←
Jeetu Patel3,897 posts
Jeetu Patel avatar

Jeetu Patel

@jpatel41

Technology Executive & Board Member. President & CPO, Cisco. Proud dad. Love design. Views are mostly my own, but sometimes not entirely my own ;-)

◷ Joined October 2008▧ 173 Photos and videos
Jeetu Patel avatar
Jeetu Patel@jpatel41·May 24

There is such a profound shift occurring in the way that Agents alter our infrastructure requirements. We are entering a networking supercycle. It’s not because humans are consuming more content. It’s because machines are beginning to think, act, and transact continuously. Cisco's latest research on AI traffic patterns points to something much bigger than incremental bandwidth growth. Enterprise WAN traffic without agentic AI was projected to grow roughly 2.5x over the next decade. With agentic AI, that projection jumps to ~9x. And here’s the craziest part! After following this data closely, I believe even those numbers may prove to be wildly conservative. This is the first time when we have published a study like this where I feel that the projections might be off significantly and what we might think takes a decade happens in 3 years. Why? Because most people are still modeling AI like software. It is not. AI behaves more like a new species of digital labor. A SaaS app waits for humans. Agents do not. Agents continuously reason, retrieve, coordinate, negotiate, execute, and loop. At software speed. Without pause. 7x24. They never get sick. Don’t need a vacation. Dont get tired. Don’t need sleep. That creates a fundamentally different traffic architecture. The industry spent decades optimizing networks for bursty downloads, video streaming, and human-paced interactions, almost all of it flowing downstream to a person on the other end. AI traffic inverts that. A single agentic task can generate 450% more traffic than a human doing the same work. Roughly 70% of that is inference. And nearly 10% of AI flows now carry more upstream than downstream data, versus 0.5% for typical web traffic, because context continuously moves back into models. Network traffic is not just increasing in bandwidth. It is fundamentally getting reshaped. This last point matters most. The internet was built as a distribution system for content. AI is turning it into an active system for cognition. The path between agents and models is becoming the spinal cord of intelligence itself. When that path degrades, the agent degrades. Networking stops being a passive transport layer and becomes part of the intelligence stack. That changes everything about how we think about resiliency, observability, security, and capacity at the edge. We may be grossly underestimating what is coming. The future will not simply have more users online. It will have trillions of digital coworkers operating continuously on behalf of humans, enterprises, applications, and eventually physical systems. Humans click. Agents swarm. That difference is what creates a supercycle. This supercycle of inference infrastructure will not just be compute bound, but also memory and network bound. Take a look at the report here: cisco.com/c/dam/en/us/soluti…

Jeetu Patel avatar
Jeetu Patel@jpatel41·Nov 4

Thanks for having me on the pod @villaumbrosia. Always a pleasure to chat with you and love what you do for the product community.

Cisco runs a $63B business with no product managers. And agents now use 5x more tokens than humans. I got @jpatel41, President and Chief Product Officer of @Cisco, to show me: 00:00 Trailer 03:17 Back for a fifth appearance 03:45 Founder mode vs. functional mode 06:10 War mode vs. peace mode in a platform shift 06:30 What you drive top-down at a large company 08:30 Reusing your route to market to go multi-product 10:22 "No product managers, only feature managers" 10:35 Build, don't buy: Cisco's M&A discipline 13:25 Killing acquired brands, and the promise to founders 16:07 Sub-2% use agents now, 30-40% within 15 months 16:33 Agents surpass humans in token consumption 18:03 The business-model pressure coming for every company 24:37 When your customer isn't human 26:50 Fighting herd mentality with counterintuitive bets 29:11 Dinner with one person outside Cisco every week

32:06
Jeetu Patel avatar
Jeetu Patel@jpatel41·Oct 30

Internalizing this deeply unlocks the human potential in ways we can’t even fully imagine. Suddenly the whole world becomes our oyster.

the most important enabling belief is that you can learn anything - a domain, a behavior, etc.

Jeetu Patel avatar
Jeetu Patel@jpatel41·Oct 30

These experiments are super important to sponsor. Some of it works and some of it doesn’t but it’s definitely worth the pursuit. The thing that we have done here is thought of this 2 to 3 person team team as someone that catalyzes the movement and gets the core team to own the final product. What this does is gets everyone to see the benefit of a different way of working. This is also important to have more memory-safe language code migrations that by definition reduce vulns. Lots of benefits to this approach. The key is to make sure that we are constantly pushing ourselves to think of new ways to accomplish much more than wasn’t possible in the past. The thing that is not obvious here is that every engineer at Cisco will benefit from this experiment and when they are part of this modernization in areas of their code, it gives them an accelerant to build what they couldn’t in the past with the speed they wanted to move at. There is a world that is not discussed as much which is that the future could be infinitely more secure, more innovative and filled with a lot more possibilities than we’ve been able to imagine due to previous constraints.

What’s changed in the last 3 months? In my conversation with Fortune 500 CIO clients, am hearing that they are willing to go 💯 percent AI First and native (versus tweaking around the AI Progressive edges). They are done with legacy and tech debt, unless AI can completely revamp. No way they can unleash the AI value, without rewriting the core systems. This comes as no surprise from @jpatel41 and @Cisco

Jeetu Patel avatar
Jeetu Patel@jpatel41·Oct 29

Also very true 👇🏾

honestly the scarier version is the opposite when you know exactly where every single lead came from because you personally dragged it in 😭 that's not pmf, that's just cardio

Jeetu Patel avatar
Jeetu Patel@jpatel41·Oct 29

The implications for retail with personal agents are so multi-layered. One, their sales shoot up because it’s easier for consumers to shop. Two, retailers will now have an intermediary so their cross-sell and upsell strategy as well as how they design the websites and acquire not just web traffic but agent traffic will completely change. Will there be a registry for websites that are optimized for agents? Or will computer use be sufficient? Will Agent algos for product selection be driven with advertising from retailers, but advertising is just preference on product placement?

Adobe data shows AI traffic to retail sites is growing rapidly and much more valuable than other traffic: chainstoreage.com/adobe-ai-e…

Jeetu Patel avatar
Jeetu Patel@jpatel41·Oct 29

This is cool 👇🏾

Introducing Reality-Aware Rendering in @Simulon, a completely new way to create with the real world. Place neural assets with your phone and act out the scene. When you stop recording, Simulon renders the take, reading everything in it (what you do, what you say and the place around you), and your assets respond to all of it. The final is regenerated from your recording: a new performance of the scene, not a copy of the footage. Available now on iPhone. Coming to Android soon.

1:20
Jeetu Patel avatar
Jeetu Patel@jpatel41·Oct 27

This is very cool and so consequential to what’s ahead 👇🏾

Excited to share Genie 4 today our most realistic and interactive world model: - steer dynamic environments in real time, shaping events, camera angles, and actions - generate native audio matched to the world, from realistic background sounds to characters who can converse in real time! - simulate photorealistic scenes and real-world physics - create more coherent multi-step events through reasoning from Gemini We've started to roll out as a research preview to trusted testers. Coming soon to Google AI Ultra subscribers in Project Genie and early access partners via the Gemini API and AI Studio.

0:50
Jeetu Patel avatar
Jeetu Patel@jpatel41·Oct 27

Hilarious and sadly true 👇🏾

It’s kinda crazy to me that we melt sand down, clean it to 99.99999% purity, literal wizards have to draw it slowly into an ingot, we cut it, inscribe a hundred billion little on/off switches on it, and when you run electric current through it you can do basic arithmetic, which when you do enough of it can do any math on the planet, and which describes reality well enough that we can teach it the ontological meaning of words and their distance from each other on a high-dimensional space, and if you ask, it can tell you what word comes next and that apparently is powerful enough that if you do it a hundred billion times it’s the smartest thing to ever exist oh yeah and we use this primarily for email

Jeetu Patel avatar
Jeetu Patel@jpatel41·Oct 26

The risk of projecting a bear market and being wrong might be higher than the risk of projecting a bull market and being off on timing. One way doors that close the option value for future maneuverability are more expensive than they seem to the naked eye. Because the price is non-participation in the future.

Jeetu Patel avatar
Jeetu Patel@jpatel41·Oct 26

Given the scale of the numbers, even being a few percentage points off on forecasts can be tricky. On the other hand, demand is not slowing down anytime soon and all supply is instantly consumed. It’s a fascinating dynamic where capital availability at the right cost of capital and forecasting accuracy have become critical elements of success in addition to a long list of other factors, largely due to the extreme scale requirements.

Ant S1: getting the compute bet right may be as important as getting the models right. Commit too little & they can’t keep up. Too much, or the wrong hardware, & the whole business can be at risk

Jeetu Patel avatar
Jeetu Patel@jpatel41·Oct 24

Indistinguishable from a human voice. Assuming this works reliably well, interfaces could change quite materially for machines. Really impressive breakthrough 👇🏾

Block@blocksOct 23

Meet Dessa, our speech model trained from scratch on human conversation. Hear it for yourself: block.xyz/dessa

0:37