously reading pan temperature and replacing the fine adjustments humans normally make. On top of that, Bosch layered Bosch Cook AI (generative AI-based), described as being in final testing. You photograph the ingredients and specify the target outcome (doneness, cooking style), and sensors plus algorithms converge on the result. What the chef stressed was not “AI tells you a recipe,” but that a sensing-and-control loop can converge toward a target state.
Translated into automotive terms, it becomes even clearer. Autochef is a controller. Pan temperature is a vehicle state variable. “Medium rare” is a target performance outcome. What matters is not showy AI, but the ability to bring a physical system reliably into a desired state. SDV is difficult for the same reason: even if generative AI speaks beautifully, it cannot make SDV real unless it can safely handle brakes, steering, suspension, thermal systems - physical reality. Bosch compressed that truth into one plate of steak.


“Hardware Keeps Evolving”: OTA in the Language of Daily Life
Immediately after the cooking demo, Rueckert reframed Bosch’s argument in daily-life language. In the past, if you wanted newer features you replaced the product; now, with connectivity and software capability, hardware can evolve long after it arrives in your home. Bosch cited examples like adding air-fry and air-steam functions to connected ovens via OTA updates - at no additional cost to users.
Automotive has heard this line a thousand times: post-sale feature expansion is SDV’s signature promise. Bosch’s point wasn’t OTA itself - it was that Bosch has already been operating the hardware - software structure and discipline that makes OTA real across multiple industries. Like LG, Bosch is not “learning SDV”; it has been training SDV-style operations in other domains.


Vehicle Motion Management: Reducing Motion Sickness with 6DoF Control
When Bosch returned to vehicles, the most striking keyword was again human. Paul Thomas introduced Vehicle Motion Management as hardware-agnostic software that coordinates braking, steering, powertrain, and suspension, allowing vehicle characteristics to be tuned at the press of a button. He added that Bosch has recently extended this to six degrees of freedom (6DoF) control.
The problem Bosch chose was not lap times or acceleration - it was motion sickness. With the note that it can affect up to one-third of adults, Bosch argued that as autonomy grows we become passengers rather than drivers, and motion sickness becomes a barrier shaping the experience of hundreds of millions. Bosch reframed SDV not as “a smarter car,” but as designing passenger well-being as a system. Here, LG’s language of space experience and Bosch’s language of vehicle dynamics unexpectedly converge.







SDV’s “Central Nervous System”: Eclipse S-CORE
Rueckert brought out Eclipse S-CORE (Safe Open Vehicle Core), led by Bosch/ETAS. She compared SDV middleware to the “central nervous system,” arguing that moving beyond fragmented, OEM-specific solutions toward a single standard would improve speed, cost, security, and innovation together. Bosch stepped beyond the frame of a parts supplier: it was a declaration to help design the order of the bottom layer itself. LG is also a key member of Eclipse S-CORE, but while LG says “connect life’s spaces,” Bosch says “rebuild the vehicle’s nervous system in open source.”


By-Wire and the AI Cockpit: Up to “AI That Sees”
Bosch then made its confidence as a systems company concrete through by-wire. It previewed production of a “true brake-by-wire” system with a major OEM, and cited a cumulative sales target of over €7 billion by 2032 for steer-by-wire and brake-by-wire. SDV’s center of gravity is moving downward - toward the digitalization of actuation.
AI connected into the cockpit. Bosch introduced an AI cockpit demo at its booth, describing a structure where one model is a text-based LLM (handling conversation and reasoning), while the other is a visual language model (VLM) that can see and interpret camera and sensor inputs. Examples included searching for parking spots upon arrival, and generating meeting minutes from online meetings attended from the driver or passenger seat. The message was less about individual functions and more about structure: binding language - vision - action into one loop.


Agentic AI Extending into Manufacturing: Collaboration with Microsoft
Bosch did not confine SDV to vehicles. In the industrial domain, it announced an MOU with Microsoft to explore a leap in productivity through Agentic AI. Dan Rodriguez, Corporate Vice President for Manufacturing and eMobility at Microsoft, joined the stage and described a factory reality where complexity increases inside legacy systems - requiring AI agents that monitor, decide, and execute as a new layer above operations.


The Final Link to Autonomous Logistics: Kodiak AI
Finally, Thomas mentioned cooperation with Kodiak AI, supporting development of a production-ready, redundant platform for fully autonomous trucking at scale. It includes sensors (cameras, radar) and steering technologies - placing “autonomous logistics at scale” inside Bosch’s portfolio. Physical control demonstrated on a frying pan becomes redundancy on the road.



SDV Is Not “Talking AI,” but AI That “Hits the Result”

If LG described “SDV completed first in life” and converged toward space experience, Bosch brought SDV back as the capability to control the physical world. What’s interesting is that these two languages point to the same conclusion. Whether you begin with “life” or with “control,” SDV ultimately becomes the way software takes responsibility for the physical world. LG designs that responsibility as a continuous living space - home, living room, robot, vehicle - while Bosch compresses it into a single moment: a steak that proves the closed loop of sensor - control - AI - outcome quality.
SDV is not AI that talks well. It is AI that safely brings physical systems into the desired result. CES 2026 was the moment when LG’s sentence of life and Bosch’s grammar of control overlapped - toward the same future.

AEM(오토모티브일렉트로닉스매거진)

---