Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

Cognition is the first customer running @NVIDIA Vera Rubin, powered by CoreWeave! Hopper in 2022. Blackwell in 2024. Now Vera Rubin: On SWE-2 inference, the new chips deliver about 4.8x more token throughput than GB200 at the same decode speed. More compute, better agents!

Cognition

187,417 subscribers

118,479 views • 3 days ago •via X (Twitter)

Education Science & Technology

Anya Rossi• Live Now

Private livecam show

35 Comments

Jared Zoneraich3 days ago

@nvidia @CoreWeave Hardware also matters!

Felix Schlegel3 days ago

@nvidia @CoreWeave more compute, better agents!

Chris W3 days ago

@nvidia @CoreWeave I'd say thanks for taking care of getting SWE-2 on this amazing hardware!

Matt Bergland3 days ago

@nvidia @CoreWeave lfg

Kartik3 days ago

@nvidia @CoreWeave So that's why the speeds got faster today!

Jay Grant3 days ago

@nvidia @CoreWeave monumental

ANUP3 days ago

@nvidia @CoreWeave cognition is not going to slow down any sooner 🔥

Jake Cosme3 days ago

@nvidia @CoreWeave Always stay shipping!

Mitchell Mecum3 days ago

@nvidia @CoreWeave Accelerate!!!

Mark Vassilevskiy3 days ago

@nvidia @CoreWeave 🔥🔥🔥

cashd3 days ago

@dabit3 @nvidia @CoreWeave Love it! But music is too loud over his speech

Justin3 days ago

@sandeepDshah @nvidia @CoreWeave ⚡️

Alex Volkov3 days ago

@nvidia @CoreWeave LFG!

J A Z I I3 days ago

@nvidia @CoreWeave now that's banger

Anderson3 days ago

@nvidia @CoreWeave 4.8x the tokens per second so the agent can now be wrong at a much higher speed

Mark Murphy3 days ago

@nvidia @CoreWeave Congrats! Been using Devin daily. Keep up the awesome work!

Ron3 days ago

@nvidia @CoreWeave @KampnerSam

y3 days ago

compute always wins

Lorenzo Porras3 days ago

@nvidia @CoreWeave

Saiyam Pathak3 days ago

@dabit3 @nvidia @CoreWeave crazy!!

Diana Zawadzki3 days ago

@nvidia @CoreWeave So epic!

diedbroke3 days ago

@nvidia @CoreWeave I bet you factory already running quantum computers. That's why y'all behind.

bnistordev3 days ago

@nvidia @CoreWeave I really love when I see Cognition name under this kind of announcements.

Aman3 days ago

@nvidia @CoreWeave this is gonna be an insane combo for agentic workloads

Dan Shamir3 days ago

@nvidia @CoreWeave You also steal board members from competitors. Losers

Aleksandr Grishin2 days ago

@nvidia @CoreWeave 🥳🥳

高嗯嗯2 days ago

@nvidia @CoreWeave swe 如果一直免费我将一直订阅

The SI Therapist3 days ago

@nvidia @CoreWeave @NVIDIA Vera Rubin delivers 4.8x more tokens on SWE-2 than previous chips. It looks like Cognition and CoreWeave have built something fast for AI coding tasks.

Ibesh3 days ago

@nvidia @CoreWeave Will the 4.8x throughput gain reduce queue times for Devin users during peak hours?

sophs j.3 days ago

@nvidia @CoreWeave Love this take

Crio Songo2 days ago

@nvidia @CoreWeave 4.8x throughput boost is really impressive. This iteration will definitely push agent performance to a new level.

Gabriel Abi3 days ago

@nvidia @CoreWeave 4.8x on SWE-2 with same decode speed — that's the unlock for agentic workloads. inference bottlenecks become a thing of the past at this rate

Damian Z. 💛🇪🇺2 days ago

@nvidia @CoreWeave LFG

MOSSE.2 days ago

@nvidia @CoreWeave So Jump Trading lied ?

Almost Tomorrow3 days ago

@nvidia @CoreWeave How much of that throughput gain translates into a lower cost per completed task, once retries and tool calls are included?

Related Videos

Nvidia $NVDA CEO just said that Vera Rubin is now in full production The supply chain Nvidia created for Vera Rubin is twice as large as Grace Blackwell

Nvidia $NVDA CEO just said that Vera Rubin is now in full production The supply chain Nvidia created for Vera Rubin is twice as large as Grace Blackwell

Evan

120,926 views • 4 months ago

Jensen Huang said Grok 5 will be 7 Trillion parameter model. On time to train, with the training window fixed at 1 month, the new Vera-Rubin GPU system of Nvidia needs 1/4 the number of systems compared with Blackwell to train the same frontier model. “factory throughput” improves by about 10x over Blackwell, and Blackwell itself was about 10x over Hopper. i.e. overall Rubin is roughly 100x Hopper in factory throughput per watt, which matters hugely, because a 1 GW, $50B data center is power-limited and revenue scales with throughput per watt. --- From Nvidia YT channel

Jensen Huang said Grok 5 will be 7 Trillion parameter model. On time to train, with the training window fixed at 1 month, the new Vera-Rubin GPU system of Nvidia needs 1/4 the number of systems compared with Blackwell to train the same frontier model. “factory throughput” improves by about 10x over Blackwell, and Blackwell itself was about 10x over Hopper. i.e. overall Rubin is roughly 100x Hopper in factory throughput per watt, which matters hugely, because a 1 GW, $50B data center is power-limited and revenue scales with throughput per watt. --- From Nvidia YT channel

Rohan Paul

165,770 views • 8 months ago

Nvidia just revealed Vera Rubin. Ships H2 2026. The numbers are wild: → 10x more performance per watt vs Blackwell → 10x cheaper inference token cost → 4x fewer GPUs to train the same MoE model Energy was the biggest bottleneck in AI. Nvidia just made it 10x cheaper.

Nvidia just revealed Vera Rubin. Ships H2 2026. The numbers are wild: → 10x more performance per watt vs Blackwell → 10x cheaper inference token cost → 4x fewer GPUs to train the same MoE model Energy was the biggest bottleneck in AI. Nvidia just made it 10x cheaper.

Min Choi

285,639 views • 7 months ago

NVIDIA Vera Rubin is in full production and arrives just in time for the next frontier of AI. The #NVIDIARubin platform uses extreme co-design across six new chips to accelerate agentic AI, advanced reasoning and massive-scale MoE model inference. NVIDIA Vera Rubin NVL72: Six new chips, one AI supercomputer — see how we built it.

NVIDIA Vera Rubin is in full production and arrives just in time for the next frontier of AI. The #NVIDIARubin platform uses extreme co-design across six new chips to accelerate agentic AI, advanced reasoning and massive-scale MoE model inference. NVIDIA Vera Rubin NVL72: Six new chips, one AI supercomputer — see how we built it.

NVIDIA Data Center

27,643 views • 9 months ago

The NVIDIA Vera Rubin platform is a multi-rack pod-scale system, built to process agentic AI and is now in full production. Built through extreme co-design, Vera Rubin unifies five connected rack-scale systems — the NVIDIA Vera Rubin NVL72, Vera CPU rack, Groq 3 LPX, Vera BlueField-4 STX, and Spectrum-6 SPX Ethernet — all purpose-built for global-scale agentic AI infrastructure. Vera Rubin delivers intelligence at scale, with the performance, efficiency and security needed to power the next industrial revolution. Congratulations to Microsoft, Dell Technologies, and CoreWeave for each standing up their Vera Rubin NVL72 engineering racks. Learn more:

The NVIDIA Vera Rubin platform is a multi-rack pod-scale system, built to process agentic AI and is now in full production. Built through extreme co-design, Vera Rubin unifies five connected rack-scale systems — the NVIDIA Vera Rubin NVL72, Vera CPU rack, Groq 3 LPX, Vera BlueField-4 STX, and Spectrum-6 SPX Ethernet — all purpose-built for global-scale agentic AI infrastructure. Vera Rubin delivers intelligence at scale, with the performance, efficiency and security needed to power the next industrial revolution. Congratulations to Microsoft, Dell Technologies, and CoreWeave for each standing up their Vera Rubin NVL72 engineering racks. Learn more:

NVIDIA

49,740 views • 4 months ago

The first-ever measured silicon numbers for NVIDIA Vera Rubin NVL72 showed 10x more tokens per megawatt than Blackwell. No projections. Real results from live hardware. The full methodology, why matched interactivity is the honest way to compare a reasoning model, and more:

The first-ever measured silicon numbers for NVIDIA Vera Rubin NVL72 showed 10x more tokens per megawatt than Blackwell. No projections. Real results from live hardware. The full methodology, why matched interactivity is the honest way to compare a reasoning model, and more:

CoreWeave

172,785 views • 1 month ago

📣 CoreWeave and NVIDIA are accelerating the build out of AI factories—token generators that directly drive revenue. This includes adoption of and support for: ✅ NVIDIA Vera Rubin deployment, starting in 2H 2026 ✅ NVIDIA Vera CPU and BlueField-4 STX ✅ NVIDIA Dynamo for inference and NVIDIA Agent Toolkit and OpenShell for long-running agents ✅ NVIDIA DSX Air for digital twins of AI factories Learn more ➡️

📣 CoreWeave and NVIDIA are accelerating the build out of AI factories—token generators that directly drive revenue. This includes adoption of and support for: ✅ NVIDIA Vera Rubin deployment, starting in 2H 2026 ✅ NVIDIA Vera CPU and BlueField-4 STX ✅ NVIDIA Dynamo for inference and NVIDIA Agent Toolkit and OpenShell for long-running agents ✅ NVIDIA DSX Air for digital twins of AI factories Learn more ➡️

NVIDIA AI Infrastructure

10,213 views • 5 months ago

NVIDIA Vera Rubin NVL72 production racks are here. The compute tray is engineered for fast compute, assembly, and serviceability. Manufacturing is 100% automated, and every tray goes together in one minute. Congratulations to Microsoft on the first operational Vera Rubin NVL72 racks, now rolling off Hon Hai Technology Group (Foxconn) Ingrasys lines.

NVIDIA Vera Rubin NVL72 production racks are here. The compute tray is engineered for fast compute, assembly, and serviceability. Manufacturing is 100% automated, and every tray goes together in one minute. Congratulations to Microsoft on the first operational Vera Rubin NVL72 racks, now rolling off Hon Hai Technology Group (Foxconn) Ingrasys lines.

NVIDIA

275,046 views • 1 month ago

The NVIDIA Vera Rubin NVL72 scale-up fabric, built on the sixth-generation NVIDIA NVLink. The NVLink 6 Switch trays connect 72 Rubin GPUs in an all-to-all configuration. The NVLink spine—over 5,000 copper cables make up the vertical NVLink traverse. It securely moves more data than the entire bandwidth of the internet. The single-wide NVIDIA MGX rack design unlocks a cable-free, horizontal NVLink scale-up traverse, running directly on PCB for resiliency at scale. This is how NVIDIA Vera Rubin NVL72 works as one, designed to deliver the lowest token cost and 10x more performance per watt than the previous generation. #NVIDIAVeraRubin

The NVIDIA Vera Rubin NVL72 scale-up fabric, built on the sixth-generation NVIDIA NVLink. The NVLink 6 Switch trays connect 72 Rubin GPUs in an all-to-all configuration. The NVLink spine—over 5,000 copper cables make up the vertical NVLink traverse. It securely moves more data than the entire bandwidth of the internet. The single-wide NVIDIA MGX rack design unlocks a cable-free, horizontal NVLink scale-up traverse, running directly on PCB for resiliency at scale. This is how NVIDIA Vera Rubin NVL72 works as one, designed to deliver the lowest token cost and 10x more performance per watt than the previous generation. #NVIDIAVeraRubin

NVIDIA AI Infrastructure

39,134 views • 2 months ago

Vera Rubin is in full production. We just kicked off the next generation of AI infrastructure with the NVIDIA Rubin platform, bringing together six new chips to deliver one AI supercomputer built for AI at scale. Here are the top 5 things to know 🧵

Vera Rubin is in full production. We just kicked off the next generation of AI infrastructure with the NVIDIA Rubin platform, bringing together six new chips to deliver one AI supercomputer built for AI at scale. Here are the top 5 things to know 🧵

NVIDIA

817,297 views • 9 months ago

NVIDIA Vera is heading to Amazon Web Services. NVIDIA’s Ian Buck hand-delivered AWS’s first Vera CPU Server and Vera Rubin GPU to Willem Visser and Supreeth Sheshadri at the AWS HQ in Seattle. Vera is purpose-built for agentic AI: more tokens per dollar, faster results for users, and the compute foundation AI factories need to scale.

NVIDIA Vera is heading to Amazon Web Services. NVIDIA’s Ian Buck hand-delivered AWS’s first Vera CPU Server and Vera Rubin GPU to Willem Visser and Supreeth Sheshadri at the AWS HQ in Seattle. Vera is purpose-built for agentic AI: more tokens per dollar, faster results for users, and the compute foundation AI factories need to scale.

NVIDIA

96,600 views • 1 month ago

Watch Nvidia CEO Jensen Huang reveal what he described as the state of the art in AI hardware: the new Rubin computing architecture. “Vera Rubin is designed to address this fundamental challenge that we have: The amount of computation necessary for AI is skyrocketing.” Huang said. “Today, I can tell you that Vera Rubin is in full production.” According to Nvidia’s tests, Rubin architecture operates three and a half times faster than the previous Blackwell architecture on model-training tasks and five times faster on inference tasks, and that's just the start.

Watch Nvidia CEO Jensen Huang reveal what he described as the state of the art in AI hardware: the new Rubin computing architecture. “Vera Rubin is designed to address this fundamental challenge that we have: The amount of computation necessary for AI is skyrocketing.” Huang said. “Today, I can tell you that Vera Rubin is in full production.” According to Nvidia’s tests, Rubin architecture operates three and a half times faster than the previous Blackwell architecture on model-training tasks and five times faster on inference tasks, and that's just the start.

TechCrunch

76,412 views • 9 months ago

NEWS: Nvidia has released a new video of its next-generation Rubin chips, which are 5X more powerful than their predecessor, Blackwell. The new Rubin chips are already in production. CEO Jensen Huang: "This is a Rubin pod. 1,152 GPUs and 16 racks. Each rack has 72 Rubins."

NEWS: Nvidia has released a new video of its next-generation Rubin chips, which are 5X more powerful than their predecessor, Blackwell. The new Rubin chips are already in production. CEO Jensen Huang: "This is a Rubin pod. 1,152 GPUs and 16 racks. Each rack has 72 Rubins."

Sawyer Merritt

798,342 views • 9 months ago

The NVIDIA Vera Rubin NVL72 compute tray. 200 AI petaFLOPs. Assembled in one minute. That's what a single-wide, third-generation NVIDIA MGX rack enables. No cables. No hoses. No fans. It's 100% liquid cooled at 45°C and brings together Vera Rubin Superchips, ConnectX-9 SuperNICs, and BlueField-4 DPUs to deliver the lowest token cost and best performance per watt. #NVIDIAVeraRubin

The NVIDIA Vera Rubin NVL72 compute tray. 200 AI petaFLOPs. Assembled in one minute. That's what a single-wide, third-generation NVIDIA MGX rack enables. No cables. No hoses. No fans. It's 100% liquid cooled at 45°C and brings together Vera Rubin Superchips, ConnectX-9 SuperNICs, and BlueField-4 DPUs to deliver the lowest token cost and best performance per watt. #NVIDIAVeraRubin

NVIDIA AI Infrastructure

334,083 views • 2 months ago

$NVDA says it has shipped its first Vera Rubin samples to customers and remains on track for production shipments in 2H. $AMD Helios vs. Vera Rubin is shaping up to be the compute Super Bowl of 2026.

$NVDA says it has shipped its first Vera Rubin samples to customers and remains on track for production shipments in 2H. $AMD Helios vs. Vera Rubin is shaping up to be the compute Super Bowl of 2026.

Shay Boloor

66,499 views • 7 months ago

From the show floor to the AI factory: Hot Chips 2026 was all about extreme co-design to accelerate agentic workloads — the most complex workload in history. Vera CPU, Vera Rubin, Groq 3 LPX, Spectrum-X Multiplane, BlueField-4 Scale-In networking. One full AI stack platform built for agents.

From the show floor to the AI factory: Hot Chips 2026 was all about extreme co-design to accelerate agentic workloads — the most complex workload in history. Vera CPU, Vera Rubin, Groq 3 LPX, Spectrum-X Multiplane, BlueField-4 Scale-In networking. One full AI stack platform built for agents.

NVIDIA

103,485 views • 1 month ago

$NVDA Vera Rubin cycle is officially moving from roadmap to reality with the first operational NVL72 rack now live at $MSFT. Nvidia says production racks are here marking the start of the next major AI infrastructure deployment wave after Blackwell.

$NVDA Vera Rubin cycle is officially moving from roadmap to reality with the first operational NVL72 rack now live at $MSFT. Nvidia says production racks are here marking the start of the next major AI infrastructure deployment wave after Blackwell.

Shay Boloor

94,910 views • 1 month ago

NVIDIA HAS TAPED OUT 6 RUBIN CHIPS AT TSMC - New CPU (Vera - 88 ARM cores with SMT) - New GPU variants (R100 with HBM4) - Scale-up NVLink switch - Silicon photonics processor - Networking silicon - Complete chipset overhaul Reminder that Rubin is on-schedule for mass production in the 2nd half of 2026

NVIDIA HAS TAPED OUT 6 RUBIN CHIPS AT TSMC - New CPU (Vera - 88 ARM cores with SMT) - New GPU variants (R100 with HBM4) - Scale-up NVLink switch - Silicon photonics processor - Networking silicon - Complete chipset overhaul Reminder that Rubin is on-schedule for mass production in the 2nd half of 2026

NIK

78,410 views • 1 year ago

🚨HUGE: Nvidia is charging MORE for its chips, and Jensen Huang says he did it on purpose. The Nvidia CEO says the company chose to "rip the band-aid off" after memory costs exploded. A single Vera Rubin rack could now cost around $8 MILLION.

🚨HUGE: Nvidia is charging MORE for its chips, and Jensen Huang says he did it on purpose. The Nvidia CEO says the company chose to "rip the band-aid off" after memory costs exploded. A single Vera Rubin rack could now cost around $8 MILLION.

Coin Bureau

46,258 views • 1 month ago

Sachin Katti says OpenAI had Astra optimize its own inference on NVIDIA's Rubin chip, adding another 2x in 72 hours "It's a testament to NVIDIA's hardware that out of the box we got a 3x improvement in throughput. We literally took the same code that we had optimized for Blackwell, ran it on Rubin, got 3x improvement." "Then, can we actually make Astra run agents using Astra as the base model, go optimize this inference of Astra itself on Rubin? It's a recursive loop. The model is figuring out how to optimize its own inference serving on Rubin chips, and over 72 hours the model produced another 2x improvement on the throughput." "Things that would take weeks, if not months, to optimize in the past, we now are getting models that are very capable of actually producing over a weekend." _______ Link and more from Sachin Katti's conversation:

Sachin Katti says OpenAI had Astra optimize its own inference on NVIDIA's Rubin chip, adding another 2x in 72 hours "It's a testament to NVIDIA's hardware that out of the box we got a 3x improvement in throughput. We literally took the same code that we had optimized for Blackwell, ran it on Rubin, got 3x improvement." "Then, can we actually make Astra run agents using Astra as the base model, go optimize this inference of Astra itself on Rubin? It's a recursive loop. The model is figuring out how to optimize its own inference serving on Rubin chips, and over 72 hours the model produced another 2x improvement on the throughput." "Things that would take weeks, if not months, to optimize in the past, we now are getting models that are very capable of actually producing over a weekend." _______ Link and more from Sachin Katti's conversation:

Fireside Alpha

40,013 views • 16 days ago