Bolt, Mercedes-Benz, Volkswagen And Volvo Build Robotaxis And Level 4 Cars On NVIDIA Hyperion

Human Verified 👁 19 0

Share to: , and

Robotaxi built on NVIDIA Hyperion driving through a European city street at dusk

NVIDIA used GTC Berlin on Wednesday to put Europe’s car industry on its Hyperion platform. Bolt with Lucid, Mercedes-Benz, Volkswagen Commercial Vehicles and Einride are showcasing autonomous-mobility projects on the computing platform, covering level 3 and level 4 vehicles and future robotaxi fleets, while Volvo Cars is adopting Hyperion together with the NVIDIA Alpamayo Stack for automated driving.

The announcement, made on October 21, also brings new tools for the developers building those vehicles: Halos SAFE, a framework for generating traceable safety evidence, plus Alpamayo 2 Edge, the Alpamayo Agentic Factory and new Cosmos 3 Nano model variants for building and evaluating self-driving systems.

Engineers validating an autonomous vehicle on NVIDIA Hyperion with simulation and safety testing screens

Hyperion’s European Line-Up: Robotaxis, Cars And Freight

Bolt has selected the Alpamayo Stack for its European robotaxi programme, starting with future Lucid vehicles built on Hyperion. Bolt plans to begin gathering real-world driving data in Madrid in 2027, subject to regulatory approval, with Germany a priority market in a planned European rollout of at least 25,000 level 4-capable Lucid vehicles.

Volkswagen Commercial Vehicles is presenting its autonomous ID. Buzz AD platform on Hyperion at the show, using the common computing and sensor design to support multiple driving systems. Mercedes-Benz is extending its Hyperion and Alpamayo partnership toward level 3 and level 4 capability, and says its MB.DRIVE ASSIST PRO system, developed with NVIDIA, will soon get approval for urban point-to-point driver assistance in Germany.

Volvo Cars is combining its own safety engineering with Hyperion, the Halos OS and the Alpamayo Stack so future Volvo cars can navigate complex real-world environments door to door. Freight is in the picture too: Einride is building its next-generation Einride Driver on Hyperion, and TIER IV with the Technical University of Munich plan to use the platform for autonomous public transport in Germany. Aptiv, Valeo and Schaeffler are building Hyperion-based computing platforms that other manufacturers can integrate.

Halos SAFE And The New Alpamayo Tools

Halos SAFE — the Safety Assurance Framework for Evaluation — sits inside NVIDIA’s Halos safety system. It connects simulation, software- and hardware-in-the-loop evaluation, on-road testing and fleet monitoring through defined scenarios and measurable criteria, so developers can generate safety evidence that regulators and partners can trace. dSPACE, IVEX and TÜV Rheinland are collaborating on the evaluation workflows, alongside an ecosystem that includes JLR, Mercedes-Benz, Volvo Cars, 51Sim, aiMotive, AVL, Deepen AI, Foretellix and Synopsys.

On the model side, Alpamayo 2 Edge is a 2-billion-parameter reasoning model built on Cosmos 3 Edge’s Nemotron-based foundation, optimised for the DRIVE AGX Thor platform and designed to run at 10Hz for in-vehicle scene understanding and motion planning. AlpaJudge, a human-calibrated evaluation model, scores driving explanations to filter weak training annotations. Foundation4D labels multicamera video without lidar, new Cosmos 3 Nano variants expand synthetic data generation and closed-loop simulation, and the Alpamayo Agentic Factory — an open blueprint expected in early 2027 — uses AI agents to connect data preparation, model post-training and simulation evaluation.

The data machinery is already being used at scale. Uber and NVIDIA are developing a Cosmos-powered data factory for autonomous-vehicle development, and Porsche is leading a Volkswagen Group workflow on Microsoft Azure that NVIDIA says can cut repeated physical test drives by up to 90% and shorten development cycles by up to 85% — figures from the company’s announcement.

Why It Matters

Hyperion is where NVIDIA’s car business stops being a chip sale and becomes the default stack under other companies’ vehicles. When a ride-hailing operator planning 25,000 robotaxis, three German car brands, a Swedish safety leader and a freight operator all build on the same platform, the platform’s interfaces — its safety evidence format, its reasoning models, its data tools — become the industry’s working standard. That is the same position NVIDIA’s data-centre business reached with CUDA, a pattern GEroundtable traced in why NVIDIA powers the AI boom. For European regulators, Halos SAFE matters just as much: robotaxi approvals turn on evidence, and a shared framework for producing it could set the terms on which level 4 fleets reach public roads. The announcements also extend the physical-AI push behind releases such as NVIDIA’s Cosmos3-H-Dreams world model, while the agent tooling echoes the safety-first framing of the OpenShell reference architecture launched at the same show.

Background

NVIDIA has sold its DRIVE platforms into the car industry for a decade, but the strategy shifted from supplying chips for driver assistance to offering a full robotaxi stack — compute, the Halos safety system and the Alpamayo family of reasoning models that can explain their driving decisions. GTC Berlin, the European edition of NVIDIA’s developer conference, has become the stage where that stack meets the continent’s car industry, and this year’s Hyperion announcement is the broadest European line-up the company has assembled in one release.

Source: NVIDIA via GlobeNewswire, Securities.io

Share to: , and

Faheem R. Khattak
Founder & Senior Editor

Senior automation engineer and AI marketing journo covering LLMs, algorithms, analytics, core updates, Bing search, and technical SEO.

✓ 10+ Years Covering Google Search & AI

0 Comments

Leave a Reply

Optional Take Tag:

🛡️ Spam protection: Honeypot + Time-Trap Active

This site uses Akismet to reduce spam. Learn how your comment data is processed.

    Add as Preferred Source on Google Follow on Google News