Advancing AI 2026 | Jeetu Patel, Cisco

Clips
More from Advancing AI 2026

Jeetu Patel

President & CPO

Cisco

play_circle_outline Inference becoming majority of compute spend; infrastructure and memory demand rising

play_circle_outline Agent-Driven Inference: Scaling Sustained, High-Volume Models on Device and Desktop Endpoints

play_circle_outline Cisco Cloud Control & AMD: Hybrid Inference, Intelligent Model Routing, Tokenomics, Observability, and Sovereign Enterprise AI Security

play_circle_outline Open vs Closed Models: Why Small Specialized Systems Excel at Vulnerability Detection

play_circle_outline Ease-of-use barrier: agents need to be simpler for widespread adoption

Info

Jeetu Patel, Cisco

Jeetu Patel

President & CPO Cisco

search

(INTRO)

Dave Vellante

>> Hi everybody, welcome back to Moscone West. You're watching theCUBE's live coverage of AMD's Advancing AI 2026. I'm Dave Vellante. We have a special guest here on our live coverage, Jeetu Patel is the President and Chief Product Officer of Cisco. Awesome seeing you, thanks for jumping in, I really appreciate it.

Jeetu Patel

>> It's so great to see you. The best ones are the spontaneous ones, man.

Dave Vellante

>> Yeah, you got it. First of all, Cisco's on a roll. we have seen you really focus the team. You're a builder, getting the product really to where it needs to be, and so congratulations on that.

Jeetu Patel

>> We have a great team.

Dave Vellante

>> Give us the quick update, and I want to get into the news.

Jeetu Patel

>> Yeah, the update is, I think, we're in a quiet period, but last quarter was fantastic. I think the general momentum of innovation has never been better than what it has been over the course of the past two years. I'd say the past two years we invented more than we had in the previous 20 years combined. And I think that's going to be dwarfed in comparison to what we have in store in the next couple years. So I think the team's excited. The only thing we're not getting enough sleep. So, we gotta every once in a while, take a Saturday off or something.

Dave Vellante

>> Well, and I know you're in a quiet period, so you can't say much. but Lisa laid out the TAM, the growth, the demand, where the demand is coming from, claiming that this is really the first year that inferencing is sucking up more compute than training, which we've been saying for a long time, it's finally happening, but there doesn't seem to be any end in sight to the demand. Certainly, the memory shortage looks like it's going to last through 27 and beyond.

Jeetu Patel

>> There's less than 1 % of the world that's using agents. Right. And if you believe that agents are actually a step function improvement from a chatbot, where you can actually have either your personal productivity or your company's productivity materially change as a result of agents. I don't see how you don't stay in a continued kind of supply shortage for a pretty long time period. And that's what I think we're seeing play out.

Dave Vellante

>> Yeah, the signals are very positive. We saw, notwithstanding IBM's disappointment, but that's really not, I don't think they're a bellwether for AI, Supermicro doubling its margins and obviously growing, you're going to probably see something similar from Dell, you're going to see something similar from NVIDIA, and clearly AMD is on a rocket ship. Okay, you got an announcement here, that you were up on stage, we're talking about distributed inference, take us through the news here.

Jeetu Patel

>> So what you're starting to see happen right now, Dave, is the pattern of inferencing is fundamentally changing as you move from chatbots to agents. So in the chatbot era, when humans were the ones that were driving the engagement, you ask a question, you get an answer. The patterns of inferencing were very spiky. what you're now starting to see in the agentic era is that these agents are working 7x24 on your behalf. They might actually go out and talk to other agents. And so the pattern of inferencing is much more durable and sustained and not as spiky. and actually the volume is higher. And so what you're starting to see is every customer I talk to right now has a couple major concerns. Hey, I'm concerned about security and safety, and I'm concerned about the cost of tokens, because these tokens can start to run up in cost, and you might not have had enough time to go out and budget for all of those. So we need to make sure that as an industry, we lower the cost of tokens quite dramatically. And so one way that that's happening is you're not just going to have inferencing run in the cloud. You'll also see a whole new set of models that emerge for inferencing running everywhere and locally. And one of those kinds of patterns is a whole new class of computing that's emerging called desk -side computing. So you work on your laptop right now, and that's a human working on a laptop. But you might have a desk-side computer where the agents are running. And by the way, this isn't really a new phenomenon. It's just taking more effect right now.

Jeetu Patel

>> I do have that, by the way.

Jeetu Patel

>> You do have that, and if you remember when OpenClaw first came out, Mac mini sold out.

Jeetu Patel

>> Right.

Jeetu Patel

>> Right? And you're starting to see that same pattern, which is like there's going to be more inferencing that happens on the desk side. Now, by the way, if you have enterprises that start to go out and deploy this, you can't just deploy and pray that everything works out right. You need to make sure that it's managed effectively, you need to make sure that it's secure, you need to make sure that the agent behavior is monitored, You need to make sure that you figure out whether the cost of tokens is out of hand. And if so, what do you need to do to make sure that you can contain the cost of tokens? All of that apparatus needs to be put in place. So what we have done is we have our unified management plane that you know well, Cisco Cloud Control. And what it does is it can go out and look at all of your inferencing capacity that you have in the cloud, all of the inferencing capacity you might have in a private data center, as well as your desk side computers, in a single management plane, and then be able to say, well, what do we need to do to ensure that you have the best cost efficiency and performance and security when you're running in this hybrid inference model, which is what I think is going to be the prominent model as we move forward.

Jeetu Patel

>> So you're doing the routing there?

Jeetu Patel

>> So in this particular case, AMD does a specific intelligent routing of the models. They have the actual compute. We provide the network bandwidth that's going to be needed. We provide the security and safety that's needed. We provide agent behavior monitoring so that if an agent's starting to go rogue, you can actually make sure that you can sequester an entire local machine. And we provide tokenomics, so you can figure out how many tokens are being used, is it more efficient than running it in the cloud, what do you need, which agents are using which amounts of tokens? Is that agent starting to show patterns of excessive token usage? If so, do you want to contain it? All of that kind of telemetry is provided in Cloud Control.

Dave Vellante

>> You guys provide that telemetry, that observability. That's fundamental to what you guys do there. Okay.

Jeetu Patel

>> We have early customers using it right now, but Cloud Control is going to go general availability in the US in early fall. And so we're excited about how customers can lower their cost, improve their efficiency for tokens.

Dave Vellante

>> It's interesting, during the pandemic everybody was working remote, they just threw everything to the cloud, it was spend, spend, spend. 2022 comes around, we're coming out of the pandemic and cost optimization was the big theme. Now you're seeing that token cost optimization.

Jeetu Patel

>> Because the volumes are higher, right? And the numbers are getting bigger. So most companies don't have a few hundred million just lying around that they can go out and start spending on tokens, they have to plan for it.

Dave Vellante

>> But at the same time, we've seen token costs come down 90 % a year. I know there's a big delta between the most expensive tokens that you're getting through APIs in the cloud from the frontier models and some of the Chinese open models. that's like orders of magnitude delta. what do you make of that? Ultimately, is this a short-term trend with this token cost optimization, or do you see this? you talk about it. Ali Ghodsi is really pounding the table on this. Do you see that as an ongoing concern? Or ultimately, are tokens going to be essentially free?

Jeetu Patel

>> The way I see it is it's an ongoing way that we will operate. Just like we have a finite set of resources with humans, and we have to plan out what those humans are going to be able to do. You're going to have a finite set of resources with tokens, and you're going to need to make sure that you can put those tokens to the best use so you can get the best outcomes for your organization. And so that's what's going to happen. And I think the reason that I don't think it's just going to be a short-term trend is because, yes, you might have seen the cost of tokens coming down by 90%, but the demand signal of how compute-intensive these agents are on infrastructure has gone up way faster than that. So your total cost is still going up within the organization. And that's going to be a finite bucket, and we need to make sure that that finite bucket is being applied in the best way possible.

Dave Vellante

>> So I want to run something by you. So the macro, if you look at spending on IT broadly across the world, organizations spend about four or five percent of revenue on IT. We forecast that that's going to double or triple by mid next decade. Why? Because you're essentially going to be scaling without the labor that you used to require to scale. We're not talking about wiping out jobs. That's really not the point. The point is you're going to shift investments to tokens, whether it's through APIs or you're generating, spending on CapEx on your own, and that is going to be where the ROI is. You're not seeing that big trend today yet. You're seeing the cost, the token cost. You're not seeing the productivity hits just yet. We're at the early days. But do you buy that premise that over time you're going to be able to scale dramatically without as much labor and that's going to justify that doubling or tripling of that revenue going to technology?

Jeetu Patel

>> So I have a slightly counterintuitive perspective on this one, which is that as AI is starting to advance and automate workflows and do things that are in a much more effective way than what was done in the past, more human bottlenecks are emerging than there were in the past. So I actually think you're going to see more jobs as a result of AI rather than less jobs. The only difference is the dexterity for AI is going to be needed in every human being. So anyone who's actually not AI fluent is going to have a harder time. But the people that are actually fluent in AI today, they're in more demand, they're making more money, they're busier than ever, and you're seeing this across the board. So I think you're going to actually see that being, so that's one dimension. The second dimension is, we have overly focused just on productivity. I think we need to focus on what class of problems could we never even dream of solving before, that we can solve as a result of AI. And I think that's a far more important question than just pondering what can we do to go out and reduce the cost for humans.

Dave Vellante

>> Yeah, so, and by the way, I would agree with that. I think we're going to have more jobs. I guess my point is, and I want to test this on you, we're going to be able to scale dramatically faster than we would in the past with a lot less labor. That labor is going to have tools and be more intelligent.

Jeetu Patel

>> Per unit output, you might have less labor, but you will have more jobs than what you had before.

Dave Vellante

>> I would totally agree with that. And when I talk about productivity, I have a very simple measure of productivity, revenue per employee. And so you're going to have much more revenue. you look at what Uber has done. Now, of course, they have a different business model. I got to ask you, it's interesting you said about only 1 % of us are using agents, building agents and using agents. We're in this bubble.

Jeetu Patel

>> We're in this bubble, especially in Silicon Valley. That's why I find it so important to go keep traveling to different parts of the world, because you want to make sure that you get a feel of how everyone's seeing this thing kind of take off, and it's not just within a very, very hyper -concentrated bubble.

Dave Vellante

>> So that brings me to, you saw, everybody saw Alex Karp's rant, which is quite interesting, and then you're seeing all the action around open weight models, of course, the Moonshot and what's happening with the Chinese open weight models. I saw somebody on TV the other day call it AI communism. It's not AI communism, in fact, it's the opposite. is that companies want to deploy open weight models so they can protect their IP, their alpha, and that's AI capitalism, but it's actually being driven by these open weight models, the likes of which China are really aggressive. Now, increasingly you're seeing US companies put out open weight models. My question to you is, it looks like certainly Anthropic and OpenAI are not capitulating, nor is Google, on closed models. and their argument is, look, we're going to have better integration, think iPhone. And I do buy some of that argument, but how do you see that playing out? Very clearly, what we published, John and I, is the power law of Gen AI way back when, where we said there's going to be a lot of these smaller language models and specialized models. That's clearly playing out. At the same time, I don't necessarily think it calls for the demise of these frontier models. You have a perspective on this. Your AI summit earlier this year was amazing. what's your thought on that?

Jeetu Patel

>> By the way, next year is February 2nd, so get ready.

Dave Vellante

>> Okay, I'm going to write that down because I don't usually travel if I can avoid it in February, but I would make a special trip for that because that was, I would say, the top event of the year. I would put it over GTC, which was amazing.

Jeetu Patel

>> We love Jensen, though, and I love Jensen's events.

Dave Vellante

>> Your thought leadership on that was fantastic. But what do you think about the prospects of the leading frontier models who right now are out of favor, they're spending all this money, but they have really smart people at the top. what's your take on their future? what do they have to do to really thrive and develop a business model that's durable?

Jeetu Patel

>> So let me take a step back, because this is a very important question, and I think we tend to get into very zero -sum discussions in this, and I think that's actually not productive. what I think is going to happen, I think firstly, what are the things that people look for as metrics to measure with models? They're looking for intelligence, right? per unit expended of cost. They're looking for cost itself, how do you contain the cost? And they're looking for control. Now, do you always want to use a frontier model for everything? No, that would be like taking a private jet to your corner store. You're not going to use a frontier model for everything. There will be certain things that make a lot of sense to use a frontier model. I think there's going to be intelligent routing which says when do you use a model that's going to be super expensive because you need that intelligence and horsepower, and when do you use a model that might not need that? So for example, we just announced we're launching a bunch of open weights models for a very, very specific purpose, which is finding vulnerabilities in code.

Dave Vellante

>> Awesome announcement.

Jeetu Patel

>> We just did that a couple days ago, thank you.

Dave Vellante

>> That's a tremendous announcement.

Jeetu Patel

>> And that particular announcement doesn't mean that we actually don't partner very deeply with Anthropic and with OpenAI. It just means that you're going to need to have SLMs that you want to have control over. And if you want to make sure that you try to find vulnerabilities in your code, you don't want to have that go in the cloud. You want to actually be able to do it right from within your environment. And if I can find 70 % of the vulnerabilities that way, great. And the remaining 30 % I can find with a frontier model.

Dave Vellante

>> So it's not a zero -sum game. And by the way, that announcement was outstanding. And I think the market is so confused right now. It took the market down, took all the cybersecurity stocks down.

Jeetu Patel

>> I do think, though, that there's going to be a very important dynamic that we have to make sure happens. It's America needs our own encouragement of open source. I think Nemotron is doing a great job at NVIDIA.

Dave Vellante

>> Thinking Machines just came out.

Jeetu Patel

>> Thinking Machines came out with one. And I think we need to make sure that we have more. We are coming out with small language models in domains that are very, very dear to us, where we have telemetry, we have a unique perspective. I think you're going to start to see more and more use of open weights models. You will see more open source come out, and you will see more closed source models. And what you have to do is get the cost of inferencing down, because inferencing is now the majority of the compute, as you heard from Lisa's keynote. But you also need to make sure that you're continuing to keep pushing the intelligence frontier, because we're nowhere near the top of the S-curve there. We need to continue to keep making sure we solve harder and harder problems.

Dave Vellante

>> I think there's another dimension of enterprise requirement.

Dave Vellante

>> Control.

Dave Vellante

>> You mentioned control, but why is it that we only have 1 % of us really driving agents? Ease of use.

Dave Vellante

>> I think it's too hard.

Dave Vellante

>> It's too hard, exactly. It's too hard. And that's why I want to bring ease of use into it. and I think this is the frontier model's opportunity. They have to make it easier. And Ray Wang's book, Everybody Wants to Rule the World, I think the frontier models want to rule the world. They have to because they have such high valuations, they're raising a lot of money, and I think they have to go after more pieces of the software stack. Now whether they do that through partnerships, or acquisition, or organic development, they've got to do more than just the end-user surface, they've got to own some of the backend context and system of record, or what Karp calls the ontology, and I think their licensing terms have to enable that sovereign protection. And I think they'll evolve to that. Your thoughts on that?

Jeetu Patel

>> I agree with you that today, using agents is still too difficult. For the average layman, my mom's no longer around with us, but if she were, she was a very heavy enthusiast of the internet. You know, if one day her internet went down, all hell would break loose. I don't think she would have been able to use agents just yet. And I think we need to make sure that we get to that kind of barometer of, when do people that don't live and breathe AI start to go out and use this for their day-to-day? And when does it get to be superbly intuitive? Because we keep talking about why this is going to go out and solve every problem, and then you need a PhD to go out and use it. So we've got to make sure that that happens. And I think that's happening. We're moving at a pretty fast pace, but we're nowhere near the level of simplicity that's needed for eight billion people all going out and activating thousands of agents.

Dave Vellante

>> Well, I have another premise that I, you said you want to come in for a quick hit. I could keep you forever. We could do an hour podcast here. But so, there was an edict.

Dave Vellante

>> You should do that sometime.

Dave Vellante

>> I'd love to. There was an edict from CEOs early on, thou shalt do AI, and what's happened is everybody started doing AI. And so now, the last time we had a sustained mid to high single digit productivity boom was back in the 90s with PCs. And that was a bottom up, right? And that's what's happening now. It's happening, the productivity is going to happen from the individuals. And so the easier that we can make it for people to do that. But the risk is you get silos of intelligence. And so that's where the industry I think has to come together to really drive that bringing together of that tacit knowledge of the enterprise. That to me is the real promise of AI where it really starts to pay off. You saw Google yesterday announced earnings surpassed the 200 billion. the magic $200 billion CapEx mark. The street's freaking out today about that.

Jeetu Patel

>> But the reason they're doing that, these are smart people at Google. The reason they're doing that is because of this thing that we're talking about, which is if the proliferation of agents is still not completely universal, then you're going to need meaningfully more infrastructure in the future than what you need today. And so you're just going to need to make sure that you prepare for that. And so while in the short term you might see some ups and downs, the secular shift that's happening around AI is going to continue to persist. And we need to make sure that we keep that in mind as you have these kind of market movements that happen on a daily basis. Because the reality is you will see cost of tokens coming down, you will see the usage go up, you will see more open source models, you will see more closed source models, you will see better training that will actually be done on the models, and you will see more inference. And inference is only going to get to be a larger and larger percentage of the total spend of compute for organizations worldwide.

Dave Vellante

>> Well we appreciate you making this all more functional, all higher performance, working within the ecosystem. you guys are amazing ecosystem partners. And making it simpler over time and safer with more control for organizations who want sovereign AI stacks. Jeetu Patel, thanks so much for stopping by.

Jeetu Patel

>> Dave, I love your show. Thank you for having me on.

Dave Vellante

>> All right, take care. All right, keep it right there everybody. This is Dave Vellante, live coverage from Advancing AI at Moscone. We're right back right after this short break.

Jeetu Patel, Cisco

search

(INTRO)

Dave Vellante

>> Hi everybody, welcome back to Moscone West. You're watching theCUBE's live coverage of AMD's Advancing AI 2026. I'm Dave Vellante. We have a special guest here on our live coverage, Jeetu Patel is the President and Chief Product Officer of Cisco. Awesome seeing you, thanks for jumping in, I really appreciate it.

Jeetu Patel

>> It's so great to see you. The best ones are the spontaneous ones, man.

Dave Vellante

>> Yeah, you got it. First of all, Cisco's on a roll. we have seen you really focus the team. You're a builder, getting the product really to where it needs to be, and so congratulations on that.

Jeetu Patel

>> We have a great team.

Dave Vellante

>> Give us the quick update, and I want to get into the news.

Jeetu Patel

>> Yeah, the update is, I think, we're in a quiet period, but last quarter was fantastic. I think the general momentum of innovation has never been better than what it has been over the course of the past two years. I'd say the past two years we invented more than we had in the previous 20 years combined. And I think that's going to be dwarfed in comparison to what we have in store in the next couple years. So I think the team's excited. The only thing we're not getting enough sleep. So, we gotta every once in a while, take a Saturday off or something.

Dave Vellante

>> Well, and I know you're in a quiet period, so you can't say much. but Lisa laid out the TAM, the growth, the demand, where the demand is coming from, claiming that this is really the first year that inferencing is sucking up more compute than training, which we've been saying for a long time, it's finally happening, but there doesn't seem to be any end in sight to the demand. Certainly, the memory shortage looks like it's going to last through 27 and beyond.

Jeetu Patel

>> There's less than 1 % of the world that's using agents. Right. And if you believe that agents are actually a step function improvement from a chatbot, where you can actually have either your personal productivity or your company's productivity materially change as a result of agents. I don't see how you don't stay in a continued kind of supply shortage for a pretty long time period. And that's what I think we're seeing play out.

Dave Vellante

>> Yeah, the signals are very positive. We saw, notwithstanding IBM's disappointment, but that's really not, I don't think they're a bellwether for AI, Supermicro doubling its margins and obviously growing, you're going to probably see something similar from Dell, you're going to see something similar from NVIDIA, and clearly AMD is on a rocket ship. Okay, you got an announcement here, that you were up on stage, we're talking about distributed inference, take us through the news here.

Jeetu Patel

>> So what you're starting to see happen right now, Dave, is the pattern of inferencing is fundamentally changing as you move from chatbots to agents. So in the chatbot era, when humans were the ones that were driving the engagement, you ask a question, you get an answer. The patterns of inferencing were very spiky. what you're now starting to see in the agentic era is that these agents are working 7x24 on your behalf. They might actually go out and talk to other agents. And so the pattern of inferencing is much more durable and sustained and not as spiky. and actually the volume is higher. And so what you're starting to see is every customer I talk to right now has a couple major concerns. Hey, I'm concerned about security and safety, and I'm concerned about the cost of tokens, because these tokens can start to run up in cost, and you might not have had enough time to go out and budget for all of those. So we need to make sure that as an industry, we lower the cost of tokens quite dramatically. And so one way that that's happening is you're not just going to have inferencing run in the cloud. You'll also see a whole new set of models that emerge for inferencing running everywhere and locally. And one of those kinds of patterns is a whole new class of computing that's emerging called desk -side computing. So you work on your laptop right now, and that's a human working on a laptop. But you might have a desk-side computer where the agents are running. And by the way, this isn't really a new phenomenon. It's just taking more effect right now.

Jeetu Patel

>> I do have that, by the way.

Jeetu Patel

>> You do have that, and if you remember when OpenClaw first came out, Mac mini sold out.

Jeetu Patel

>> Right.

Jeetu Patel

>> Right? And you're starting to see that same pattern, which is like there's going to be more inferencing that happens on the desk side. Now, by the way, if you have enterprises that start to go out and deploy this, you can't just deploy and pray that everything works out right. You need to make sure that it's managed effectively, you need to make sure that it's secure, you need to make sure that the agent behavior is monitored, You need to make sure that you figure out whether the cost of tokens is out of hand. And if so, what do you need to do to make sure that you can contain the cost of tokens? All of that apparatus needs to be put in place. So what we have done is we have our unified management plane that you know well, Cisco Cloud Control. And what it does is it can go out and look at all of your inferencing capacity that you have in the cloud, all of the inferencing capacity you might have in a private data center, as well as your desk side computers, in a single management plane, and then be able to say, well, what do we need to do to ensure that you have the best cost efficiency and performance and security when you're running in this hybrid inference model, which is what I think is going to be the prominent model as we move forward.

Jeetu Patel

>> So you're doing the routing there?

Jeetu Patel

>> So in this particular case, AMD does a specific intelligent routing of the models. They have the actual compute. We provide the network bandwidth that's going to be needed. We provide the security and safety that's needed. We provide agent behavior monitoring so that if an agent's starting to go rogue, you can actually make sure that you can sequester an entire local machine. And we provide tokenomics, so you can figure out how many tokens are being used, is it more efficient than running it in the cloud, what do you need, which agents are using which amounts of tokens? Is that agent starting to show patterns of excessive token usage? If so, do you want to contain it? All of that kind of telemetry is provided in Cloud Control.

Dave Vellante

>> You guys provide that telemetry, that observability. That's fundamental to what you guys do there. Okay.

Jeetu Patel

>> We have early customers using it right now, but Cloud Control is going to go general availability in the US in early fall. And so we're excited about how customers can lower their cost, improve their efficiency for tokens.

Dave Vellante

>> It's interesting, during the pandemic everybody was working remote, they just threw everything to the cloud, it was spend, spend, spend. 2022 comes around, we're coming out of the pandemic and cost optimization was the big theme. Now you're seeing that token cost optimization.

Jeetu Patel

>> Because the volumes are higher, right? And the numbers are getting bigger. So most companies don't have a few hundred million just lying around that they can go out and start spending on tokens, they have to plan for it.

Dave Vellante

>> But at the same time, we've seen token costs come down 90 % a year. I know there's a big delta between the most expensive tokens that you're getting through APIs in the cloud from the frontier models and some of the Chinese open models. that's like orders of magnitude delta. what do you make of that? Ultimately, is this a short-term trend with this token cost optimization, or do you see this? you talk about it. Ali Ghodsi is really pounding the table on this. Do you see that as an ongoing concern? Or ultimately, are tokens going to be essentially free?

Jeetu Patel

>> The way I see it is it's an ongoing way that we will operate. Just like we have a finite set of resources with humans, and we have to plan out what those humans are going to be able to do. You're going to have a finite set of resources with tokens, and you're going to need to make sure that you can put those tokens to the best use so you can get the best outcomes for your organization. And so that's what's going to happen. And I think the reason that I don't think it's just going to be a short-term trend is because, yes, you might have seen the cost of tokens coming down by 90%, but the demand signal of how compute-intensive these agents are on infrastructure has gone up way faster than that. So your total cost is still going up within the organization. And that's going to be a finite bucket, and we need to make sure that that finite bucket is being applied in the best way possible.

Dave Vellante

>> So I want to run something by you. So the macro, if you look at spending on IT broadly across the world, organizations spend about four or five percent of revenue on IT. We forecast that that's going to double or triple by mid next decade. Why? Because you're essentially going to be scaling without the labor that you used to require to scale. We're not talking about wiping out jobs. That's really not the point. The point is you're going to shift investments to tokens, whether it's through APIs or you're generating, spending on CapEx on your own, and that is going to be where the ROI is. You're not seeing that big trend today yet. You're seeing the cost, the token cost. You're not seeing the productivity hits just yet. We're at the early days. But do you buy that premise that over time you're going to be able to scale dramatically without as much labor and that's going to justify that doubling or tripling of that revenue going to technology?

Jeetu Patel

>> So I have a slightly counterintuitive perspective on this one, which is that as AI is starting to advance and automate workflows and do things that are in a much more effective way than what was done in the past, more human bottlenecks are emerging than there were in the past. So I actually think you're going to see more jobs as a result of AI rather than less jobs. The only difference is the dexterity for AI is going to be needed in every human being. So anyone who's actually not AI fluent is going to have a harder time. But the people that are actually fluent in AI today, they're in more demand, they're making more money, they're busier than ever, and you're seeing this across the board. So I think you're going to actually see that being, so that's one dimension. The second dimension is, we have overly focused just on productivity. I think we need to focus on what class of problems could we never even dream of solving before, that we can solve as a result of AI. And I think that's a far more important question than just pondering what can we do to go out and reduce the cost for humans.

Dave Vellante

>> Yeah, so, and by the way, I would agree with that. I think we're going to have more jobs. I guess my point is, and I want to test this on you, we're going to be able to scale dramatically faster than we would in the past with a lot less labor. That labor is going to have tools and be more intelligent.

Jeetu Patel

>> Per unit output, you might have less labor, but you will have more jobs than what you had before.

Dave Vellante

>> I would totally agree with that. And when I talk about productivity, I have a very simple measure of productivity, revenue per employee. And so you're going to have much more revenue. you look at what Uber has done. Now, of course, they have a different business model. I got to ask you, it's interesting you said about only 1 % of us are using agents, building agents and using agents. We're in this bubble.

Jeetu Patel

>> We're in this bubble, especially in Silicon Valley. That's why I find it so important to go keep traveling to different parts of the world, because you want to make sure that you get a feel of how everyone's seeing this thing kind of take off, and it's not just within a very, very hyper -concentrated bubble.

Dave Vellante

>> So that brings me to, you saw, everybody saw Alex Karp's rant, which is quite interesting, and then you're seeing all the action around open weight models, of course, the Moonshot and what's happening with the Chinese open weight models. I saw somebody on TV the other day call it AI communism. It's not AI communism, in fact, it's the opposite. is that companies want to deploy open weight models so they can protect their IP, their alpha, and that's AI capitalism, but it's actually being driven by these open weight models, the likes of which China are really aggressive. Now, increasingly you're seeing US companies put out open weight models. My question to you is, it looks like certainly Anthropic and OpenAI are not capitulating, nor is Google, on closed models. and their argument is, look, we're going to have better integration, think iPhone. And I do buy some of that argument, but how do you see that playing out? Very clearly, what we published, John and I, is the power law of Gen AI way back when, where we said there's going to be a lot of these smaller language models and specialized models. That's clearly playing out. At the same time, I don't necessarily think it calls for the demise of these frontier models. You have a perspective on this. Your AI summit earlier this year was amazing. what's your thought on that?

Jeetu Patel

>> By the way, next year is February 2nd, so get ready.

Dave Vellante

>> Okay, I'm going to write that down because I don't usually travel if I can avoid it in February, but I would make a special trip for that because that was, I would say, the top event of the year. I would put it over GTC, which was amazing.

Jeetu Patel

>> We love Jensen, though, and I love Jensen's events.

Dave Vellante

>> Your thought leadership on that was fantastic. But what do you think about the prospects of the leading frontier models who right now are out of favor, they're spending all this money, but they have really smart people at the top. what's your take on their future? what do they have to do to really thrive and develop a business model that's durable?

Jeetu Patel

>> So let me take a step back, because this is a very important question, and I think we tend to get into very zero -sum discussions in this, and I think that's actually not productive. what I think is going to happen, I think firstly, what are the things that people look for as metrics to measure with models? They're looking for intelligence, right? per unit expended of cost. They're looking for cost itself, how do you contain the cost? And they're looking for control. Now, do you always want to use a frontier model for everything? No, that would be like taking a private jet to your corner store. You're not going to use a frontier model for everything. There will be certain things that make a lot of sense to use a frontier model. I think there's going to be intelligent routing which says when do you use a model that's going to be super expensive because you need that intelligence and horsepower, and when do you use a model that might not need that? So for example, we just announced we're launching a bunch of open weights models for a very, very specific purpose, which is finding vulnerabilities in code.

Dave Vellante

>> Awesome announcement.

Jeetu Patel

>> We just did that a couple days ago, thank you.

Dave Vellante

>> That's a tremendous announcement.

Jeetu Patel

>> And that particular announcement doesn't mean that we actually don't partner very deeply with Anthropic and with OpenAI. It just means that you're going to need to have SLMs that you want to have control over. And if you want to make sure that you try to find vulnerabilities in your code, you don't want to have that go in the cloud. You want to actually be able to do it right from within your environment. And if I can find 70 % of the vulnerabilities that way, great. And the remaining 30 % I can find with a frontier model.

Dave Vellante

>> So it's not a zero -sum game. And by the way, that announcement was outstanding. And I think the market is so confused right now. It took the market down, took all the cybersecurity stocks down.

Jeetu Patel

>> I do think, though, that there's going to be a very important dynamic that we have to make sure happens. It's America needs our own encouragement of open source. I think Nemotron is doing a great job at NVIDIA.

Dave Vellante

>> Thinking Machines just came out.

Jeetu Patel

>> Thinking Machines came out with one. And I think we need to make sure that we have more. We are coming out with small language models in domains that are very, very dear to us, where we have telemetry, we have a unique perspective. I think you're going to start to see more and more use of open weights models. You will see more open source come out, and you will see more closed source models. And what you have to do is get the cost of inferencing down, because inferencing is now the majority of the compute, as you heard from Lisa's keynote. But you also need to make sure that you're continuing to keep pushing the intelligence frontier, because we're nowhere near the top of the S-curve there. We need to continue to keep making sure we solve harder and harder problems.

Dave Vellante

>> I think there's another dimension of enterprise requirement.

Dave Vellante

>> Control.

Dave Vellante

>> You mentioned control, but why is it that we only have 1 % of us really driving agents? Ease of use.

Dave Vellante

>> I think it's too hard.

Dave Vellante

>> It's too hard, exactly. It's too hard. And that's why I want to bring ease of use into it. and I think this is the frontier model's opportunity. They have to make it easier. And Ray Wang's book, Everybody Wants to Rule the World, I think the frontier models want to rule the world. They have to because they have such high valuations, they're raising a lot of money, and I think they have to go after more pieces of the software stack. Now whether they do that through partnerships, or acquisition, or organic development, they've got to do more than just the end-user surface, they've got to own some of the backend context and system of record, or what Karp calls the ontology, and I think their licensing terms have to enable that sovereign protection. And I think they'll evolve to that. Your thoughts on that?

Jeetu Patel

>> I agree with you that today, using agents is still too difficult. For the average layman, my mom's no longer around with us, but if she were, she was a very heavy enthusiast of the internet. You know, if one day her internet went down, all hell would break loose. I don't think she would have been able to use agents just yet. And I think we need to make sure that we get to that kind of barometer of, when do people that don't live and breathe AI start to go out and use this for their day-to-day? And when does it get to be superbly intuitive? Because we keep talking about why this is going to go out and solve every problem, and then you need a PhD to go out and use it. So we've got to make sure that that happens. And I think that's happening. We're moving at a pretty fast pace, but we're nowhere near the level of simplicity that's needed for eight billion people all going out and activating thousands of agents.

Dave Vellante

>> Well, I have another premise that I, you said you want to come in for a quick hit. I could keep you forever. We could do an hour podcast here. But so, there was an edict.

Dave Vellante

>> You should do that sometime.

Dave Vellante

>> I'd love to. There was an edict from CEOs early on, thou shalt do AI, and what's happened is everybody started doing AI. And so now, the last time we had a sustained mid to high single digit productivity boom was back in the 90s with PCs. And that was a bottom up, right? And that's what's happening now. It's happening, the productivity is going to happen from the individuals. And so the easier that we can make it for people to do that. But the risk is you get silos of intelligence. And so that's where the industry I think has to come together to really drive that bringing together of that tacit knowledge of the enterprise. That to me is the real promise of AI where it really starts to pay off. You saw Google yesterday announced earnings surpassed the 200 billion. the magic $200 billion CapEx mark. The street's freaking out today about that.

Jeetu Patel

>> But the reason they're doing that, these are smart people at Google. The reason they're doing that is because of this thing that we're talking about, which is if the proliferation of agents is still not completely universal, then you're going to need meaningfully more infrastructure in the future than what you need today. And so you're just going to need to make sure that you prepare for that. And so while in the short term you might see some ups and downs, the secular shift that's happening around AI is going to continue to persist. And we need to make sure that we keep that in mind as you have these kind of market movements that happen on a daily basis. Because the reality is you will see cost of tokens coming down, you will see the usage go up, you will see more open source models, you will see more closed source models, you will see better training that will actually be done on the models, and you will see more inference. And inference is only going to get to be a larger and larger percentage of the total spend of compute for organizations worldwide.

Dave Vellante

>> Well we appreciate you making this all more functional, all higher performance, working within the ecosystem. you guys are amazing ecosystem partners. And making it simpler over time and safer with more control for organizations who want sovereign AI stacks. Jeetu Patel, thanks so much for stopping by.

Jeetu Patel

>> Dave, I love your show. Thank you for having me on.

Dave Vellante

>> All right, take care. All right, keep it right there everybody. This is Dave Vellante, live coverage from Advancing AI at Moscone. We're right back right after this short break.