This is my conversation with Michael Kagan, the founder of Mellanox and current CTO of NVIDIA We speak about his life, the progression of networking over the span of his career, and how the work that NVIDIA is now doing to build "Planet scale" computing I hope you enjoy the conversation as much as I did 00:00 Highlights 01:16 Growing Up in Russia and Moving to Israel 03:42 Intel, the PC Revolution, and a Passion for Technology 06:14 Sponsor: Liquid 06:33 Lessons from Intel: Software and Networking 10:10 Founding Mellanox and the Vision for Cloud Computing 14:34 From Supercomputers to the Cloud 17:46 How GPUs Sparked the AI Revolution 20:52 Sponsor: Compute Exchange 21:14 NVLink and the Rise of Rack-Scale Computing 24:57 The Engineering Challenges Behind NVL72 27:39 Designing Hardware for Future AI Workloads 29:43 BlueField DPUs and the AI Factory 32:55 Making AI Agents Safer 35:37 What Mellanox Could Have Done Differently 37:03 NVLink, InfiniBand, and Spectrum-X 39:24 Connecting AI Factories at Planetary Scale 42:42 Keeping Massive AI Systems Running Through Failures

Update on Instinct use case ~ Currently working on scheduling 2 important group dinners 1 with my new friends at SF compute ( @_nikta_khani ) , the other a ultra super duper top secret illuminati dinner Mission: find place with good food that doesn’t make @carmenli think I’m spending too much money. Has to be excellent vibes, good food. Private space preferred. instonct : plz find me some places that work for this in month of Nov, ideally early Nov, seek and destroy It finds all the emails, sends out, its managing the reposes, coordinating into a carousel of options. We can compare/contrast “Make no mistakes” prompt submitted. “Computer, make sure we have a great time — Interestingly, one of the ~7 places i had instinct hit up in sf blocked me with Captcha , immediately no. Off the list no questions asked. We do not have time for that anymore If you want to block my bots I’m done doing business with you as a restaurant. Simple as that. But other than that, I basically have a personal EA coordinating this so we have an epic time… pretty cool

What a turnout at the Compute Trading Happy Hour for #SFTechWeek#. 850+ people requested a spot and the Chotto Matte rooftop was packed. Thanks to everyone who came, and to @Silicon_Data and @robotventures for hosting with us.

We’re looking for an exceptional Product Manager for @computeexchange Salary is $150-$230k depending on experience level / background. Comes with equity Preference to be based in the Bay Area or NYC To apply DM me or @computeexchange your background. Let us know why you’d be a good fit working with Compute Exchange, and share your CV/Linkedin

There’s a common misconception that older hardware is becoming plentiful and losing value rapidly. At @computeexchange we’re not seeing that happen Buyers looking for H100s (4 year old hardware) are having an extremely difficult time. There’s virtually no supply available

@Etched founder just said they (Etched) is one of the first companies creating a “memory swap” that they can raise money against - they’re betting that memory prices will be going up Etched is betting that memory prices will be going up for “awhile” Wow…

Well, this is fast. We tested @nvidia's new Vera CPU head-to-head against our most sought-after CPUs. 8x faster to spin up 1,000 @daytonaio sandboxes. 2,000 done in 27.4s, before Zen 5 finished 1,000. Like a Rimac Nevera hitting 200 km/h before a Porsche 911 clears 100. Full results soon 👀

Thank yoi @a16z for citing @Silicon_Data rental indices and residual value data! The key point: GPU depreciation and obsolescence don’t have to be assumptions anymore. We now have real market data to measure them.

One week into my new role as a compute deals lead, I’ve already seen tons of buyers and sellers discussing, going back and forth, determining the market in real time One things clear Everyone wants NVIDIA gear. There is absolutely zero mention of anything else. CUDA moat (or whatever the moat is now) is alive and well, and business is indeed booming

NV Vera CPU on display again at the Coreweave event Nvidia does not get the credit it deserves for the foresight it had to build a CPU designed for agentic use cases Now, everyone in the industry is claiming they have an “agentic CPU” but Vera was first to market and first to scale by a long shot Vera is the product of tremendous foresight and hard work, and is extremely well suited for high value agentic workloads