Sit back and enjoy the show. You’re watching computer graphics and simulation powered by NVIDIA CUDA-X, the software helping scientists and engineers put their ideas to the test. From #NVIDIAGTC Berlin:

Oct 30, 2026 · 6:30 PM UTC

132
129
11
1,003
140,137
RelevantRecentLikes
Simulation is the underrated half of AI. Testing an idea in a virtual world before building it changes the economics of engineering.
1
1
3
825
Somewhere a render farm is quietly taking full credit for this.
1
793
CUDA powered sims keep looking better
654
That simulation reel looks unreal 🔥
1
392
Holy shit if only we could get a hold of cards instead of you guys selling out to big Ai
1
174
every sim demo looks effortless right up until you try to walk around inside one
1
567
Amazing seeing how far this technology is coming. Can see this being so useful for product demos and presentations.
135
Science that looks this pretty could make anyone excited about homework 🤣
328
All we need to do is imagine it, And the tools to create it are catching up. Ray Harryhausen’s jaw would drop.
1
474
The simulation visuals in that clip look really smooth
392
Fluid movement looks amazing!
2
420
Reality could never afford this budget.
231
thanks for everything I use Nvidia since 2002 - 2003 to 2026 i growing with it ( Gaming-students-working Now ) (render and simulation) I have family now And it hard to buy new One and i need More of Cuda 🫀 @NVIDIA Can you help now ?
223
Computer graphics with an engineering day job: helping test ideas before anyone has to order the expensive parts.
1
77
visuals look insanely detailed
46
That GTC Berlin clip makes CUDA-X click — an idea turning into a simulation you can watch.
45
man, F Nvidia. What a bunch of clowns.
1
16
Would be useful to see how closely these simulations match physical test results.
1
52
Fuck this company! It’s holding 50 billion in cash, and investors have not made one penny in a year.
58
Incredible simulation quality! NVIDIA CUDA-X keeps pushing the boundaries of what's possible in computer graphics and engineering
92
Thats a lot of heavy lifting 🏋️
14
CUDA 生态的真正护城河不是性能,是十几年攒下来的工具链和踩坑文档。
10
That Berlin clip makes the graphics and simulation work feel a lot more tangible than a spec sheet ever would.
30
Watching graphics and simulation powered by CUDA-X from GTC Berlin is a good reminder that GPUs now serve scientists and engineers as much as gamers. Faster simulation speeds up how ideas get tested. Which field do you think benefits most from this next?
109
CUDA 真正的护城河不是算力,而是让科学家把想法直接变成看得见的现实。
23
图形和仿真原本是两套管线,如今同一套软件栈就能跑。真正的变化在于:研究者不用为了写一个物理模拟先去学图形学,想法到结果的距离被压短了。
30
CUDA-X is turning ideas into reality.
33
视觉上越来越像电影,成本却一年一年往下走。真正的转折不是画面更好看,而是小工作室也能负担得起这种算力。
19
渲染和仿真共用一套算力底座,这才是 CUDA 真正的护城河——不是芯片,是二十年积累的工具链和习惯。
32
Simulation at this fidelity is no longer entertainment — it's where robotics learns physics before touching hardware. The open question: when does synthetic training stop being a complement to reality and start being a substitute for it?
76
仿真做到这个程度,工程师画完图就能先在屏幕上跑一遍失效场景。省下的不只是打样费,是几个月来回试错的时间。
14
仿真这块的价值一直被低估:训练和推理贵在算力,真正省时间的是先在虚拟世界里把方案试错完一遍,再动真机。CUDA-X 这类工具链才是隐性壁垒。
26
Geforce now next upgrade?
39
仿真这条路走得慢,但沉下来的都是别人绕不过去的东西
18