Starbase, Texas — The future of artificial intelligence has just touched down where rockets soar. In a moment blending space-age innovation and AI revolution, NVIDIA CEO Jensen Huang personally delivered the first-ever DGX Spark AI supercomputer to Elon Musk at SpaceX’s Starbase, marking a defining moment in the evolution of edge AI computing.
The event wasn’t just a product handoff — it was a symbolic meeting of two technological forces shaping humanity’s future. Amid the hum of SpaceX engineers and the backdrop of the towering Starship, Huang walked through the facility, greeted by grinning engineers, before sharing a casual moment with Musk in the cafeteria. Musk, ever the down-to-earth visionary, was seen chatting with staff and opening snacks for kids before sitting down for pizza with Huang.
“Imagine delivering the smallest supercomputer next to the biggest rocket,” Huang quipped, capturing the poetic contrast of the day — the power of compact AI meeting the grandeur of interplanetary engineering.
A Petaflop Powerhouse the Size of a Book
With the DGX Spark, NVIDIA is redefining what’s possible in the world of AI performance and portability. Barely the size of a hardcover book and weighing just 1.2 kilograms, the DGX Spark packs a full petaflop of AI performance — enough to run models with 200 billion parameters locally.
At its heart lies the NVIDIA GB10 Grace Blackwell Superchip, offering up to 1 petaflop at FP4 precision. This tiny powerhouse brings together 128GB of unified CPU-GPU memory, NVMe storage, and NVIDIA ConnectX networking for seamless performance. It even features HDMI output for instant visualization, making it a true “AI lab in your backpack.”
In Huang’s words, Spark represents “supercomputer-class performance that you can grab and go.”
From the Data Center to the Desktop: A Revolution Beyond the Cloud
Nine years after launching the groundbreaking DGX-1, NVIDIA’s latest innovation pushes the AI frontier beyond massive data centers — straight into the hands of developers, researchers, and creators. The DGX Spark bridges the gap between raw computing power and personal innovation, making petaflop AI accessible anywhere ideas strike — from robotics labs to creative studios.
The supercomputer runs on NVIDIA’s full AI software stack, including frameworks, pretrained models, and NVIDIA NIM microservices, empowering users to:
- Customize image-generation models like FLUX.1
- Build advanced vision search and summarization agents using NVIDIA Cosmos
- Deploy optimized chatbots powered by Qwen3
This isn’t a dev box — it’s a launchpad for human imagination. Whether fine-tuning large language models or designing next-gen robotics systems, DGX Spark makes local, high-performance AI a tangible reality.
Partners Power the Ecosystem: From Palo Alto to Creative Studios Worldwide
The DGX Spark rollout is already sparking waves of excitement across the AI and tech ecosystem. Top technology partners — including Acer, ASUS, Dell Technologies, GIGABYTE, HP, Lenovo, and MSI — are bringing their own Spark-powered systems to market, transforming desktops into AI launchpads.
Each partner is helping deliver NVIDIA’s full AI stack in a compact form, fueling the rise of agentic and physical AI development right at the user’s fingertips.
Across industries, early adopters are already putting Spark to work:
- Ollama (Palo Alto) – Redefining how developers run large language models locally.
- NYU Global Frontier Lab – Prototyping algorithms for privacy-sensitive AI applications.
- Zipline – Expanding the boundaries of autonomous drone delivery.
- Arizona State University – Accelerating robotics simulations and edge AI research.
- Refik Anadol’s Studio – Merging art and artificial intelligence with breathtaking creativity.
Each of these installations marks a new milestone — AI no longer confined to massive data centers but living where creation happens.
When Supercomputing Meets Space Exploration
The DGX Spark’s debut at SpaceX’s Starbase was timed with another monumental moment — the 11th test flight of Starship, the world’s most powerful launch vehicle. It was a fitting parallel: while SpaceX pushes the boundaries of space travel, NVIDIA is expanding the frontiers of AI accessibility.
Together, these two innovations symbolize the next leap for humanity — where interplanetary dreams and artificial intelligence converge.
Availability: Global Rollout Begins October 15
Starting Wednesday, October 15, the NVIDIA DGX Spark will be available globally via NVIDIA.com and authorized partners. As units begin to reach labs, studios, and startups worldwide, NVIDIA is documenting each stop — from California to Europe — as part of its growing “Spark Stories” series, showcasing how creators everywhere are harnessing this portable AI powerhouse.
A Spark That Ignites the Future
From Elon Musk’s SpaceX facility to university robotics labs and AI art studios, the DGX Spark represents a monumental step toward democratizing AI. It’s not just a piece of hardware — it’s a symbol of how far computing has come, and how close it can now be to those shaping tomorrow.
As Jensen Huang said while handing over the device to Musk, “This is where the next AI revolution begins — right where humanity reaches for the stars.”
Read More: Netflix Taps Google’s Varun Sharma to Drive Global Go-to-Market and Ads Growth