Robots may eventually learn millions of hours of experience before ever touching the real world.
That’s what tools like NVIDIA Warp make possible: fast physics simulation on GPUs, at huge scale.
The robot of the future may spend most of its “childhood” inside simulation.
10 million downloads for NVIDIA Warp 🎉
Warp started with a simple idea: you shouldn’t have to leave Python to get real GPU performance for physics and simulation.
Since then, developers have used it to accelerate work across physics simulation, computational engineering, geometry processing and robotics.
Thank you to everyone who downloaded it, broke it, filed an issue or sent a PR. On to the next 10M!
For most of human history, intelligence was one of the scarcest and most expensive things on Earth.
The best doctors, engineers, scientists, lawyers and strategists took decades to train.
Their knowledge lived inside their heads. Their time was limited. And you couldn’t copy them. That is why high-level intelligence carried enormous economic value.
AI is starting to break that scarcity.
We may be only a few years away from extremely capable intelligence being available 24/7, instantly, in millions of copies, at a tiny fraction of today’s cost.
That would be a strange economic reversal: intelligence itself becomes cheap. And when intelligence becomes abundant, the expensive things may shift elsewhere:
> Compute
> Energy
> Unique data
> Distribution
> Physical execution
> Trust
> Taste and judgment
> Human attention
In that world, simply knowing the answer may no longer be special. What matters more is knowing what to ask, what to build, what to trust, and how to turn intelligence into real-world action.
For thousands of years, humanity tried to make people smarter through education, books and the internet. Now we are doing something fundamentally different. We are beginning to manufacture intelligence itself.
And if superintelligence eventually becomes cheap enough for almost anyone to access, the future may not belong to the smartest person.
It may belong to the person who knows what to do when everyone has intelligence.
This is probably how quantum becomes useful in practice.Not as a replacement for GPUs, but as another specialized compute layer tightly connected to AI supercomputers.
The future may be hybrid compute, not one winner.
Quantum systems only become useful when integrated with AI supercomputing, and we're building the platform to make that happen.
Across the world, supercomputing centers are connecting quantum processors with NVIDIA GPUs.
✔️ NVIDIA CUDA-Q lets developers write once and deploy across all major physical quantum processors.
✔️ NVIDIA Ising, our open-source model family, brings AI to quantum by cutting logical error rates by over 300x and compressing qubit calibration timelines from days to hours.
✔️ NVIDIA NVQLink connects quantum processors to GPUs for real-time error correction with microsecond QPU-GPU latency.
Learn more about how we're advancing quantum computing ➡️ https://t.co/UDa7mdlEgC
AI video is quietly moving from “generate me a cool clip” to an actual filmmaking tool. Control the first frame. Control the last frame. Extend the scene. Add references. Upscale to 4K.
The more control creators get, the less this feels like an AI demo and the more it starts feeling like a real production workflow.
This is moving fast.
introducing Gemini Omni 1.1 Flash
this model brings a new suite of creative controls and generative video capabilities to developers
- extend scenes for longer storytelling
- specify first and last frames
- draft videos more efficiently in 360p
- upscale up to 4K resolution
- add video references in your multimodal input
now available via the Gemini API and in AI Studio
This is how I imagine the next generation of technology districts.
AI labs, robotics, clean energy, advanced manufacturing, all built into one ecosystem.