NVIDIA AI Infrastructure
@NVIDIAAIInfra
AI factories for the era of AI reasoning.
1.7K Following 81.1K Followers
Ten years ago at NVIDIA GTC 2016, our CEO @JensenHuang introduced DGX-1, our first AI supercomputer.
Since then, NVIDIA DGX has transformed from a single system into a blueprint for AI factories. What hasn’t changed is the reason why we build DGX: to test the systems and work through the details of interconnects, cooling, and software so our partners and customers have a proven foundation to build on.
Charlie Boyle, our VP of DGX Systems, looks back at how DGX got its start, what’s changed, and what comes next.
Thanks to everyone who’s been on this journey with us. #DecadeOfDGX#
Watch the full video:
Show more
With multi-generational CUDA compatibility, continuous software optimization, and CUDA-X support for diverse workloads, NVIDIA GPUs keep delivering value for years after deployment.
Scale-in is the networking layer that connects the AI factory to the outside world.
Gilad Shainer, our SVP of Networking, breaks down how scale-in brings users and models into the AI factory while enforcing security across the entire compute and storage stack.
Read the full analysis with @sdxcentral ➡️
Show more
Older NVIDIA GPUs can remain productive and revenue-generating long after newer generations arrive.
@OrnnExchange's analysis of NVIDIA A100 rental pricing offers third-party evidence of a multi-year earning life for prior-generation NVIDIA infrastructure.
CUDA’s support for diverse workloads, combined with continuous software optimization across generations, helps keep that hardware productive longer.
Learn more:
Show more
New third-party benchmark results for NVIDIA Vera CPU are in!
@duckdb put the Vera CPU to the test and measured 1.5x higher TPC-H performance compared to a leading x86 CPU.
Check out the results ⤵️
Show more
🤝 @OpenAI moved Astra's code, optimized for NVIDIA Blackwell, to NVIDIA Vera Rubin, delivering 3x higher throughput out of the box.
In just 72 hours, AI optimization unlocked an additional 2x gain.
Together, the right hardware and tooling can compress weeks or months of optimization into a single weekend.
Watch the full conversation with @business’s Dina Bass at AI Infra Summit 2026 ➡️
Show more
AI factory efficiency takes a full-system approach.
The NVIDIA DSX AI factory platform brings compute, power, cooling, and grid flexibility together to make the entire AI factory more efficient and generate greater AI output from available energy.
Show more
More productive AI capacity from the power already available.
In a proof of concept, @LambdaAPI used NVIDIA DSX MaxLPS to run 19 nodes within the power budget of a 16-node baseline, delivering 24% more token throughput and 23% higher performance per watt.
Read the success story ➡️
Show more
@Pinterest Learn more about our collaboration with Pinterest and @vllm_project:
Our compute is designed for the long run. @OrnnExchange breaks down the data:
🤝 @dMatrix_AI is adopting NVIDIA NVLink Fusion to connect its next-generation Raptor XPUs with NVIDIA AI infrastructure, joining a growing ecosystem of partners using our rack-scale architecture for AI factory-scale deployment.
By bringing together NVIDIA NVLink scale-up, Spectrum-X scale-out networking and the NVIDIA MGX rack architecture, d-Matrix can build specialized inference systems on a proven, rack-scale foundation spanning compute, racks, networking and software.
Learn more:
Show more
What does it take to deliver high-quality results in seconds at massive scale?
@perplexity_ai serves 50M+ queries daily across 15 models, including NVIDIA Nemotron 3 Ultra. That takes high-performance infrastructure and smarter routing without compromising quality, speed, or cost.
That’s why Perplexity has built on NVIDIA generation after generation. With NVIDIA Blackwell, the larger NVLink domain enables models to scale across multiple nodes without sacrificing latency.
Learn more:
Show more
We're heading to #SEMICONWest# 2026!
Join us to see how we're enabling the semiconductor ecosystem to accelerate innovation across chip design, manufacturing, and production.
Learn how accelerated computing and AI technologies are powering smart factories and scaling semiconductor workloads. @SEMIconex
📆 October 13-15
📍San Francisco, CA
🔗 Learn more and register:
Show more
Congratulations to @OpenAI on GPT-6 Astra! 🌌
Its most intelligent and aligned model to date, trained and deployed using NVIDIA AI infrastructure to advance computer use, agentic coding, professional work, scientific reasoning and cybersecurity.
Excited to see what developers, researchers and enterprises build with Astra.
Show more
Thank you to everyone who joined us at @hotchipsorg 2026!
While there, Dion Harris broke down our announcements and explained how the NVIDIA Vera Rubin platform is purpose-built for the demands of long-context inference and agentic AI.
Show more
AI factories are becoming a new class of infrastructure. 🏭
Join our CTO Michael Kagan at #SEMICONTaiwan# 2026 on September 2 as he explores how our extreme co-design, from compute and networking to systems and software, enables intelligence at industrial scale.
Learn more:
Show more
Announcing the expansion of NVIDIA NVLink Fusion with NVHBM, a next-generation high-bandwidth memory technology that brings higher memory performance and efficiency to XPUs.
Amazon's @AnnapurnaLabs will be the first to work with us on NVHBM, combining @awscloud custom silicon with our memory technology and the NVLink scale-up architecture to enhance performance and efficiency for AI workloads.
Learn how we're helping hyperscalers and AI innovators build the next generation of AI infrastructure:
Show more
Delivery day at our Microsoft DCs as the first production Vera Rubins arrive. A huge thank you to our partners at @nvidia and our Azure hardware and datacenter teams for all the incredible work that brought us to this milestone!
Show more
Agentic AI demands the right infrastructure for every workload.
@UiPath built on @googlecloud AI Hypercomputer, using NVIDIA Hopper GPUs for training and NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs for inference, to match the right GPU to the right job at scale.
The result: 99.5% accuracy on 100M+ transactions for Omega Healthcare, and 70% faster invoice processing for Thermo Fisher Scientific.
Learn more ⤵️
Show more
Get up to 40% more GPU capacity within the same power budget with NVIDIA DSX MaxLPS.
It helps turn each fixed megawatt into more AI output through dynamic power allocation, advanced performance per watt techniques, and 45°C thermal efficiency and site design.
Learn how to maximize AI factory performance per watt ➡️
Show more
