Nile Makes Enterprise Networks AI-Ready

Hey guys, thanks for the intro. We're here with Sasikaran, who's the Chief Marketing Officer for Nile, and we're having a little chat about, well, what's going on with networking and in general infrastructure these days. Because, well, it's getting harder to predict how much we're going to need, and budgeting is becoming a major challenge, and there's all kinds of other interesting things on the horizon.

But Sashi, welcome to the show. Well, thanks for having me, Mike. All right.

So what do you see people struggling with these days? Because a lot of the AI workloads that are coming out and being deployed seem to me, at least, to be much less predictable, and it's getting harder to kind of figure out what to budget for and how much I'm going to need of what when. But what are you seeing in here?

Well, I think the last 12 months or so have been a nightmare for many CFOs across enterprises that we've spoken to. And a large portion of that is due to, I guess, the uncertainty that's crept in into budget and forecasting, and the unpredictability and the volatility that we see in pricing structures across the industry. And I think some of these kind of started out maybe 12 months ago when we saw these tariffs start to come in.

That was sort of the first shock effect into the supply chain ecosystem because there was a lot of unpredictability across different countries. And then we saw the AI tailwind start to pick up, particularly in the enterprise. And the AI value chain is extremely interesting because any one single component can actually hamper the entire ecosystem's ability to come around and deliver value.

We saw things like memory companies skyrocket, and supply was very constrained. And so we've seen different vendors come out and be unable to hold pricing for more than a week sometimes. And these were quotes that typically were about 90 days in duration from a validity perspective, but now you're getting a week, or that price is no longer there.

And if you look at traditional enterprise planning cycles, typically a budget is allocated to IT over the course of the year, and they have to plan around their activities. And not only did we see this unpredictability and volatility, but you also saw something that was going for, let's say, a dollar suddenly shoot up to $4. So your ability to go procure that thing meant you had to defer on procuring something else.

And so we've seen all of these interesting dynamics happen, and I think many companies are still navigating their way through this chaos right now. So the cost of the underlying equipment has gone up and tends to vary widely. At the same time, there's a bit of a perfect storm because the demand for throughput is also rising in ways that are difficult to predict.

So I may need to add a new router or switch here and there along the way, but I don't know what it's necessarily going to cost me. What's the challenge there for if you're in charge of IT, what are you supposed to do about this? Is this just kind of the way life is, or are there things that I can do to kind of mitigate some of this?

Yeah, look, I think a lot of those have to really rethink on what their traditional modus operandi is and see how do you kind of go down a path where they have greater control over their destiny. And you see some companies getting ahead of it in terms of rethinking their architecture, their supply chain, and things like that. Some of the things that we have seen is, with this fallout is, clearly, if you're looking at investing in AI, some of the investments have gone into the data center.

And that meant they have had to starve out other areas of their infrastructure across the rest of the enterprise, which means delaying upgrades, deferring refreshes. And this has a negative impact on maybe security issues that creep in if your upgrades are not all on time or if you're out working with equipment that needs to be refreshed. So some have tried to stretch the dollar by cutting corners where, in the long run, it hampers them.

And others have poured in more investment. And each have their pros and cons. But I think the longer-term value is in about thinking more holistically.

And what we have seen is also now the war situation in Iran. If you think at it from a global perspective, a lot of these companies are really interconnected. That has also caused additional issues that are impacting the cost of goods as well.

And many companies are thinking about shifting from CapEx to OpEx because anytime you want to refresh, if you have to come up with a boatload of investment, and that investment is 4x what you had budgeted for just six months ago, then clearly you have to go and cut corners. So one of the tailwinds, at least, that we're seeing is a lot of customers are rethinking CapEx versus OpEx, and they're rethinking whether they need to own the infrastructure, or they look at infrastructure being delivered as a service. And then it's also a question of human capital.

The more you need to buy the infrastructure and have people manage it manually, there are a lot of things that creep up that can lead to inadvertent errors. So how do you run these autonomously? So these are some thoughts that are surfacing as patterns and are top of mind.

So people are maybe taking a harder look at the total costOf running these networks, and they're now just looking at one thing because some parts of that I may not have any control over, so I got to go focus on the things I can control. Is that about the right assessment? Yeah, I think that's one way of looking at it.

Because typically in a normal scenario, you see that every $1 that's invested in CapEx is about $5 to $7 in terms of OpEx over the life cycle of the infrastructure. But if your $1 of CapEx is now becoming $4, the only way they can keep their budget intact is by starving operations or making it much more efficient. So I think that's really the organic evolution that we're seeing in terms of rethinking how do you make operations more efficient, because you can't defer CapEx for inadvertently long period of time.

So you have to look at autonomous operations and maybe shrink the OpEx cost lower. So how autonomous can things get at this point? Of course, everybody and his brother is talking about AI, but when it comes to networking, what is realistic in terms of what I should expect to be able to automate?

Well, the good news is a lot can be automated. Automation is not new to the industry, and if you look at it, there are varying forms of automation that have been coming along for the past two decades, and I think each is building on top of the other. But the more important thing now is not just going to automate a bunch of different widgets, but what I was alluding to earlier, which is rethinking what does the architecture not require today.

And if you can simplify the Jenga blocks, if you will, because in enterprise, infrastructure has been built block by block because it's all brownfield environments, and nobody has gotten rid of the previous complexity. They've just stacked on top of it, and they've tried to automate that complexity. And now with AI becoming a force multiplier there, you are sort of amplifying the positivity or negativity of the underlying infrastructure and its capabilities because applications have to run on that infrastructure.

And so I think it's a great equalizer for many organizations now as they come and adopt AI to rethink, especially as they go through refresh cycles on whether the architecture on which they're making their employees and their applications more efficient and productive, is that the right architecture. Do we still need to have protocols that were needed maybe two, three decades ago, or are there some things that are not required today? And that's sort of the architectural discussion that we are seeding as well, particularly in branch and campus environments, which is where 70, 80% of the workforce really are engaged in.

Data centers are more machine centric, but then you come into branch and campuses, this is where users, devices, it's more people centric, and you want them to be much more productive. Unfortunately, there we see the infrastructure still dating back from the '90s, the protocols from the '90s, and this is the Jenga stack that I was talking about. And simplifying that and running AI is really a huge force multiplier, both for IT as well as a productivity multiplier for the employees and partners.

How do you go have that conversation? Because I think it's one thing to say we need more bandwidth, so we need to buy another switch or another router. It's another thing to show up and say, "Hey, we need to rethink the entire architecture of this thing," and that's a bigger upfront cost, so how do I explain that to people in a way that will get them on board?

Well, the beauty of it is it isn't bigger upfront cost because one of the things that we are, as I said, seeing the tailwinds about is this whole notion of as a service. So when you think about before cloud happened, everybody was buying software, installing it on a server, floppy disk, CD-ROMs, you name it, then upgrading the licenses, patching it themselves. Then came the cloud, and you were able to take software as a SaaS, buy it, and consume it as a subscription, and you didn't have to incur your own server costs, if you will.

And I think it's the same way now that we're seeing with infrastructure and particularly networking, which has been lagging behind. So instead of you owning all of the switches, you investing your IT team to go manually configure it, maintain it, operate it, troubleshoot it, can you consume your network as a service? Can you not pay for all the costs up front, but on a subscription basis?

And a provider like Nile, for example, takes care of all of that headache for you. So you don't have to deal with infrastructure upgrades. You don't have to deal with the security issues.

You don't have to deal with the upfront CapEx investment that you need to make. So these are some of the big anchor roadblocks that many people have not known that there is a better way. And so I think the perfect storm that you and I talked about in terms of budget crunches, AI coming in, infrastructure supply chain bottleneck, escalating price volatility, these are all causing triggers for people to think, is there a better way?

And so I think that's really where we're seeing us having conversations aligned with, hey, is an as a service a better option for you? Is a subscription OpEx model better for you? Is something which is architecturally baked in with security natively instead of use tacking security and AI differently better for you?

Is autonomous operations better for you? And the answer to a lot of these appear to be yeses now, far more than what you see maybe a few years ago. Does this conversation depend upon what's happening in maybe a given vertical industry?

Because I think in some industries, they're cash rich, and they seem to prefer to spend capital versus consume operating expenses. In other places, things are a little tighter, and maybe the operating model makes more sense. Or do you think eventually everything is just going to move to as a service?

Look, I think it is a function of the business reality at the end of the day, right? And there will always be organizations that will look to procure equipment, and they prefer having depreciation write-offs on their books. And those are companies that are functioning a certain way.

But when you are dealing with not one or, but rather three or four scenarios of unpredictability, the overarching business guidance will always to bring predictability, bring certainty. " Right? Because it's always a risk to reward.

" But in reality, they are the ones that are guiding many organizations to say, "Look, there's so much fluctuation, volatility. We're putting cash up front, and then we are investing in people to manage it. " And they all realize the shift towards autonomous is happening.

The degree varies, but that shift is certainly hap- Last question. What is the relationship going to be between the networking people and the rest of the IT crew in this age of as a service and OpEx budgets and... Because there's other things that go into that thinking.

So is that all going to be mellowed together in one budget, or are we still managing all this stuff in our various silos? Look, I think the silos are blurring to a large extent. Typically, if you look at it, the network budget has been about 7% to 10%.

Obviously, the operational costs vary a bit. But increasingly, networking and security are intertwined, so you can look at it as network security. So these two previously siloed functions are coming together.

And any application is only as good as the network it runs on and how it manages to be breach-free in some sense. So I do think there is a lot more, and you see this top-down as well, in many organization where the CISO was previously reporting to the CIO, now it's a peer-level reporting structure. Or in some cases, it's a function where they're looking at the IT, which is usually the spend and the governance with the CISO, being brought in with board oversight.

So a lot of these, especially with AI now being a natural trigger, it's forcing the rethink of a lot of these traditional boundaries. Because AI is becoming the new democratic neutralizer, which is able to cut across different functions. And it amplifies the positivity or the negativity that each function has in the way that I see it.

It's like holding a mirror to yourself, and AI becomes this reflection that is either portraying the positive or negativity that you see. And so I think many companies are using that to take a fresh look at everything. And certainly, the network becomes a control point.

You don't want it to be the least common denominator and impact the higher value investments that are made in other aspects of the company. So I do see the strategic relevance of the network being amplified over the next few year as infrastructure investments get a fresh look. Well, folks, you heard it here.

All these things are related. It starts with a simple increase in the price of memory. It starts to cascade all the way up through the networking infrastructure, and before you know it, you got an entirely different model for running your network.

But hey, the good news is it might be easier at the end of the day.