AI-Native Networking – Make Every Connection Count

Daniel Newman:
I’m Daniel Newman, your host, founder, and CEO of The Futurum Group. Very excited for this conversation. I have a returning guest, someone I’ve had on a few of my different shows over the years, Rami Rahim. He’s the CEO of Juniper Networks, and today we are going to be talking about Juniper AI-Native Networking, making every connection count. Rami, first of all, thank you so much for joining me again. It’s good to see you. It’s been a minute.

Rami Rahim:
Thanks so much, Daniel. It’s a pleasure to be here.

Daniel Newman:
How are things progressing overall at Juniper?

Rami Rahim:
Well, things are going really great, quite frankly, Daniel. We have been leveraging artificial intelligence now to drive our business and customer results for over five years. Learning from data, telemetry, deriving insights, and then using that insight to deliver incredible value for our customers.

Daniel Newman:
As an analyst on the outside looking in, it’s been very exciting to watch all the innovation in the space and of course to watch the innovation you’ve been leading over at Juniper. Now, you’re not a recent appointment in terms of your role at Juniper. You’ve been there, what, I think you were employee 32, Rami.

Rami Rahim:
32, yes.

Daniel Newman:
So you’ve been through a lot.

Rami Rahim:
Yes.

Daniel Newman:
I want to talk about the future, but before we really dive deeper into this, into what you just shared, give me a little bit of the rundown, give the audience a little bit of that kind of rundown of some of the innovation and these different periods that you’ve been through since your time and being the 32nd employee at Juniper.

Rami Rahim:
Yes. So Juniper itself was a big bet on IP being the protocol of choice for the internet, and also is an architectural bet on the need for purpose-built silicon that was required, we believe back then to enable the internet to scale to what it is today. That was a massive inflection, and honestly, had we not made that bet correctly, I wouldn’t be here having this conversation. Juniper wouldn’t exist as a company. And since then there have been a number of really big inflection points in the market.

Obviously mobile computing, cloud computing, which essentially separated the consumers of infrastructure and applications and where those applications and the infrastructure is actually hosted. And that essentially stretched networking and made networking even that much more important. Software defined networking was an architectural shift that enabled compelling new solutions like our missed AIOps capabilities that essentially takes control and management of a networking solution and puts it into the cloud.

And of course, we’re just beginning on this journey of artificial intelligence, and my prediction is that AI is going to be bigger than all of the other previous inflections combined. And it’s great that we’ve already been investing in it, learning from those investments and deriving real value for our customers and achieving financial success with the early stages of AI just over the last few years. But I think the best is yet to come.

Daniel Newman:
To your point, how does AI become bigger than every inflection in the past? Well, it starts to become part of everything that we do across technology. And so that has kind of left every person in your role, every CEO on the planet asking and answering a more complicated question, and I’d love to hear what you have to say. But okay, so everything changed around November 2022 after the generative AI boom, and every company now needed an AI story. What’s the big problem that you are solving for Juniper now that AI became the rotation to the epicenter of focus for every tech company and basically every business on the planet?

Rami Rahim:
Right. It’s a great question, Daniel. I kind of view it in two different directions. There’s first AI for networking. This is the use of artificial intelligence and in particular AI operations to augment human operations, to make running networks super easy, less expensive, more power efficient, and also to delight the end user with an incredible experience. That’s what AI for networking is all about, but that’s not where the opportunity stops.

There’s also networking for AI, and this is around the recognition that AI doesn’t just appear out of thin air. You need the processing power to enable the accumulation of information, the learning, and ultimately the processing necessary to derive incredible value for practically any AI application out there. That can’t be achieved with just a single GPU, as you and I know. You need to do this by basically clustering hundreds if not thousands of GPUs together.

That is a high performance networking problem. And this is where the need for not just industry leading software, but also silicon capabilities that can keep up with the immense performance capabilities necessary to build these clusters of GPUs. So those are the two dimensions that we are investing in to capture the full AI opportunity that’s in the market today.

Daniel Newman:
Yeah, there’s all of this inference that we want to do, and of course we hear all about the training that’s required, which is kind of the really cool LLMs and everyone is focused on that right now, but in the end, this data has to move. It has to move in the cloud, in the branch, in the edge network, on the devices. These are going to be really interesting problems to solve, and the explosion is exponential. This is not small orders of magnitude. This is going to completely revolutionize the world, and that’s kind of where that multi-trillion dollar spend number probably came from. But let’s get a little bit more particular, Rami, about what you’re doing at Juniper. So in January you launched a new solution, you call it the AI-Native Networking Platform.

Rami Rahim:
Yep.

Daniel Newman:
Talk to me a little bit about what that means for the business and the impact you expect it to have on your customers.

Rami Rahim:
Yeah, we’re taking a page out of the success story that we have seen with missed AIOps in the Campus and Branch, and essentially making it into an end-to-end AI-Native Platform that runs across every networking domain. And as I mentioned, Daniel, the success is not new for us. So if you look at our mystified AIOps powered solutions, they grew 70% in revenue 2022 to 2023, and the momentum remains incredibly strong for us.

So we believe that by taking these capabilities and expanding them across the data center, the wide area networking, even up into the application layer, we can see similar types of results in our business from doing that. Why? Because our customers have seen the light, quite frankly. I’ve seen them go from AI skeptics when we initially introduced our Campus and Branch AI capabilities, into AI believers as they started to see the immense value that this delivers to their business. And I suspect that… I believe that this is going to happen also in these other domains, in the data center, at wide area network as well.

Daniel Newman:
Yeah, I don’t believe there’s room for deniers anymore. I’m just going to be really straightforward about this.

Rami Rahim:
Yes.

Daniel Newman:
I think there are varying levels of enthusiasm right now, and I do think that we are in a catch-up phase. I think you’ve probably heard yourselves and some of your partners and competitors have talked about the backlog of now implementing. ’23 was the year of selling AI chips and infrastructure, and now is the year for many of your partners in the world’s biggest SIs to get it integrated. But I also like what you said, Rami, you said something about AI being everywhere. And look, I mean our business, even as an analyst, we used to have someone that was like the AI analyst.

And our new perspective is there’s no such thing. AI is everywhere. It sits on top of every technology discipline. It’s AI and silicon. It’s AI and cloud, it’s AI and networking, it’s AI and… You know what I’m saying? So I think the way you’re going to have to do business is AI becomes a core part of everything you do. So let me ask you that. Are you going to be able to flip and get the market to buy into this idea that Juniper is an AI company now?

Rami Rahim:
Yes. Well, it’s a great question, and not only do I believe we can do this, we have already done it in the Campus and Branch. We are true leaders in demonstrating to our customers the power of AI and transforming the experience of both the network operator and the end user. I don’t need to prove to you, other than to just show you the financial results that we have achieved with our AI powered solutions, that this is something that is a reality today. But I also believe we’re just scratching the surface and the opportunity for us is to expand that capability to every networking domain.

And that’s what we’re doing imminently in the data center, eventually in the wide area network. And then finally, as I just mentioned, this is going to become a key part of leveraging our pedigree, our history of being able to build high performance, highly resilient network solutions and applying those solutions to the AI cluster data center. That is a massive multi-billion dollar, very high pace, high growth opportunity that’s for grabs today. And we’re investing to capture more than our share.

Daniel Newman:
Yeah. Look, there’s something that you said there about the proof, and of course, as someone that watches markets very closely, everybody out there wants to see the clear revenue inflection, meaning how is AI adding, how is it net revenue expansion? But I think there’s two parts of it. There’s almost the table stakes aspect that you have to first prove that AI is core to your solution, and then you have to prove that the market’s willing to pay something more. Whether that drives up margin, drives up productivity, drives up the total number of skews and demand. And of course for you drives market share.

Rami Rahim:
Yeah.

Daniel Newman:
And so these are all things that of course, we’re watching, and I think the proof does come in the numbers, but it comes in many formats. And I think the market’s still really trying to digest what formats that should come in other than just this is how much AI revenue is in this particular business’s numbers. Now, having said that, you also indicated that you see this networking opportunity. A lot of the focus on AI today has been about compute. It’s been on the infrastructure especially about compute and on the business apps and on all the LLMs, it’s really more of an application. But what about networking? I mean, it sounds like you’ve come to the realization that AI can completely and fundamentally change networking. Explain what you mean by that.

Rami Rahim:
Well, there is a number of different things. As I mentioned, the internet was the first big bet that resulted in the birth of Juniper Networks, and I believe the internet is the greatest vehicle for human advancement and innovation the world has ever seen. It turns out that connecting people and things in scales of billions is a very useful thing to do. It’s also a very hard problem to solve, and we’ve solved that incredibly well.

AI is going to be even bigger for networking than what I just described for the internet. And the reason is that the processing power that’s necessary to achieve these unbelievable outcomes that we’re seeing, result from AI in practically every single application that you and I can imagine, as well as applications we haven’t even imagined before, requires more than just one GPU, more than just a dozen GPUs. It literally requires thousands, if not tens of thousands of GPUs connected together in a high performance way that can operate as a collective single system.

That is a networking problem more than anything else. That is the opportunity that is before us today. Today, that market opportunity for networking is actually not that large. It’s measured in the few number… Billions of dollars, but the growth rate is immense. And I’m referring here specifically to the growth rate of ethernet networking for clustering of GPUs, for processing of AI workloads, either learning or inference. A massive opportunity with huge growth rates that we’re investing to capture.

Daniel Newman:
So I love to ask those sitting in your seat. I say, what makes your AI special? Like, okay, every networking company right now, every cloud company, they’re all saying something. What’s the Juniper… Like this is what makes Juniper and its AI story special or different.

Rami Rahim:
Yeah, so I love answering that question because I do acknowledge there is a lot of AI washing out there and a lot of lofty claims made by our peers, our competitors in the industry. And I would just ask whoever’s evaluating to see what is real from what is fluff and PowerPoint slides, is look at the results from actual customers. So Gap Inc reduced the number of truck rolls to their stores by 85% by leveraging our AI operations.

ServiceNow achieved a 90% reduction in trouble tickets in their network operations. Fasthosts, an 85% faster network service provisioning as a result of our artificial intelligence. So I’ve just always point those who are evaluating and trying to distinguish between the fluff and the real to look at real customer outcomes. And we can point our prospects, our customers, to real examples across the board around the world.

Daniel Newman:
So let’s sit in the chair of your customer for just a minute, because AI is changing their world. It’s changing the businesses fundamentally, whether it’s the network partners you have, the enterprises, ’cause you’ve had a big shift in your business to being more enterprise and cloud centric. How are you seeing it playing out for your customers?

Rami Rahim:
When I talk to IT professionals, CIOs, CTOs of our customers and our prospects around the world, I think there are a couple of recurring themes. First and foremost, it’s just insufficient talent, right? There’s a talent shortage of skilled IT professionals that are out there. And second, it’s budgets. These IT professionals are being tasked with achieving strategic growth for their companies, keeping the disruptors out of their business, digital transformation initiatives. And that requires resources, it requires investment that they’re just not seeing at the pace that they need.

At the same time, 70, 80% of IT staff today is focused on just keeping the lights on, making sure that the network is just running smoothly. And I think that the answer to all of the above problems is to augment the IT professionals that exist today, the humans with robots, and that robots is in the networking, just software that’s running artificial intelligence to essentially alleviate all of these problems. This, I believe, is why there are so many skeptics that are transitioning to believers right now. There are no better alternatives today. That’s the challenge, but then it’s of course the big opportunity for us.

Daniel Newman:
Hey, Rami, I want to thank you so much for joining. Going under the covers a little bit about the AI journey and the story. It’s a big opportunity and I expect Juniper to continue to be a big contributor to the innovation in AI and networking. And I hope I’ll be able to have you back soon.

Rami Rahim:
Thanks so much for having me on, Daniel.