Autor: MK

A car pulls up to the curb. The app says, “Your ride is here.” No one’s in the driver’s seat. For people who live in one of the dozens of cities now hosting robotaxi services, this is already a reality.

The robotaxi industry has moved from prototype milestones to commercial operations, with an expanding ecosystem accelerating the pace of deployment. New collaborations announced at NVIDIA GTC Taipei reflect robotaxi programs spinning up around the world:

• Uber and Autobrains are launching a robotaxi program in Munich on the NVIDIA DRIVE Hyperion platform, using Autobrains’ agentic AI to support scalable operations. 
• Foxconn is expanding its collaboration with NVIDIA to deploy robotaxi fleets, combining its services with NVIDIA DRIVE Hyperion for rapid integration and scaling in Taiwan.
• VinFast is working with Autobrains to bring level 4 vehicles built on DRIVE Hyperion to the Southeast Asia market.
• HUMAIN is working to bring DRIVE Hyperion-powered robotaxis to Saudi Arabia, expanding the platform’s global footprint into the Middle East.

Building a Safe Software Foundation

As the robotaxi industry scales, safety is paramount.

Regulators, certification bodies and developers are scrutinizing what safe deployment at scale requires. 

Industry discussion on level 4 autonomy often centers on what the vehicle can perceive and decide. 

That discussion is well-founded. Accurate perception, sound decision-making and handling the unexpected are difficult problems, and real progress toward solving them is being made.

But perception and decisions alone are not the whole story. Regulators require something more: proof that the overall system behaves reliably, isolates faults before they escalate and never operates outside the boundaries it was designed for. 

Robotaxi safety requires solving four distinct challenges simultaneously:

• A safety-certifiable operating system
• Safe, standardized hardware and software interfaces
• AI that operates within verifiable guardrails
• Validation at scale before vehicles touch public roads

To help solve these challenges, the recently introduced Halos Operating System (OS) — a component of the NVIDIA Halos full-stack, comprehensive safety system — offers a unified, production-ready safety foundation for AI-driven vehicles, built on NVIDIA DRIVE Hyperion. It comprises: 

Halos Core: A Certified OS Foundation

At the foundation of NVIDIA Halos OS is Halos Core, which is the next generation of NVIDIA DriveOS and certified to automotive safety standards. It’s audited, documented and proven to behave predictably under fault conditions, with a hypervisor — a specialized software layer — that isolates safety-critical functions so failures can’t reach vehicle controls. 

Halos Core is compliant with ISO 26262 ASIL D, includes safety-certified support for NVIDIA CUDA and TensorRT, and provides the TensorRT Edge-LLM open source framework for high-performance large language model inference.

Halos SDK: Standardized and Safe Interfaces

A robotaxi integrates cameras, radar, lidar and other sensors, each streaming data in a different format at a different rate. Without a standardized middleware layer, every hardware change forces teams to manually rebuild those integrations. 

Halos SDK removes that burden. Its sensor abstraction layer decouples the autonomous driving stack from individual sensor drivers, so adding or swapping a sensor no longer causes ripples through application code, while a vehicle abstraction layer connects the autonomous driving stack to the rest of the vehicle through a single, consistent interface. 

On top, Halos SDK provides the runtime building blocks that safety-critical software demands: a deterministic application-level scheduler for predictable timing, zero-copy inter-process communication that moves data without added latency, a comprehensive system error-handling framework and a robust scenario data recorder — delivering the foundation for highly reliable and low-latency automotive applications.      

Halos Applications: Safety Guardrails for AI

AI models can match human driving behavior, but regulators require more than performance. 

The Halos Applications layer provides safety guardrails for AI through deterministic, rule-based functions, analyzed and designed to behave within defined bounds. It includes world model perception and the top-rated NVIDIA DRIVE active safety stack featuring automatic emergency braking, lane departure warning, blind spot monitoring, collision warning and more. 

In addition, in Halos Applications, Halos OS can be combined with end-to-end AI models for which explainability and transparency are essential. This includes the NVIDIA Alpamayo family of open models for autonomous vehicle development, which enables chain-of-thought reasoning, continuously evaluating the road, planning next steps and adapting to changing conditions.

The Halos Safety Evaluation Framework

Halos Infra is the cloud-side development infrastructure that enables autonomous vehicle training, simulation and validation at scale. It’s the foundation for the recently released NVIDIA Halos Safety Evaluation Framework (SEF).

SEF provides the tools and guidelines needed to build a credible safety case, from L2 driver assistance to L4 robotaxis. It draws on more than 330 research papers and 1,000 patents developed within NVIDIA Halos OS.

Halos Infra runs on NVIDIA’s three-computer autonomous driving solution: 

• NVIDIA DGX systems for training the AI stack in the data center
• NVIDIA Omniverse on NVIDIA OVX systems for simulation and synthetic data generation
• The NVIDIA AGX in-vehicle computer for real-time sensor processing and safety

Halos OS spans the full development lifecycle — from training and simulation in Halos Infra to inference in the vehicle itself.

Learn more about NVIDIA Halos.

Today, Google DeepMind released DiffusionGemma — an experimental open model built for exceptionally fast text generation. NVIDIA has optimized DiffusionGemma to run even faster across NVIDIA GeForce RTX GPUs, the NVIDIA RTX PRO platform and NVIDIA DGX Spark systems, from local PCs to the cloud. 

Rather than generating text one word at a time, DiffusionGemma generates multiple words in parallel to output whole blocks of text, opening a new, low-latency frontier for the kind of single-user workloads that developers, researchers and AI enthusiasts run every day. 

Features of the new model include: 

• Parallel generation: DiffusionGemma denoises up to 256 tokens per step instead of predicting one at a time. 
• Built on Gemma 4: DiffusionGemma is built on Gemma 4, a 26-billion-parameter mixture-of-experts model that activates just 3.8 billion parameters per step, pairing a diffusion head with Google’s Gemma 4 architecture. 
• Up to 4x faster performance: The boost means fast text generation, where single-user generation usually stalls — at batch size 1, on local hardware. 
• Open and local: DiffusionGemma is open weights under a permissive Apache 2.0 license and runs entirely on RTX and DGX Spark — no cloud, no per-token cost — with day-zero support in Hugging Face Transformers, vLLM and Unsloth. 

A Different Way to Generate Text 

Almost every large language model (LLM) in wide use today is autoregressive — meaning it generates text one token at a time, with each new word depending on the one before it. That sequential process is what makes interactive AI feel like it’s typing. 

DiffusionGemma takes a different path. Built on the Gemma 4 26B mixture-of-experts architecture, it generates text the way diffusion models generate images: by starting from noise and refining a whole block of text at once. Each step denoises up to 256 tokens in parallel rather than emitting a single token and waiting to compute the next. 

The result is a model that thinks in blocks instead of sequentially. For latency-sensitive, single-user work — such as interactive chat, agentic loops or on-device assistants that plan and act — that parallelism translates into responses fast enough to keep pace with how developers think and iterate.

DiffusionGemma Flies on NVIDIA GPUs 

Generating one token at a time is fundamentally a memory-bound problem — at batch size 1, a traditional LLM spends most of its time waiting on memory bandwidth, not doing math, which leaves a lot of compute on the table. 

Diffusion flips the equation. Pulling a full 256-token block through the transformer in parallel is a compute-bound workload — exactly what NVIDIA GPUs are built for. NVIDIA Tensor Cores accelerate the dense parallel math, and the CUDA software stack lets the model run efficiently from day one without bespoke tuning. In short, the model’s design plays directly to the GPU’‘s strengths. 

That shows up in the numbers. DiffusionGemma delivers 1,000 tokens/sec at batch size 1 on a single NVIDIA H100 Tensor Core GPU, 150 tokens/sec on NVIDIA DGX Spark and fastest local inference on NVIDIA DGX Station — roughly 4x faster than an equivalent autoregressive model running in the same single-user regime. 

That advantage holds across NVIDIA’s full lineup, running: 

• Locally on the NVIDIA DGX Spark deskside personal AI supercomputer — powered by the NVIDIA GB10 Grace Blackwell Superchip with 128GB of unified memory — with the preinstalled NVIDIA AI software stack ready for prototyping, fine-tuning and fully local agent workflows. 
• On NVIDIA RTX PRO 6000 workstations, providing developers, researchers and AI professionals with the headroom to run local low-latency generation and agentic loops as part of a professional workflow. 
• On DGX Station, delivering best-in-class, high-speed inference at up to 800 tokens/sec for low-latency text generation and agentic loops with 748GB of coherent memory. 
• On GeForce RTX GPUs, with llama.cpp support coming soon. 

Get Started Locally

The fastest way to start testing and prototyping the model is through Hugging Face Transformers, which runs DiffusionGemma on a GeForce RTX 5090 or DGX Spark out of the box. For higher-throughput inference, vLLM provides day-zero serving support.  

For adapting the model to a specific task or domain, fine-tuning is available through Unsloth and NVIDIA NeMo framework, with ready-made DGX Spark playbooks to get a local environment running quickly. Check out the vLLM playbooks for DGX Spark , RTX PRO and DGX Station. 

Try Diffusion Gemma on Hugging Face or test it for free using NVIDIA-hosted application programming interfaces at build.nvidia.com. 

Go deeper on the architecture and local deployment by reading the NVIDIA technical blog and the Google DeepMind announcement.

#ICYMI: The Latest From RTX AI Garage 

 NVIDIA researchers released SANA-WM, an open source world model that turns a single image and a camera path into a minute-long, 720p video with precise 6-DoF control. At just 2.6 billion parameters, its distilled version generates a full 60-second clip in 34 seconds on a single NVIDIA GeForce RTX 5090 GPU using the NVFP4 format — delivering up to 36x higher throughput than comparable open models while running on one GPU. Read the paper. 

 Building Windows agents just got a full toolset — NVIDIA and Microsoft rolled out turnkey agent sandboxing on native Windows — Microsoft eXecution Containers plus the NVIDIA OpenShell runtime — alongside up to 2x faster agentic inference and native Windows support for Hermes Agent. 

DGX Spark goes from unboxing to a running agent in minutes — A streamlined NVIDIA NemoClaw install gets developers to a working local agent fast, with Qwen3.6-35B running up to 2.6x faster on vLLM. And the new cluster assistant in NVIDIA Sync links up to four DGX Spark units into one 512GB pool — enough for ~400-billion-parameter models. 

Plug in to RTX Spark on Facebook, Instagram, TikTok and X — and stay informed by subscribing to the RTX Spark newsletter. 

See notice regarding software product information.

NVIDIA GPUs with Confidential Computing are now used for confidential inference in Apple’s Private Cloud Compute (PCC), as it expands beyond Apple’s data centers to Google Cloud. 

Unveiled during Apple’s annual WWDC gathering for developers from around the globe, NVIDIA GPUs will support server-side inference for Apple Foundation Models, custom-built by Apple and Google, leveraging the technologies behind the Gemini family of models.

NVIDIA is collaborating with Apple and Google to support some of the next-generation Apple Intelligence features, using NVIDIA Blackwell GPUs with Confidential Computing integrated into Private Cloud Compute’s hardware security architecture running on Google Cloud.

Confidential Computing Matters for the Era of AI Experiences 

NVIDIA Confidential Computing provides a hardware-based security layer for accelerated AI workloads. The technology protects data while it’s being processed by isolating workloads in trusted execution environments and enabling systems to cryptographically verify that the infrastructure has not been tampered with before any sensitive data is sent to the server. 

For end users, NVIDIA Confidential Computing means that no one, not even the system’s builders, can look at their data, chats or conversations.

Adoption of NVIDIA Confidential Computing at this scale reflects a broader shift in AI infrastructure: As AI experiences combine on-device and cloud-based processing for their tasks, there’s a need for high-performance, server-side inference while maintaining strong privacy and security guarantees. 

How Confidential Computing Enforces Privacy and Trust

NVIDIA Confidential Computing reflects NVIDIA’s commitment to trustworthy AI and includes these key capabilities:

• Hardware-rooted trust, helping establish that systems are running on genuine, untampered NVIDIA GPUs.
• Encrypted communication paths, helping protect data as it moves between components.
• Remote attestation, enabling software to verify the security state of the platform before releasing sensitive data.
• Support for accelerated AI inference and training, helping organizations run privacy-sensitive workloads without moving away from GPU performance.

These capabilities are increasingly relevant for AI services that need to process sensitive information while maintaining strong user privacy controls.

Learn more about NVIDIA Confidential Computing and NVIDIA AI cybersecurity solutions. 

A year ago at London Tech Week, NVIDIA founder and CEO Jensen Huang and U.K. Prime Minister Keir Starmer made a declaration: the U.K. would be an AI maker, not an AI taker. 

At this year’s event, NVIDIA and its partners are showcasing how that commitment is producing real momentum across the nation’s infrastructure, startups and enterprises. 

U.K. technology leaders are innovating across healthcare and life sciences, coding, agentic AI, inference and more — all running on sovereign AI deployments.

Commitment to Compute

Over the past year, the number of AI cloud providers planning to deploy AI infrastructure on U.K. soil has doubled. 

Nebius has announced plans to expand customers and cloud capabilities with three new deployments of advanced NVIDIA AI infrastructure, as the NVIDIA AI Cloud ecosystem partner continues to build out its commercial and AI R&D hub in London. Combined, the deployments are expected to reach 65 megawatts when fully ramped up in 2027.

CoreWeave is building in the U.K. Government’s AI Growth Zones, and seven more NVIDIA AI Cloud ecosystem partners have plans in the pipeline. BT and Nscale announced plans to build sovereign AI data centers across three existing BT sites in the U.K., combining NVIDIA AI infrastructure, Nscale’s full stack and BT’s trusted nationwide connectivity backbone. 

From Fund to Frontier

Central to that sovereign compute story is Isambard-AI — the U.K.’s most powerful computer. Built on 5,400 NVIDIA GH200 Grace Hopper Superchips and running entirely on zero-carbon electricity, it’s the engine behind some of the U.K.’s most ambitious AI research. 

The U.K. government’s Sovereign AI Fund is putting that capability to work by backing homegrown companies and providing the domestic infrastructure needed to scale their ambitions. 

Among its first recipients is Ineffable Intelligence, which recently announced a collaboration with NVIDIA to build the future of reinforcement learning infrastructure. 

Other recipients include four U.K.-based NVIDIA Inception startups, each pushing the AI frontier using Isambard-AI. These startups are:

Cosine Builds Sovereign Coding Platform

Cosine is building an end-to-end sovereign AI coding platform for highly regulated industries such as financial services, critical infrastructure and national security. Using Isambard, Cosine is training a new, large-parameter, mixture-of-experts, multimodal agentic LLM for natively handling data types beyond text and image. 

“Access to Isambard enables the project, full stop,” said Alistair Pullen, cofounder and CEO of Cosine. “We already have the people who know how to do this. We have the data. We have the infrastructure and the training. The thing we’ve never had is this level of compute.”

Cursive Trains Self-Improving AI Systems

Cursive is building self-improving AI systems that learn continuously from real-world data, enabling them to operate autonomously over long periods of time. This is unlocked through new memory-augmented architectures with dramatically larger context windows, currently in development using the Sovereign AI Fund resources. In addition, the team recently adopted the NVIDIA Megatron-LM framework for distributed training at scale.

“The Sovereign AI Fund is more than just processing power — it’s a statement about investing in AI in the U.K.,” said Talfan Evans, cofounder and CEO of Cursive. “Sovereignty is actually now a buying criterion — and it’s a challenge to tap into the resources we uniquely have as U.K. and European companies.”

Doubleword Optimizes Inference to Deliver Abundant Intelligence Tokens

Doubleword, the U.K.’s first dedicated inference lab, optimizes every layer of the AI stack to maximize what it calls “IQ per dollar.” The company deploys open models including NVIDIA Nemotron 3 Super 120B and builds on the NVIDIA Dynamo inference framework. 

On Isambard, Doubleword’s early results achieved 70x faster model cold starts — aka model loading times — and 4x lossless KV cache compression, critical advancements for long-running agentic workloads. The result: inference at 90-95% lower costs than other leading inference providers.

“Sovereign AI is most impactful at the inference layer,” said Meryem Arik, cofounder and CEO of Doubleword. “Inference is when you’re actually getting the value from the model — we want that value created in the U.K., with U.K. compute and U.K. data centers.”

Prima Mente Uses Foundation Models to Study Alzheimer’s and More

Prima Mente builds biological foundation models to identify new biomarkers, subtypes and drug targets of Alzheimer’s, Parkinson’s and ALS. With its Isambard allocation, the company is developing Pleiades 2, a foundation model combining five biological data modalities. 

Achieving nearly 3x speedups in model training with NVIDIA Blackwell GPUs, Prima Mente also uses NVIDIA Parabricks for genomic data processing and NVIDIA Transformer Engine for model optimization.

“Research shows Alzheimer’s might be 25 different subgroups of disease, and we want to help by using AI to identify these subtypes and the biology within the cells as they change,” said Hannah Madan, cofounder of Prima Mente.

Video courtesy of Nebius and Prima Mente.

AI Talent, Policy and Production

NVIDIA’s £2 billion investment in the U.K. startup ecosystem — in collaboration with leading venture capital firms — is bringing new capital and advanced AI infrastructure to major U.K. hubs including London, Oxford, Cambridge and Manchester. 

U.K. membership in the NVIDIA Inception program has increased by 50% over the past year. AI-native companies like Doubleword, Synthesia and PolyAI are scaling globally from U.K. roots. 

At last year’s London Tech Week, NVIDIA announced a collaboration with the U.K Department for Science, Innovation and Technology on 6G and AI skills. The 6G collaboration has seeded testbeds at four U.K. universities. In May, the NVIDIA Deep Learning Institute (DLI) delivered two new courses — added to support the nation’s wireless research community — to participants from over 30 U.K. universities.

Plus, as part of this AI skills collaboration, NVIDIA DLI courses are offered as part of QA’s AI Apprenticeships in England. 

And the NVIDIA Developer Program now includes more than 200,000 U.K. developers. 

The Sovereign AI Forum, which launched last year with seven charter members, convened the country’s AI leadership to turn policy into deployment roadmaps. Over the past year, the Forum has welcomed dozens of participants across government, industry and the startup community — turning policy into deployment roadmaps.

And enterprise AI is moving from pilot to production:

• Apian is building digital twins of two National Health Service hospitals, combining autonomous devices, ground robots, computer vision and robotic simulation.
• Deliverance AI is helping regulated enterprises to run, govern and scale AI agents inside their own environment — through a single control plane. The Agentic Operating System is built for organizations where data sovereignty is non-negotiable.
• Orbital Industries has announced codesigned, NVIDIA Vera Rubin DSX AI Factory-compliant AI infrastructure that accelerates time to first token.
• Reading Football Club is partnering with Stelia to establish an AI Centre of Excellence, combining Stelia’s full-stack AI platform with accelerated compute infrastructure from NVIDIA and Lenovo.

It all reflects momentous progress in U.K. AI leadership — and offers a glimpse of where it’s heading.

Join NVIDIA at London Tech Week.

NVIDIA and LG Group are building an AI factory to accelerate LG Group’s next wave of AI-driven businesses, spanning robotics, autonomous driving, data center technologies and GPU cloud services.

The AI factory will provide LG Group with accelerated computing infrastructure to train, simulate, validate and deploy AI-based applications across its key businesses. 

The collaboration brings together NVIDIA’s full-stack, end-to-end AI factory platform with LG Group’s global leadership in consumer electronics, robotics, mobility components, smart spaces and data center technologies.

Together, the companies are connecting AI model development, physical AI data generation, robot simulation and training, edge deployment and factory-scale digital twins into a unified workflow for building physical AI systems. 

Advancing Physical AI and Robotics

The combination of LG’s production technology data and know-how from global manufacturing sites with NVIDIA’s AI infrastructure and digital twin technologies will help enhance AI-driven manufacturing AI competitiveness. The two companies will collaborate to build an autonomous manufacturing ecosystem in which the entire process — from raw material procurement to production, logistics and customer delivery — is connected in real time through data and AI, and establish it as a new global smart factory standard.

LG Electronics is developing home-based robots like CLoiD to help with a wide range of indoor household tasks, enhancing everyday convenience and improving quality of life. 

By integrating the NVIDIA Isaac Sim and NVIDIA Isaac Lab open robotics frameworks into their development workflows, LG can simulate, train and validate these home cobots in physically accurate virtual environments before deployment. 

The company is exploring using the NVIDIA Isaac GR00T open, reasoning vision action language model for both its home robots and modular robotics platforms. The GR00T model will provide LG robots humanlike reasoning and the ability to execute complex tasks. NVIDIA and LG Electronics also plan to jointly develop reference robots, positioning LG’s robots as part of the NVIDIA Isaac GR00T ecosystem.

To help overcome the training data challenge for robotics, LG Electronics is developing a physical AI data factory poised to help Korean and global companies accelerate physical AI projects. By turning compute into data, LG will be providing high-quality training data for robotics and industrial AI projects, using NVIDIA Cosmos world foundation models for synthetic data generation and augmentation.

LG Innotek, harnessing its world-class optical expertise, plans to provide state-of-the-art robotics components, including sensing solutions, specifically optimized for NVIDIA’s development environments and GPU architecture.

LG CNS is building an ecosystem that enables anyone to easily adopt AI robots in manufacturing and logistics sites. By integrating NVIDIA’s robotics technologies including Isaac open robotics frameworks, NVIDIA Cosmos open world models and Isaac GR00T robotic foundation models into its PhysicalWorks industrial robot platform, the company is accelerating the AI transformation of logistics and manufacturing floors.

Building an NVIDIA DSX-Aligned AI Factory Infrastructure 

The two companies will also expand cooperation in the field of next-generation AI factories, which will support the AI era.

Beyond its certification cooperation with NVIDIA on cooling solutions for AI factory thermal management — including cooling distribution units (CDUs) and cold plates — LG Electronics is further elevating its AI factory capabilities through technical collaboration on prefabricated modular design technologies. This initiative aligns with the NVIDIA DSX AI factory platform, enabling the rapid deployment of scalable, high-performance supercomputing infrastructure.

These technologies include CDUs, cold plates and prefab modular design capabilities to help address the power, thermal and deployment requirements of next-generation liquid-cooled AI factories.

In collaboration with LG Electronics and LG Energy Solution, LG Uplus — a telecommunications provider under LG Corp. — plans to build scalable, power-efficient AI factories based on NVIDIA DSX. The effort is expected to combine NVIDIA accelerated computing and AI factory reference architectures with LG’s infrastructure, energy and telecommunications capabilities to support future AI cloud and GPU service opportunities. 

LG CNS plans to build scalable, power-efficient, high-performance AI factories powered by NVIDIA GPUs based on NVIDIA DSX.

LG Uplus plans to build a large-scale AI data center capable of accommodating the latest NVIDIA GPUs.

LG Energy Solution plans to collaborate with NVIDIA on emerging 800 volt-direct-current data center energy solutions, in alignment with NVIDIA’s BESS Self-Qualification guidelines, to keep pace with next-generation GPUs. 

Accelerating Autonomous Driving and Mobility AI

In mobility, LG Electronics works with NVIDIA to align its advanced driver-assistance systems (ADAS) and in-vehicle AI systems with the NVIDIA DRIVE platform. 

The collaboration will focus on aligning sensor, compute and software architectures with the NVIDIA DRIVE Hyperion architecture, supporting LG Electronics’ roadmap for autonomous driving, ADAS and software-defined vehicles. 

LG Electronics also plans to use NVIDIA DRIVE AGX accelerated compute for its future mobility applications, including AI-powered cockpits and edge AI processing. Through this work, LG Electronics aims to strengthen its automotive electronics portfolio and accelerate the development of AI-driven mobility solutions for global manufacturers.

LG Innotek is rapidly cementing its leadership in the autonomous driving market, using its core portfolio of world-class sensing, connectivity and lighting solutions. LG Innotek plans to collaborate with NVIDIA on next-generation components engineered specifically for NVIDIA architecture. 

Advancing Sovereign AI With EXAONE

NVIDIA and LG AI Research are collaborating to advance EXAONE, one of Korea’s leading sovereign AI models and an open model family available to developers, enterprises and researchers. 

LG AI Research used NVIDIA Blackwell GPUs, NVIDIA NeMo framework and NVIDIA Nemotron open datasets to support EXAONE model development, as well as NVIDIA TensorRT-LLM software to build high-performance inference engines for optimized deployment.

LG Group is exploring broader adoption of EXAONE and agentic AI technologies across its businesses through platforms such as ChatEXAONE — LG Group’s EXAONE-based enterprise chatbot service. NVIDIA will help power LG AI Research’s sovereign AI models, so LG Group can accelerate enterprise AI transformation, software-defined operations and productivity across its business portfolio. 

Learn more about the NVIDIA DSX platform.

(Image courtesy of LG Group)

NVIDIA and Doosan Group are expanding their collaboration to advance new opportunities across physical AI, robotics and AI factory infrastructure, spanning Doosan Robotics, Doosan Bobcat, Doosan Enerbility and Doosan Corporation Electro-Materials BG.

The collaboration will bring together NVIDIA’s full-stack accelerated computing platforms with Doosan Group’s capabilities in industrial automation, power generation and advanced electronics materials to support next-generation AI infrastructure.

Doosan Group’s businesses span several layers of the AI factory ecosystem, from intelligent robotics systems to the full spectrum of large-scale power solutions and advanced electronics materials for AI data center equipment. 

NVIDIA and Doosan will explore how NVIDIA’s physical AI stack, NVIDIA DSX AI factory platform, NVIDIA MGX and accelerated computing platforms can support these areas.

Advancing Physical AI and Robotics

Doosan Robotics is integrating NVIDIA Isaac Sim and NVIDIA Isaac Lab open robotics frameworks, NVIDIA Cosmos open world foundation models, the open source Newton physics engine and NVIDIA Jetson Thor to advance its Agentic Robot OS — an AI-powered platform connecting perception, reasoning, simulation, learning and on-device inference. 

By integrating NVIDIA’s physical AI technologies, Doosan Robotics aims to help industrial robots better perceive, reason and act in complex and dynamic environments. Simulation-to-real workflows, physics calibration and AI reasoning will make collaborative robots more adaptable, task-specialized and ready for scalable deployment. 

The companies are also looking to develop reference use cases for high-value industrial tasks such as depalletizing and sanding, as well as new robot form factors including dual-arm and humanoid platforms.

Built on Agentic Robot OS, these capabilities aim to help Doosan Robotics evolve from a robot arm provider into a full-stack AI-first robotics solution company. The work is part of a broader, Doosan Group-wide direction for physical AI that extends beyond robotics into areas such as construction machinery and power equipment.

Doosan Bobcat also plans to explore integrating NVIDIA physical AI technologies into equipment used across construction, landscaping, agriculture and material handling applications. This work will help accelerate the development of specialized world models that enable Doosan Bobcat’s equipment to perceive diverse operating environments, reason about changing conditions and perform tasks more autonomously. The companies also aim to help establish an industry-standard ecosystem for compact autonomous equipment.

Exploring AI Factory Power Solutions

Doosan Enerbility is exploring opportunities to support NVIDIA AI factories and the NVIDIA DSX AI factory platform through its large-scale power infrastructure portfolio, including gas turbines, steam turbines and small modular reactors, together with Doosan Fuel Cell’s hydrogen fuel-cell systems. These technologies are relevant to AI data centers that require reliable, high efficiency and continuously available power.

Future collaboration could include power supply design for AI factory deployments, optimization of generation equipment and evaluation of low-carbon power sources such as small modular reactors. By aligning AI infrastructure requirements with energy system expertise, Doosan Enerbility could help address the growing power demands of accelerated computing.

Supporting the NVIDIA MGX Ecosystem With Advanced PCB Materials

Doosan Corporation Electro-Materials BG is supporting next-generation AI data center infrastructure through copper clad laminate, or CCL, a key foundational material for printed circuit boards. 

High-performance CCLs are used in printed circuit boards (PCBs) for networking equipment, AI accelerators and AI server motherboards, where low signal loss and high reliability are critical.

NVIDIA MGX provides a modular reference architecture for accelerated systems, helping system manufacturers and ecosystem partners build servers and rack-scale AI factory infrastructure. As AI servers and networking systems increase in performance and bandwidth, advanced PCB materials such as CCL can play an important role in enabling high-speed signal integrity across the data center equipment ecosystem.

Learn more about NVIDIA DSX and MGX.

Featured image courtesy of Doosan Group.

At GTC Taipei at COMPUTEX last week, NVIDIA unveiled RTX Spark, the superchip that reinvents Windows PCs for the era of personal AI agents. On the heels of this announcement, NVIDIA founder and CEO Jensen Huang headed to South Korea, where he introduced RTX Spark to the nation’s passionate gaming community.

Leading game developers — including Korea’s KRAFTON and NC — are already working to bring their titles to RTX Spark-powered systems. 

Designed for local AI, creating and gaming, RTX Spark brings together 30 years of NVIDIA innovation to slim Windows laptops with all-day battery life and small, ultraefficient desktop PCs. 

With the superchip, gamers can play AAA games at 1440p resolution and over 100 frames per second with NVIDIA ray tracing, DLSS and Reflex technologies. In addition, RTX Spark supports all NVIDIA RTX technologies, including the recently announced DLSS 4.5 Ray Reconstruction, which features a second-generation transformer model for realistic image quality. 

RTX Spark Ignites Korea’s Gaming Community

Korea has played a major role in spearheading esports and driving the boom in PC bangs, or internet and gaming cafes. With longstanding collaborations rooted in the country, NVIDIA in October celebrated 25 years of GeForce in Korea with a free festival for gamers, highlighting the rich gaming ecosystem that has been built over decades.   

On Friday, Huang headed to T1 Base Camp — a PC bang owned by T1, one of Korea’s top esports teams. There, he met with T1’s reigning League of Legends World Champion team, including six-time World Champion Lee “Faker” Sang-hyeok to unveil RTX Spark. 

NVIDIA and Riot Games — developer of League of Legends — are collaborating to bring the title as well as VALORANT to RTX Spark, expanding gamers’ access to high-performance gaming on slim laptops. 

To mark the occasion, T1 Base Camp attendees had the chance to win RTX Spark laptops, League of Legends and T1 merch signed by Huang and Faker, as well as GeForce RTX 5090 GPUs. 

Surprising PC-Bang Gamers 

Later, Huang headed to Seoul’s Gangnam district, where he surprised PC-bang gamers with a first look at RTX Spark with KRAFTON and NC. 

At the first stop, Optimum Zone PC, Huang and KRAFTON Chairman Byung-gyu Chang showcased PUBG: BATTLEGROUNDS and Subnautica 2 on RTX Spark to a captivated crowd of gamers. 

Gamers then got the surprise chance to play with the unreleased PUBG Ally, a co-playable character built with NVIDIA ACE technologies on RTX Spark laptops. PUBG Ally resulted from AI research and development at KRAFTON and NVIDIA, part of an initiative to create next-generation game characters that act like teammates and enable more meaningful, immersive engagements with players.

Next, Huang stopped at another PC bang, Portal PC, where he showcased NC’s CINDER CITY and AION 2 on RTX Spark, with support from Taekjin Kim, co-CEO of NC. 

NC and NVIDIA began working together in the early 2000s on the Lineage franchise and have since collaborated to integrate RTX technology into many of NC’s flagship games, including Lineage 2, AION, Blade & Soul, AION 2 and CINDER CITY. 

Gamers at Portal PC were given the chance to play a demo of NC’s highly anticipated open-world massively multiplayer online tactical shooter CINDER CITY on GeForce RTX-powered PCs.  

CINDER CITY will support the DLSS 4.5 Dynamic Multi Frame Generation and Super Resolution features at launch. Plus, gamers will be able to experience the title on slim RTX Spark laptops and compact desktops when the game is released later this year. 

In addition to KRAFTON, NC, and Riot Games, 100+ Windows software providers and game developers are embracing RTX Spark. These partners include NetEase, Remedy Entertainment and XBOX. 

Learn more about RTX Spark and its launch partners.

Home to cutting-edge sovereign AI infrastructure and robotics innovators, as well as one of the world’s most passionate gaming communities, South Korea is one of the world’s centers of AI. NVIDIA founder and CEO Jensen Huang is in Seoul this week to meet the partners and builders behind that work.

Stay tuned here for live updates.

Thursday, June 4, 10:30 p.m. PT

Touchdown in Seoul

On the heels of GTC Taipei at COMPUTEX, NVIDIA founder and CEO Jensen Huang touched down in Seoul Friday afternoon, greeted by fans and media as his visit got underway.

A key focus of the trip, Huang said: to align the AI supply chain ahead of a busy second half of the year.

“We have a very significant, very large AI infrastructure buildout — already a very successful first half,” Huang told media. “Grace Blackwell, our system, is doing very well, and Vera Rubin is in full production — so we are going to be very busy the second half [of the year].”

Huang also touched on the huge potential for robotics and physical AI in Korea.

“Robotics is going to be the next major sector here in Korea — this is a great opportunity for Korea to invest in AI,” he said.

From memory manufacturing to robotics and gaming, Huang is off to a packed schedule with partners — but not without leaving time to enjoy some Korean fried chicken and BBQ. “It’s all delicious,” Huang said.