How Ciena makes the whole internet work, with CEO Gary Smith

At present, I’m speaking with Gary Smith, CEO of the networking firm Ciena. You most likely aren’t acquainted with Ciena — the corporate isn’t actually a family title. However each web consumer has relied on the corporate’s merchandise; in actual fact, each web consumer has used its merchandise. 

Ciena makes the {hardware} and software program that make the fiber optic cables that join the world gentle up with knowledge. That’s every part from native fiber networks for broadband ISPs to the huge undersea cables that join continents. There’s a excessive chance that this very podcast got here to you over a Ciena-powered community, in actual fact — this firm is in all places.

Gary himself is an interesting character on this story. He joined Ciena in 1997, the yr the corporate had an astonishingly profitable IPO in the midst of the dotcom growth. And he grew to become its CEO in 2001, proper after the dotcom crash sent the company’s stock price — and the rest of the economy — into a tailspin. So Gary has had a front-row seat to the event of the web as we all know it, and Ciena has been there each step of the best way — as telecoms and undersea cables first introduced the planet on-line, all the best way to the rise of cloud computing and, now, generative AI. 

You’re going to listen to Gary and I discuss quite a bit about one thing known as WDM, or Wavelength Division Multiplexing, which is the elemental technical innovation Ciena is constructed round. WDM is completely key to how the trendy web works, and most of the people do not know it exists. The fundamental thought is that WDM makes use of a number of wavelengths of sunshine to suit extra knowledge onto a single fiber optic cable, which permits these cables to ship an increasing number of info over time — you’ll hear Gary name it “virtualizing” a fiber optic cable, such as you’re turning a single cable into two, or 5, or 10. That is clearly vastly vital because the world’s web utilization will increase and data-hungry purposes like streaming video and generative AI ramp up.

Ciena didn’t invent WDM, but it surely was the primary firm to deploy it commercially within the ’90s — and I needed to understand how Gary was managing his funding in pushing that expertise ahead and whether or not he was interested by the following type of tech that may disrupt it.

I additionally needed to find out about Ciena’s prospects. There are solely so many corporations constructing deep sea cables and large knowledge facilities, and it received’t shock you that the shopper set has largely shifted from telecom suppliers and ISPs to massive cloud computing corporations like Meta and Google and Microsoft over the previous few years. Only a few months in the past, Meta announced it might pay $10 billion to construct its personal unique subsea cable stretching almost 25,000 miles. Ciena sits proper in the midst of that sophisticated relationship between trade and infrastructure, and I believe you’ll discover Gary’s perspective to be fairly distinctive. 

We additionally obtained into the fracturing state of the web and what it signifies that individuals in numerous geographic places, depending on varied ranges of presidency management, expertise the net world in a different way. Connecting cables stretching throughout continents means the individuals on both finish of these cables might need very totally different concepts concerning the web, and Ciena has to take all of that under consideration.

See, I’m guessing you didn’t find out about Ciena earlier than, however virtually each single Decoder thought is correct right here, sitting on the spine of the web.

Okay, Ciena CEO Gary Smith. Right here we go.

This interview has been calmly edited for size and readability. 

Gary Smith, you’re the CEO of Ciena. Welcome to Decoder.

Thanks. Good to be right here.

I’m very excited to speak to you. The Decoder crew and I have been speaking earlier about how we love tales of hidden corporations that run the world, and Ciena very a lot appears like a kind of corporations. It’s a large participant in networking. It’s an enormous participant in the way forward for our networks. There’s some AI to speak about. There’s undersea cables to speak about, which is at all times enjoyable. Shortly, for everybody, describe what Ciena is and what you do.

Mainly, Ciena is the main expertise firm on the planet for offering high-speed connectivity. So if you consider all of the networks which have been enabled across the web and world wide, or, as you alluded to, submarine cables, Ciena is actually the driving drive behind that high-speed expertise. Even with cellular connectivity, terrestrial connectivity, and submarine cable connectivity, we’re mainly the most effective on the planet at transferring optical bits.

I at all times give it some thought as two internets, proper? There’s the web we expertise — the digital web —  after which there’s the bodily web: the precise cables and wires that join our networks collectively and join us collectively. That’s Ciena’s enterprise, operating these cables, determining the best way to transfer knowledge throughout them extra rapidly, and determining the best way to transfer extra knowledge throughout them. 

Ciena’s been round for a minute. How has the dynamic between what persons are doing on the digital web most customers expertise and what you must do on the bodily web modified? It looks as if one thing we don’t ever actually get to speak about. I’m very curious to ask you about it.

I believe we devour a lot of the web now in a cellular type, on our iPads, iPhones, or no matter. We nonetheless have some desktops, but it surely’s all laptops. The primary level of connectivity there’s cellular. It’s accessible to us on a regular basis. However what we don’t take into consideration is that it’s solely cellular till it’s not. It must get terrestrialized to the closest space the place it comes down into terrestrial —  sometimes fiber as a result of to get actually excessive pace, you must use fiber. Wi-fi has some limitations round that. So mainly, till it will get to any cellular utility, its actual goal is to get terrestrial into that cell tower as rapidly as attainable.

Then, from the underside of the cell tower is when all of the magic begins, with the bodily fiber cables and the networks that go world wide. You probably have an inquiry, relying on what it’s, that would require a distinct continent. That goes down by means of the cell tower and out throughout the ocean and again to you. However that’s invisible to us all. We’re in a position to do it at such excessive speeds now that it’s clear to the purposes that you simply’re utilizing.

That’s the half I actually wish to dig into. Intellectually, I believe I’ve at all times understood that. I believe most individuals perceive that there are servers and knowledge facilities elsewhere and we’re speaking with them by some means. However in some unspecified time in the future we have to hit a fiber line and really discuss to these servers. 

What I’m making an attempt to grasp is that there’s the buyer utility layer of the web — we’re all going to look at a bunch of Instagram reels or one thing. Then, there’s the web we have to construct to help that. What’s the connection between these two? How typically do you assume, “I see the following set of purposes which are approaching telephones, I see the following set of units that individuals might wish to use. or I see the rise of streaming providers. We higher begin constructing an web that may help it?” That dynamic typically appears subsumed in some cell service or knowledge middle conversations. It very not often involves the forefront.

As a result of it’s the piece that connects all of these things. Everyone focuses on issues like AI, Netflix, or no matter proper now. How do you get these issues to manifest themselves to the buyer and to enterprise? The fact is that it required an enormous build-out over the past twenty years for a fiber infrastructure that mainly connects all these knowledge facilities collectively, that connects them by means of submarine cables, and connects them throughout international locations. It has intelligence on it so it is aware of the place to go to get that info. That has advanced over the past twenty years. It continues to evolve as a result of if you consider a number of the enterprise fashions that we now take without any consideration, like Uber, that wouldn’t be accessible if you happen to couldn’t get immediate response to it.

Twenty years in the past, that may not have been attainable as a enterprise mannequin. There was an fascinating level within the early 2000s the place there was this large overbuild of fiber capability. It was type of the telecom nuclear winter the place we had extra capability than apparently we’d want for the following twenty years. Individuals constructed out all of those giant fiber networks and nobody got here. However that enabled enterprise fashions to take giant quantities of bandwidth at low price. That created enterprise fashions like Amazon, which weren’t in a position to get stimulated and be sustainable till that they had giant quantities of reliable connectivity. So it’s continuously evolving with purposes which are demanding a lot larger bandwidth. We’re in an period of that now with AI for positive. The networks must catch up and facilitate that.

Let me ask concerning the networks. Let’s say I wish to construct a community in some rural a part of the nation. Do I simply name Ciena and say, “All proper. We’re going to get to work. Go dig some trenches”? Do I run the community? Do you run the community? How does that work?

We don’t run the community. These are the service supplier carriers. In the event you take North America for instance, there are a whole lot of carriers that do the native final mile piece out within the rural areas. Then, there’s the Verizons, Comcasts, and AT&Ts of this world that then facilitate all of that connectivity to the buyer. We offer the expertise that powers their networks mainly. For them to get that high-speed connectivity, Ciena’s expertise goes on the ends of these fiber cables that enables the pace and throughput to proceed to drive the capability.

What we mainly do is virtualize a single piece of fiber. So as a substitute of getting that little thread, which is a few hair thread when it comes to weight and distance, we are able to multiply that many, many instances. That’s a whole lot of digital fibers due to the expertise we use round it. That’s enabled the entire scaling of the web. It’s enabled the scaling of this world connectivity phenomenon that we have now, and it’s on the middle of scaling for AI connectivity now and into the longer term.

That core piece of expertise the place you’re multiplexing and virtualizing the single piece of fiber, that was Ciena’s massive expertise guess. I imagine it’s known as WDM. Clarify that to individuals in a short time.

Mainly, it’s virtualization. The place you’ve obtained one piece of fiber, [WDM] makes it appear like a number of fibers. Relying on the space and what you’re placing down, it may be a whole lot of fibers, which is definitely bodily manifested as a single fiber. So that you don’t must put extra fibers into the bottom. With the expertise we have now on both finish of it, you’ll be able to simply scale it up. 

So, it’s the virtualization of fiber that has enabled this virtually seamless marriage between the purposes, to your level, and the driving of all of the community visitors with the precise community. Via our expertise growth, we’ve grown the power to go try this sooner than Moore’s Regulation. We now have a brand new piece of expertise, known as WaveLogic, that may do 1.6 terabits of information down a single piece of fiber.

It’s at all times been defined to me that the fiber is the fiber, and we’re simply sending gentle backwards and forwards. It’s the items on the finish that truly create the extra bandwidth to create that further functionality. Is that roughly true? Are you simply centered on the bits on the finish that discuss to one another?

Completely. There are expertise developments throughout the fiber that help that, for positive, and there’s been numerous that growth from corporations like Corning within the fiber world. However mainly, the magic is the expertise that goes on the top of these fibers. When you consider it, if we needed to scale by digging up the place each time we needed to place extra fibers in, it’s simply not sensible. We are able to get an incredible quantity of growth and digital capability even on previous fibers simply by placing our expertise on the top of it.

You have been speaking about scaling sooner than Moore’s Regulation, which is roughly a doubling each 18 months. Decoder listeners are very pedantic. They are going to inform me that’s not really what Moore’s Regulation means, however I believe that’s the way you’re utilizing it. I respect our listeners very a lot, however I don’t assume that’s what applies right here. You’re speaking a few doubling of bandwidth each 18 months. 

You’ve been doing WDM since 1997, proper? That is the expertise guess of the corporate. Is it the capability of WDM the place you’re utilizing a number of wavelengths of sunshine throughout a single fiber? Is that what’s doubling?

Yeah, it’s our potential to proceed to drive that. Once we first began, we might multiplex and maintain it quite simple. It was a single piece of fiber. We thought we might put 4 digital fibers down there. That was revolutionary on the time. Most of the carriers we went to with that mentioned, “Oh, that’s implausible. We’ll by no means want greater than that.” Then, we found out the best way to do eight, then 16, then 32, then 64. And we’re approach, approach past that now. 

So what which means is you might have that single piece of fiber that appears like 64 fibers. That’s enabled this large scaling of structure world wide. Then, we deploy that expertise not simply with native connectivity, however  throughout the nation and complete continents — 8,000, 10,000-mile submarine cables. We are able to help that type of capability throughout the submarine cables.

Do you simply promote these packing containers to the telecom and knowledge middle suppliers and stroll away? Do you might have ongoing help?

No, it’s a really built-in partnership. Many of the main carriers —  I believe 96 p.c of the foremost carriers world wide — are Ciena prospects. These are collaborations that we’ve labored on for a number of a long time. We assist handle the networks. We do numerous the set up work for them. We do the planning. We’ve the software program that truly manages the community and appears for predictive points or challenges and switches robotically. If there are issues on sure fibers, it’s going to detect that and change it by means of. So, there’s numerous underlying expertise and collaboration between us and all of the service suppliers. 

Additionally the entire hyperscalers. Just about the entire hyperscalers use Ciena to attach their knowledge facilities world wide. We companion with them as nicely on expertise developments which are particular to their networks. You have a look at the hyperscalers, a number of the largest networks on the planet, they usually’re very sophisticated, multi-continent, metro, long-haul, submarine cables. The largest gamers and the most important homeowners of capability on submarine cables now are literally the cloud gamers. That was once the normal service suppliers, however now it’s the likes of Google and Meta that actually personal numerous the worldwide capability on these submarine cables.

That is the factor that I used to be speaking about at the start. It appears so laborious to grasp, and it’s so fascinating to speak to you about it since you’re in it. The web we expertise is clearly altering the dynamics of the web we construct. So. I expertise my cellphone and there’s a bunch of stuff occurring on some knowledge middle owned by Google, Microsoft, or no matter, and which means these corporations now personal a lot of the capability of undersea cables. 

I don’t assume that chain of occasions is intuitive except you’re taking note of it. However simply speaking to you for 13 minutes right here, that appears very intuitive to you. It is smart to you that they might have the capability as a result of I’m not sending almost as a lot knowledge as a client throughout the ocean. Do you see that altering? How briskly has that grown? Describe that dynamic to me.

Take the submarine cables for instance. It’s a great proxy for what’s occurred right here. From the Eighties to 2010, this was all funded by the service suppliers. It was a approach of getting transatlantic cellphone conversations largely to Europe and just a little bit to Asia. These corporations, like British Telecom and AT&T, would have these cables, they might get collectively a consortium, and, as a syndicate, fund these submarine cables. With the appearance of cloud, you had these gamers that needed to have very high-speed connectivity. Take for instance Meta, Google, Microsoft, or any of those gamers. In the event that they have been going to be within the European market, they wanted to have very excessive pace, devoted capability. So in about 2010, you started to see the cloud gamers buying the capability of these submarine cables.

Within the final 5 years or so, you’ve seen Google start to personal many of those cables, with Meta and others too. So, it advanced from, “Hey, it’s all this service supplier and it’s simply cellphone calls.” Once you wish to attain your aunt in Europe, in India, or no matter, then it’s only a cellphone name, which is a really small quantity of information. The quantity of information that’s now being required by these cloud gamers is a number of extra instances than that. You’ve obtained numerous graphics going over it as nicely, which takes numerous bandwidth. So the necessity for bandwidth has grown dramatically. It is smart that the cloud gamers wish to join their knowledge facilities. In the event that they’re placing an information middle in India and wish to join it to North America, they wish to personal the connectivity for that, and it’s a a lot larger pace. In order that they’re not simply shopping for the capability. They’re really taking management and proudly owning the cables.

So, there’s been type of an evolution of that over the past 5 years. The primary 10 years was, “Hey, we’ll take the capability now.” Their networks are made up of a hybrid of proudly owning the cable, proudly owning some, or they hire it from the service suppliers. However you’ve now obtained, I believe, 600-plus submarine cables world wide. You’ve obtained about 1.5 million kilometers of fiber on the backside of the ocean facilitating this, and the cloud gamers are actually driving that market now.

That’s going to steer us into the inevitable dialog we’ll have about AI, however I simply wish to keep on this for one second. How has the verticalization, from “what we promote is virtualized compute capability” all the way down to “we’ll now personal the cables within the ocean and verticalize our enterprise and ensure we personal it tip to tail,” modified your organization? 

We’re , “There’s WDM in 1997. That is the core innovation that lets us get extra capability out of the fiber networks. We’re initially going to promote that to telecom corporations to do extra cellphone calls or the early web.” Now, a number of generations in, you’re promoting it on to Google to attach knowledge facilities in order that it’d run AI workloads.” It appears like that may have some stress in your roadmaps and your go-to-market. How is that expressed?

The evolution has been such that the primary appreciators for that expertise have been clearly the service suppliers. Because the web started to achieve traction in 2003, 2004, you started to see the cloud gamers get extra within the community as a result of in the event that they’re going to monetize and get to the eyeballs that they needed, the community really turns into vital to them. So, there’s been a gradual improve in appreciation of that even previous to the AI piece, they usually’ve change into way more energetic. I imply, we actually didn’t promote to the cloud gamers till about 10 years in the past, and we had zero market share there. Now, we’ve obtained the primary market share with a number one participant in connecting knowledge facilities and cloud gamers world wide.

Submarine cables are the identical type of factor. We actually didn’t deal with the submarine cables with our expertise as a result of there wasn’t that a lot capability being required. We’ve now obtained primary share in submarine cable connectivity world wide as a result of they worth very high-speed capability. So the cloud gamers have gone from being a really small share of our enterprise to over 50 p.c of our enterprise now being, immediately or not directly, cloud and hyperscalers. So it’s a really vital shift within the enterprise, and that’s actually accelerated within the final 5 years.

I believe this can be a nice time to ask the basic Decoder questions. You’re describing a shift within the enterprise and a shift in your prospects. How is Ciena organized now? What does that org chart appear like?

It’s been simpler for us in some methods as a result of we’ve at all times been the main expertise firm, so it’s actually about who appreciates that expertise and collaborates. This shift to the submarine markets and to cloud gamers has usually been a great match for Ciena as a result of we transfer in a short time, and we’re at cutting-edge expertise. The newest expertise we have now is definitely on three-nanometer chips. We have been the primary firm on the planet to have a 3nm chip really be used within the telecom house, and we’re now really engaged on our second. We’ve obtained that out, it’s carrying visitors world wide, and we’re engaged on our second model of that. We’ve at all times been at the forefront of expertise, and it truly is a pure match with these cloud gamers.

So, we’re organized very merely. We’ve a large-service supplier crew world wide, a devoted crew for the hyperscalers and cloud, and an engineering crew that works with them on co-development and stuff with that type of expertise. I’m an amazing believer in simplicity of organizational design. So, we have now a quite simple organizational construction that’s advanced to deal with that. That’s how we’re in a position to scale as rapidly as we have now.

You’ve obtained the 2 massive markets — the telecoms and the hyperscalers —  and you’ve got the engineering group. I’m assuming the telecoms and the cloud suppliers need various things. How is that expressed in what the engineering group goes off and builds? 

It’s fascinating. There’s numerous commonality about what they need. If you consider it from an engineering standpoint and the most recent WaveLogic expertise we simply developed, we really take that very same expertise and we deploy it in barely totally different variants to each markets. So, we get a greater scale of growth over that. Clearly, we develop a single chip after which do variants of it for particular purposes that they need, however that very same expertise is basically consumed throughout the service supplier world as a result of all of them need the identical factor: higher economics throughout the fiber that they’ve obtained and elevated bandwidth.

The actual sensitivity, I might say — what’s actually the distinction between the 2 markets — is the cloud gamers are way more centered on energy, which all of us hear about from an information middle standpoint. It manifests itself precisely the identical approach within the telecom networks. We do variants which are energy optimized for the cloud gamers, however that’s the predominant distinction. That’s simply taking the identical core expertise and doing a distinct variant of it that’s prioritized round energy. That allows us to scale the event and the remainder of it. So, we have now a single world engineering crew centered on this core expertise.

Ciena has at all times been a really centered participant. We’ve been very centered on simply high-speed connectivity. This has enabled us when most of our rivals are attempting to do cellular or other forms of issues. For the previous three a long time, we simply needed to be the most effective on the planet at transferring optical bits at a excessive pace. The chance we have now now could be more and more into the campus and doubtlessly even inside the information middle with that very same core expertise, and that’s all in entrance of us. We get zero income from there proper now. Nevertheless it’s the identical prospects, in order that’s an thrilling alternative for us sooner or later.

I wish to speak about that just a little bit extra, however can I simply ask yet one more query? You mentioned energy. That means to me that you simply’re constructing an information middle, you might have an power finances, and then you definitely maintain all of the distributors to some quantity on the road to be sure to hit your power finances. You’re feeling that too, even because the particular person is like, “Right here’s the field that connects to the fiber line.”

Sure. It manifests itself in a different way clearly, however everyone talks concerning the knowledge facilities and the ability and the remainder of it. Everyone may be very centered on that for apparent causes. However that very same problem deploys after they come out onto the community too. You place amplifiers on the fiber to do this, and round these amplification websites, they need tremendous low energy as a result of they’re constrained round energy and house.

I’m assuming they might somewhat spend their energy finances on operating the GPUs as scorching as they’ll. 

Precisely. They don’t wish to spend it on the communications half, or they wish to spend as little as attainable on that. They need it optimized. We’re in a position to do this for them when it comes to the footprint of the expertise. We are able to now get down on a single plug what was once a rack of cords. Fairly actually, a six-foot-wide rack on a single plug, that very same capability. That’s what we’ve advanced over the past twenty years with. So, we’re nicely positioned with these guys to proceed to drive down the ability utilization on the community, not simply within the knowledge middle.

How massive is Ciena? How many individuals are on all of those varied groups?

We’re near about 10,000 individuals worldwide. We’re very distributed. This yr’s consensus income is about $4.4 billion, one thing like that. We’re worthwhile, and we have now a powerful stability sheet. We very a lot make investments immediately into this expertise. As a result of we’ve been such a centered participant at it, we felt that this high-speed connectivity can be appreciated finally, and it’s actually coming into its personal now. It’s important to make these investments for the long run. A few of our applied sciences take five-plus years to evolve. We’re lucky on this time proper now. It’s a great intersection level the place high-speed expertise is tremendous appreciated.

Is that an excellent cut up? Is it 3,300 in engineering, 3,300 in hyperscaler sales-

No, it’s mainly predominantly engineering. As you’ll be able to think about, this type of expertise is extra about engineering, so greater than half of our crew are engineering expertise. We’ve obtained purposes engineers on the entrance finish who assist make sure that we perceive the purposes and work with the shoppers on that, however greater than half is engineering. As I mentioned to you earlier on, it’s a quite simple construction.

One of many issues that actually strikes me is there’s this core enabling invention, WDM. The corporate is constructed round it. Ciena’s made just a few acquisitions through the years. You’ve led just a few of them. There was one in 2019 for Centina Systems. Is WDM nonetheless the guts of it? Is that proprietary Ciena we’re constructing round it? Is there a subsequent approach to make use of fiber that’s coming?

No, it’s mainly the journey that we’re on. We’ve made a number of acquisitions, but it surely’s all been concerning the core mission, which is high-speed connectivity and the issues we want round that. Numerous the acquisitions have been round vertical integration of that, so we have now extra management over it. It’s not simply the core expertise. It’s the issues round it that you must do. Most of these surrounding applied sciences have been the place we might focus and supply aggressive benefit by doing one thing to them. 

Over time, we’ve labored very fastidiously at acquisitions and natural developments that make us vertically built-in so we are able to management that. There’s benefits to manage and scale, however what actually drove us was that we expect we are able to put aggressive benefits in every of the applied sciences that we collect across the core expertise, if that is smart.

The rationale I’m asking that’s as a result of typically you would possibly wish to purchase an organization that’s early to the following flip, but it surely looks as if WDM continues to be the guts of what you’re doing, and also you don’t see a end line there. 

Mainly, it’s this expertise. The precise variant that we deal with is the coherent expertise, and that’s what has enabled this DWDM to scale so much. However there are different applied sciences round it that we’ve developed, just like the SerDes — the Serializer / Deserializer — and the sorts of issues which are for analog conversion, which is extremely tough to do. We’ve obtained a number of the greatest expertise on the planet round that. It’s the moat that you simply put round that, but it surely’s not the following factor. We expect there’s numerous headroom, significantly as a result of this expertise is all deployed proper now. Give it some thought exterior the information middle. In order quickly because it desires to current exterior the information middle, that’s Ciena.

The alternatives we have now with the identical sorts of expertise at the moment are changing into obvious contained in the campus — these campuses are getting tremendous massive now — and finally can be inside the information middle. That’s as a result of the  GPU and compute funding that’s occurring requires larger and better pace over longer and longer distances, even in a few of these knowledge facilities. To start with, they’re monumental. You’ve obtained fiber going round them, typically a number of kilometers inside a single knowledge middle with that fiber getting wrapped round. 

We expect there’s an intersection level for our expertise there. The expertise there proper now, which is simply known as direct-connect, is required to attach the entire racks throughout the knowledge middle and the intra-data middle. The physics of that’s getting challenged as these GPUs require larger and better pace. So, we expect there’s an actual alternative for us to intersect that, to your level, with the identical type of core expertise that requires some variations over the following 1–3 years.

Let me ask you the basic Decoder query, after which I wish to get into that. You’ve been at Ciena for a very long time. I believe you’ve been there for 30 years. You’ve been the CEO for nearly 25 now. How do you make selections? What’s your framework?

Relies on what sort of determination it’s, proper? Not all selections are equal. I very a lot really feel that it’s a crew recreation. Once we make strategic selections — which I believe is what you’re  alluding to with how we make investments and what long-term issues we’re centered on — we’re strategically aligned round what we wish to go do. We’re a really centered firm. Again to the purpose I used to be making earlier, I believe individuals respect the most effective of breed, and we wish to be the most effective on the planet at what we do. The north star may be very clear to us. Then, it’s about what issues we have to do to get out into the longer term on that.

From that standpoint, the choice making is considerably simpler as a result of we all know we’ve chopped the sector, so we all know the place we’re enjoying on. There’s not some nice dialog about adjacencies and issues that we would wish to get into; it’s all primarily based on our core expertise. To reply your query, if it’s one thing strategic round these investments, it’s at all times a crew recreation. We’re tremendous collaborative. It’s how we work and get cross-functional enter into all of these issues. So, it’s a course of. 

However I might say that I believe the position of a CEO typically is to say, “Okay. We’ve had the controversy. That is the general perspective of what we wish to do. Let’s go.” I believe a part of the position of the CEO is to facilitate the most effective conversations and discussions after which assist drive to a conclusion.

Let me offer you a hypothetical that I believe you’ll shrug off, however I’m going to strive anyway. You’re two massive teams of purchasers. You’ve obtained your telecom corporations, and also you’ve obtained your massive cloud suppliers who’re all racing into AI. You have got some alternative to develop that enterprise by going into the information facilities, not simply touchdown the fiber on the doorstep. That looks as if the a lot richer alternative. 

I have a look at the state of the telecoms in the USA, for instance. They’re not rising a lot, proper? They’re churning prospects between them. It will appear very apparent that your complete crew would say, “Don’t fear about them a lot. We have to put all of our deal with the AI suppliers as a result of that’s the place there’s development utilizing our identical core expertise.” However then you definitely nonetheless must service the telecom prospects, I think about. How do you stability that out? How do you make that call to say, “These investments are going to go over right here?”

It’s really a lot simpler for us as a result of, again to the purpose I used to be making, it’s the identical type of core expertise. Once we transfer the ball down the sector for the cloud gamers, we are able to apply that to the service supplier house as a result of it’s the identical core expertise. By way of their wants, there are literally numerous similarities. One other mind-set about that is after we begin speaking concerning the knowledge middle and the remainder of it, we offer optical methods the place we do the entire thing for the service suppliers. And we try this just about for the cloud gamers too. So we do the entire methods for the software program, what we name the WDM modem expertise, and the road methods. We do all of that, and sometimes we’ll set up it, help it, and keep it. We try this for service suppliers and we try this for the cloud gamers.

Then, to the purpose across the alternative within the knowledge middle, it’s not essentially about offering these full methods. It’s taking them, placing them in plugs, after which placing them in elements. So it’s a distinct type of play, but it surely’s nonetheless the core expertise. So, something you do to advance that additionally applies to our methods enterprise, and our methods enterprise is each service supplier and within the cloud. You don’t must make the “that are your favourite kids?” type of alternative, which is the place you have been going with that. I believe that’s one other good thing about an organization staying tremendous centered on its core expertise and understanding what worth we convey to the world. That may be very tough for anyone else to copy,

Effectively, can I ask you about that? As a result of it feels just like the Trump administration is incoming. By the point individuals take heed to this, he may have been inaugurated and we’ll be off to the races with a brand new type of enterprise local weather right here in the USA. Mergers can be in trend once more. They haven’t been in trend over the previous 4 years. If I used to be operating Google or Microsoft, I might say, “Effectively, that is the core expertise that’s going to make my knowledge middle extra environment friendly and transfer knowledge sooner on my cable. I’m simply going to purchase this firm.” Have you ever been approached to be acquired?

Effectively, I’m not going to–

I’m simply questioning as a result of I can see why over the previous a number of years, all of that dialog has chilled. Nevertheless it’s additionally outstanding to me that an organization like Ciena has remained unbiased for 30 years.

That was once the dialog on a regular basis. Take into consideration 20 years in the past. I’m unsure a few of your viewers will bear in mind a few of these names: Lucent, Alcatel, Nortel, even Nokia. The mannequin was all what I name the “generalist mannequin.” They have been really spin outs from their cellphone corporations like Marconi. Northern Telecom got here out of Canada. So with the mannequin, we needed to be finish to finish. You needed to put switches, routers, handsets, cellular infrastructure, and this factor known as transport. We mentioned after we began, “No, we’re going to only deal with the transport piece as a result of we expect it will possibly get to scale.” 

What’s occurred over the past twenty years is that the generalists, as a result of they couldn’t afford to be on the innovative in every of these areas, have all struggled. You hint the genesis of all of these corporations now and mainly they’re all in Nokia. With Alcatel and Lucent and all the remainder of them, the music stopped and that’s the place they ended up.

The purpose I’m making about that’s the world’s too aggressive now for the entire end-to-end factor and being a generalist provider. It simply completely is. They used to say to us, “Effectively, when are you going to get purchased?” We used to say, “Hear, we’re a public firm. We’ve all of the fiduciary tasks that you’d anticipate with a public firm, however we expect we are able to get to scale and be an actual firm simply doing what we’re doing.” Now, these items take some time and we’ve been public for 25 years, however we’re at some extent now the place we completely have scale. We’re very robust financially. We are able to put money into this type of expertise. We’ve obtained the primary market share in all the important thing markets that we wish to play in. So, we’ve obtained a great alternative to drive shareholder worth from right here.

Do you assume that DWDM, the multiplexing you do to virtualize the fiber, is sufficient of a moat? You have been describing the opposite acquisitions you’ve made and the way you encompass the core expertise with different issues to enter totally different markets because the moat. However is that core expertise one thing that… I don’t know if Mark Zuckerberg awoke one morning and was like, “screw it, I’m open sourcing it,” he might simply do it? Or is that extra specialised?

There are different corporations which have variants of this expertise, for positive. It’s simply that we’re in a position to implement it sooner and faster and higher than anyone else, and we have now the core competencies round it. A little bit of an analogy is TSMC. Individuals type of know the best way to do [what TSMC does], however they’ll’t do it, proper?

You possibly can go to ASML and ask for a machine and that’s not the place it stops.

Yeah. As a result of it’s individuals, and it’s the parents that perceive that it’s the 5, 10 p.c of every of those areas that actually make the distinction to issues while you put it and all of the issues we’ve constructed round all of it collectively. So like with any of these items, it’s a confluence that makes up the aggressive benefit. It’s while you get all the way down to the actual applied sciences. 

It’s all of these issues that present the moat for us. It’s the truth that we’re centered and the truth that we’re now so embedded with these key cloud and expertise corporations and have been a part of the collaboration course of across the evolution of their networks. It’s the relationships that you simply construct up, that trusted relationship and the expertise collaboration over time. That’s why we expect we’ve obtained a chance to take that into a complete new marketplace for us with the campus and finally inside the information middle.

I wish to speak about two issues that appear like they’re going to occur to the hyperscalers, particularly, over the following 4, 10 years. One in all them has quite a bit to do with subsea cables and the opposite one is clearly AI. Let’s begin with subsea cables. If I look broadly at our nation’s massive corporations — Amazon, Google, Meta, Apple — they’re about to enter right into a world the place an emboldened Donald Trump desires to only reset the worldwide order. They personal cables that join our continents collectively. There’s lots of people who wish to break the web aside in numerous methods. We might ban TikTok. We’d not ban TikTok. China doesn’t enable our apps there.

There’s numerous curiosity in what is occurring on our networks that join our international locations collectively and who will get what the place. Then there’s simply the fact that they’re wires within the floor. Unhealthy issues occur to them. They get lower. There are wars. How a lot of that do you interact with personally? How a lot do you consider, “Okay. There are some fraught geopolitics which are about to vary. I must make some selections to get forward of it, concentrate on it, insulate myself in opposition to it, or insulate Ciena in opposition to it?” How a lot of your time do you spend on that?

I imply, there are the headings of issues you’ll be able to management after which issues you can’t management however should be conscious of, proper? It’s positively in that second checklist. I might argue that over the past 5 or 10 years we’ve had various geopolitical dynamics occurring. We’re conscious of that, significantly as a result of we function in a number of international locations, so 80 international locations world wide. Our expertise is broadly deployed. We’re conscious of these sorts of dynamics, and I’ll offer you an instance.

We decided within the early 2000s that we might not go in and be a supplier to the Chinese language marketplace for its main service suppliers. There was a complete set of thought processes that went into that. One in all our rivals on the time was Huawei and it was simply rising. I don’t assume individuals respect this, however the first actual iteration of expertise that Huawei got here out with was really in our house, so approach earlier than all of the cellular stuff and all the opposite all-powerful issues that it’s now doing. So round 2003 we mentioned, “What’s this firm, what are they doing, and the way is it attaining value factors which are simply decrease than our price?” Seems that we’re being sponsored by the Chinese language authorities, proper?

It’s like we’ve seen this film a number of instances since 2003. 

So one of many thought processes that we had, simply to present you an illustration , we mentioned, “You realize what? We is not going to compete with it immediately within the China market. It clearly desires us there as a result of we’ve obtained main expertise and the remainder of it. We’re really not going to fabricate in China both. So we’re going to be the antithesis, if you happen to like, of Huawei in each single approach. We’re going to be the choice that’s not a Chinese language producer, and we received’t function there.” 

That was a really controversial factor. You think about a younger firm that’s simply gone public and going, “Grasp on a minute, why are you not going into the biggest market and quickest rising telecom market on the planet?” However we had the power of our convictions. It’s taken some time, however I believe largely that was a great determination for us as a result of we didn’t get distracted by that Chinese language market.

They’ll offer you 10 p.c of that market, however they’ll be sure to received’t get the rest, and also you’ll companion with a neighborhood companion. I spent numerous time in Asia. I lived in Asia for quite a few years earlier than I got here to Ciena. And I noticed that film, to your level, play out in so many various methods. I used to be tremendous delicate to that, as was the crew. So, that’s an occasion of us doing one thing due to the geopolitical dynamics on the time and us making a choice. You don’t understand how that’s going to play out as a result of issues can change., You’ve simply obtained to be sensible round what you assume is the most effective factor for the shareholders and your organization and keep on with it.

Do you principally comply with the lead of the purchasers? Do you might have your personal factors of view on geopolitics? I’m considering of issues like while you land the cable in Europe and your units there decoding all of the visitors that there are a variety of European spy companies that may wish to sit proper subsequent to that. 

The NSA in our nation would love to take a seat proper subsequent to that. Once you land in different international locations which have totally different sorts of privateness and safety issues, they could wish to be part of that. Is that one thing the place you say, “Look, Meta owns this cable. You discuss to Meta. We’re going to do what the consumer says,” or is that, “We’re merely not going to construct these cables in sure units”?

We’re really fairly lucky that we don’t must make a few of these robust selections as a result of it’s not our cable. So to your level —

However the knowledge going over the cable, that’s you, proper? You’re encoding and decoding it?

Yeah, that’s what we’re doing, however even that expertise is owned by the shopper. We bought that to the shopper, and what they do with it’s fully as much as them. We don’t have management over that. It’s their community.. It’s their property. You concentrate on these submarine cables. Somebody owns the cable, and we put the expertise on the top of it. It’s their expertise. They personal that, they usually run it as nicely. We don’t run it. In order that they run these networks. We’re lucky in that we’ve at all times been in a position to keep one step behind a few of these dilemma questions. It’s not one thing we have now to opine on.

When you consider the precise cables, they’re in worldwide waters. What’s the longest one that you simply function?

I believe it’s about 10,000 kilometers. A few of these cables are across the Pacific the place you get the large distances, so they’re very, very lengthy cables. Numerous them cease off on totally different islands and locations and spur off. A few of the new cables getting in are spurring off in other places. We’ve additionally developed a expertise that overlays that gives for resilience. So if you happen to get a lower on a specific cable, then we are able to reroute it relying on the structure. 

It’s known as GeoMesh. It’s software program and expertise that’s embedded into our core choices and permits us to offer resilience throughout a few of these submarine cable hyperlinks. Within the regular run of issues, you get I believe one thing like 100-plus faults a yr, primarily trawlers dragging their anchors on a fiber cable as a result of they’re buried on the backside of the ocean till they’re not. I imply, they’ve obtained to return up someplace. Sometimes, it’s within the shallower waters the place you’ll get some fishing vessel come throughout it.

Do you ever get offended sharks? Do you ever get a textual content message like “shark assault introduced down the community?”

Really, that’s a standard query I get requested round it. There’s a part round fish assaults. I believe it’s zero. I’m positive there are some examples of it that somebody can level to over the past three a long time, however I’ve by no means heard of it personally. I believe they’ve obtained different issues to fret about than these little tiny… it’s concerning the measurement of a giant hosepipe, so I don’t assume it’d be significantly fascinating to them.

We’ve this phrase that we’ve been utilizing at The Verge, my pal Casey Newton got here up with it ages in the past. He calls it the “Splinternet,” the place we understand one cohesive web, however actually we’re fracturing. We’re fracturing into continent-specific, country-specific internets. 

As I talked to you concerning the cloud suppliers, it’s very apparent to me that we now have quite a few personal internets working at excessive pace world wide which are simply owned and managed by single corporations. You have got a protracted view of the web. You might be one of many few friends on Decoder who’s ever simply straight up talked about Alcatel. You have got a protracted view of the web, how we’ve grown it, and what it’s for. The telecoms, by and huge, have been nationwide champions, proper? That’s how they’ve thought of themselves. Even in the USA, that’s how they’ve thought of themselves. That’s the place they got here from.

There’s a brand new type of nationalism in America round our tech giants within the Trump administration. Mark Zuckerberg says, “I need the Trump administration to go fight the European regulators on my behalf.” That’s new, it hasn’t come up earlier than. We’ll see if this jacket matches. He’s making an attempt it on. However the best way we construct networks is an expression of nationwide functionality. We haven’t skilled that an excessive amount of because the web. It’s been one world web that connects all, after which we do no matter we wish and set our borders. However we haven’t really cut up the networks aside. However now you’ve obtained massive cloud supplier personal networks. You do have rising nationalism world wide. You simply have a pulling aside of the web. Is that affecting how you are constructing networks or how your purchasers are asking you to construct networks?

I completely perceive the notion, and I believe what you’ve obtained is {that a} disproportionate quantity — that is type of an English understatement — of the innovation and growth for all of this has been accomplished within the US, amongst all these expertise corporations, for positive. Now, the large cloud gamers are investing in that community as a result of that’s their enterprise mannequin, and the community is changing into an increasing number of vital. Once we discuss concerning the community, to your level, we speak about it on this homogeneous type. 

But when you consider this within the context of a number of the giant markets that the US cloud gamers are attempting to get into… take locations like India. Once they enter the Indian market, they’re dependent upon connectivity from the native carriers. To allow them to tip up with the submarine cables and the touchdown stations, however then it’s, “How do I get related to all of the individuals I must get related to in India enterprise and client?” They’re then depending on the native carriers, be it Bharti, Jio, Vodafone Thought, a number of the wholesale carriers which are additionally there, and fiber. There’s at all times going to be an interrelationship as a result of as all-powerful as all of those cloud gamers are, they’re not going to have the ability to connect with everyone’s dwelling throughout the planet. So that you’ve at all times obtained this interaction between service suppliers and the cloud gamers.

What typically occurs there’s that in a few of these international locations, the service supplier will present a particular community for the cloud participant, they usually name it a Managed Optical Fiber Community, or MOFN. And we assist facilitate that. It’s our expertise, so it matches all of their world networks. Nevertheless it’s owned and operated by the native service supplier to help Meta, Google, Microsoft, and many others. 

More and more, I’d argue you’re seeing that interdependency work. So, I wouldn’t say it’s the counter-argument to the concept we’re all splitting up as a result of on the finish of the day, to get to the customers world wide, these service suppliers will not be going away in that regard. That’s not going to vary. They’re the one ones that may present that connectivity in the end, be it the enterprise or the buyer.

I’m fascinated by this as a result of it does appear to be the general geopolitical dynamics of the world are altering, and the web will change with it. I simply maintain coming again to the query I began with, which is about how the expertise we have now on the digital web clearly impacts the bodily web that you’re liable for constructing. I’m at all times looking for the touchpoints between the 2. The best way we architect networks in India has a direct relationship to the Indian web that these customers expertise, and it looks as if these issues are, if not diverging utterly between totally different international locations, definitely starting down totally different paths. That, to me, appears completely fascinating.

I wish to finish with AI right here. As you consider getting into their knowledge facilities and saying, “We are able to envision some new excessive bandwidth, excessive capability workloads in these knowledge facilities,” which for the previous 15 years we have now probably not provide you with large new ones exterior of streaming video. Streaming video is quite a bit, however that’s mainly what we got here up with. Now there’s some new ones in AI that may hit all types of latest industries. 

Do you assume that that’s simply an exponential rise in how a lot capability we want, how a lot energy we’ll want to make use of, how a lot networking capability that we’ll maintain? Or is it, to make the comparability earlier to darkish fiber, constructing out numerous capability that we’ll finally use over time?

That’s an amazing query. Having lived by means of the earlier interval of the telecom nuclear winter the place we constructed out all this capability — and because it turned out, nobody got here — it did facilitate all these enterprise fashions that wouldn’t have accelerated with out that glut of capability. I get requested the query if we’re in one other bubble round all of those knowledge facilities. Particularly across the knowledge middle piece, I believe there’s an insatiable demand for compute, to construct out compute energy. This stuff at all times take longer than all of us assume to monetize and the remainder of it, so we’re most likely going to undergo some iterations. However to reply your actual query, I’m really of the other view this time spherical. I believe we’re underestimating the quantity of visitors and bandwidth that can be required to go do that as a result of we’re already seeing a clip up in cloud visitors. We’re fairly nicely positioned to see that.

Now, how a lot of that would I attribute to AI? A few of it’s, however I’ve obtained no approach of actually discerning which bit is which, and it’s not our community. However there’s a particular uptick in cloud visitors simply to connect with all of those knowledge facilities. We’re completely seeing that, and I believe that’s going to proceed since you’ve obtained much more knowledge facilities coming on-line within the subsequent 18 months, or two years world wide. In the event that they’re going to monetize any of that, they’ve obtained to all be related for studying, coaching, and the remainder of it.

That’s earlier than you even see the purposes for medical use or verticalization, no matter you wish to name it. Utilizing your ChatGPTs is producing visitors, positive. However then you definitely begin to get to any of that medical utilization, for instance, and the quantity of graphics which are required. Graphics are about 10–20 instances extra bandwidth-consuming than regular knowledge. So, if you happen to begin to get to the monetization of that by means of these vertical purposes, I believe we’re underestimating the quantity of visitors that’s going to be there.

Once you plan out development for the corporate — clearly, you’re a public firm, so it’s a must to do it out in entrance of everybody and speak about threat — in your most optimistic view, is it sustainable, the speed of demand that you simply see now? You’re saying we’re underestimating it. I have a look at it perhaps a bit extra cynically, which is, “Wow, we’re constructing numerous capability in opposition to enterprise fashions that don’t exist.” OpenAI has but to show $1 in revenue, and it desires the entire energy that the world can generate. It desires each GPU that Nvidia can construct. That’s quite a bit, proper? There’s an inherent promise there that this may all repay. 

Perhaps you’ll name {that a} bubble or perhaps you received’t. I’m simply it extra mathematically. Numerous corporations are going to promote numerous issues to those knowledge facilities. We’re going to construct numerous capability, and perhaps in some unspecified time in the future we may have constructed sufficient or not sufficient or one thing will occur, however that may come to an finish. Then an organization like yours is saying, “Effectively, we’ve had this charge of development and now perhaps it’s plateaued or perhaps it’s happening.” How are you interested by that?

I give it some thought like this: Don’t confuse compute capability with the community. The community is at all times the afterthought to all of this, proper? It’s, “Oh my gosh, I do really want to attach these knowledge facilities.” So what’s the hierarchy of their thought course of? It’s, “I want the compute. I want the ability. I want the house.” These are the issues which are driving their determination. Then by the best way, “Oh my gosh, I want to attach these knowledge facilities proper now.” I’m being flippant, however it’s down there. I’d wish to inform you it was proper up there, but it surely isn’t. Within the reply to this query, I believe that’s really a great factor. We solely get there after they want it. They’re not constructing out these monumental capacities as a result of we see it. We’re placing capability on the market, they usually’re placing visitors down it.

So I take actual consolation from that, having lived by means of 2002 and 2003 when our revenues went down from something like $1.6 billion to about $300 million in one year. I’ve nonetheless obtained the scars from that, so I’m significantly delicate to this. This isn’t the identical dynamic. They’re really placing capability out, and as quick as we put it on the market, visitors is coming by means of it. To your level and to what numerous people are saying, is it going to get overheated when it comes to the GPU stuff and the information facilities and the remainder of it? In my private opinion? Completely. Will there be a little bit of a tail off in all of that? Completely. Will it take longer for purposes within the AI world to actually generate the economics? Completely. However we’re not banking on that. We’re simply seeing this good regular development in bandwidth, which I believe will pervade all through.

Right here’s one other approach of answering your query. Over the past twenty years, bandwidth has grown about 20, 30 p.c a yr and isn’t cyclical. When you consider it, it shouldn’t be cyclical, ought to it? We’re simply placing an increasing number of purposes out, we’re utilizing our telephones for extra stuff, extra persons are getting telephones, and all these units at the moment are related. That has been very regular and constant all through. Even while you get an financial downturn, it continues at about that tempo. What we’re speaking about right here is the appearance of AI on prime of that. Precisely what’s it going to be? Nobody actually is aware of. There’s all these fashions, however nobody actually is aware of. Nevertheless it’s affordable to imagine that it’s going to return on prime of that and will push us to 30, 40 p.c bandwidth development per yr, and I believe that’s going to be fairly sustainable. When you consider all of the promise of AI, the actual trick to it’s you begin placing any graphics down it.

Each time you hear [Google CEO] Sundar Pichai say multi-modal AI, I can see the greenback indicators floating by means of the places of work over there.

I like it. We’re mainly planning our enterprise on what we see, and we’ve taken our development charges up. Sometimes, this enterprise has grown. We’ve grown about 7 or 8 p.c. We’re now saying it’s going to develop 8 to 11 p.c over the following three years, and that’s with none large entry into the information middle house or the rest. It’s simply what we see from a cloud development standpoint.

You’ve talked quite a bit about getting into the information middle utilizing the identical core expertise that you simply use for the undersea cables and the telecom suppliers. What are the limitations to doing that and what have been the stuff you assessed while you thought, “Okay. That is the market we have now to deal with”?

The limitations to entry there are that you simply’ve obtained a expertise known as direct-connect, which has been nicely ensconced there actually for the reason that begin of communication in knowledge facilities. You’ve obtained an ecosystem there that’s going to be designed to elongate that for so long as attainable, and the shoppers wish to maintain it for so long as attainable. With the analytics we’ve accomplished on it and dealing with them, they assume that there’s going to be an intersection level the place the physics of that will get to the purpose the place it’s not sustainable. We expect we’ve obtained to the purpose with energy and house the place there’s a great intersection. So it’s the analytics round that and dealing with two or three of those gamers who’re already our prospects however exterior the information middle.

Is that them coming to you or are you going to them?

A bit of little bit of each. We’ve robust relationships with them, in order that collaboration has been there. We’re reasonable about it. The prevailing approach of working and that direct-connect? That’s not going away. That’s most likely nonetheless the vast majority of purposes for it. So that you’re what the subset of corner-case purposes are for this type of high-speed connectivity. However even a small a part of that’s large for us and big from a scale standpoint given the quantity of spend that’s gone on in that house. It’s at all times about taking dangers when that market isn’t there proper now, however if you happen to wait till that market seems, it’s too late. In the event you’re ready for, “Oh, what’s the market share of this and what’s the expansion charge?” On the finish of the day, you’ve simply obtained to go along with what your instinct is and what you’re listening to and seeing from the shoppers and the remaining.

So we’ve obtained our first iteration of expertise for this popping out on the finish of the yr. That’s a guess that we positioned two, three years in the past. As you recognize, the timing takes some time with a few of these chip designs and stuff. It’s important to guess the place the market’s going to go, not the place it’s proper now. There’s at all times a component of threat to it. Hear, having gone by means of a few of this earlier than, you’re by no means going to get it utterly proper. You simply must be in the correct vector, concentrate, and tweak alongside the best way round it. So we’re listening intently.

Gary, this was an unimaginable dialog. You’re going to have to return again someday. I might discuss to you about this for hours and hours.

Like to. I loved it, Nilay. Good questions. Great things. 

We’ll see. We must always have you ever again a few years from now and see how harassed you’re about these undersea cables. I believe that may be very fascinating.

We’re definitely in for some leisure, aren’t we?

I do know what we’re writing about for the following 4 years. Gary, that is superb. Thanks a lot.