What's Up with Tech?

How Megaport Built A Thousand-Data-Center Backbone In Seconds-Driven Software

• Evan Kirstel

Share episode

Interested in being a guest? Email us at admin@evankirstel.com

What if spinning up a 100G connection took less time than making coffee? We sit down with Megaport’s CEO Michael Reid and technical evangelist Alexis Bertholf to explore how a fully automated, software-defined backbone turns global networking into a click-and-go experience across 1,000+ data centers.

We trace their unconventional paths from aerospace to IT, then dive into the engine room: why no one logs into devices, how code drives every change, and how that design makes it possible to provision cloud on-ramps, private VXCs, and global WAN in under a minute. Michael shares how the team jumped from 10G to 100G services backed by 400G backbones, launched an automated internet product that exploded in demand, and built an AI exchange connecting next-gen inference providers to the places latency matters most. Alexis breaks down complex ideas with practical clarity and tells a standout story: a space company facing a launch deadline messaged her on TikTok when its carrier fell short—and Megaport delivered in time.

We also tackle the biggest myth in cloud: leaders assume connecting to cloud is easy. The reality is that rigid telco contracts and long lead times turn networking into the long pole. Megaport’s answer is flexibility at scale—automated, diverse, and instantly adjustable—so teams can pivot as workloads shift, AI features roll out, and regions change. We unpack edge networking strategies using distributed compute stacks to host SD-WAN and security functions near cloud on-ramps, landing local traffic onto a private global backbone for consistent performance and sub-150ms AI voice targets.

If you care about multicloud networking, SASE, SD-WAN, AI inference, and turning connectivity into a reliable product instead of a risky project, this conversation offers a clear blueprint. Subscribe, share with a teammate who owns the WAN, and leave a review with the cloud myth you want retired for good.

Inspiring Tech Leaders - The Technology Podcast
Interviews with Tech Leaders and insights on the latest emerging technology trends.

Listen on: Apple Podcasts   Spotify

Support the show

More at https://linktr.ee/EvanKirstel

SPEAKER_02:

Hey everybody, really excited for this uh fun and entertaining informative chat with Megaport, who are empowering a new generation of elastic connectivity and more. Michael, Alexis, how are you? Good to see you both. Uh really a big fan of yours, both of you professionally, as well as the mission and vision at Megaport. Maybe we'll start with introductions. Michael, as as the CEO of Megaport, uh give us a bit into your history and the massive expansion that currently is underway.

SPEAKER_01:

Yeah, world domination uh on a mission four. Yes. No, very quickly, uh actually Alexis and I share a slightly similar background. We're both aerospace engineers uh by by trade, I suppose, and then sort of failed at that to get a job in the industry from my side because uh I'm Australian and there was no aerospace industry in Australia. So you that's you fail into a thing called IT. These days you can take a degree and get into IT. Uh then I ended up at a company called Cisco in Australia for some period of time, running banking and finance, ran the northern region of Australia. Then I moved to the Bay Area and ran acquisitions for Cisco for a long period of time, uh, for about five years, taking on a whole range of different companies, scaling them through the business. Uh, ended up with a company called Thousand Eyes, which was really cool. Tech, for those who are listening, you might be familiar with Thousand Eyes, and then was the CRO for that business for a few years, scaled that massively, and then ended up as the CEO of Megaport and relocated back to Australia, but spend most of my time, as you know, on a plane back to the United States. So uh, yeah, most of our businesses in the United States, something like 60 something percent. But our head office and engineering team and HR and finance is in Australia in Brisbane and we're publicly traded on the Australian Stock Exchange. So that's me.

SPEAKER_02:

Brilliant. And Alexis, I'm a big fan of yours as a technical evangelist at Megaport and a content creator, engineer turned content creator. Um, how do you describe yourself these days?

SPEAKER_00:

I like to pretend that I'm still an engineer. Just throwing it out there. Um, but similar to Michael, studied aerospace engineering out of college, uh, tripped and fell into IT. I was lucky enough to get into an accelerator program with Cisco. So I spent a year um down in RTP taking my C CNA, CCNP, and doing all my technical training. I got pushed in front of customers shortly after that. And I spent about five years as a Cisco SE. And uh in the middle, towards the end of COVID, I started posting content online just about the things my customers were asking me uh under the premise of if one customer has this question, I'm sure more do too. And what I realized is there's a lot of um people out there that don't have dedicated support, right? And still have these questions and aren't getting them answered. So that's kind of how my channel or channels got rolling and it snowballed into a full-time job. Um, I was lucky enough to be adopted uh by the Megaport team. I was so lucky I actually got to write my own job description, which was a really fun adventure. So I filled it with all of the things I love doing. And I I actually just crossed my year anniversary with the company um last week. So very happy to be here.

SPEAKER_02:

Wonderful. Well, congratulations. And again, big fan, as are many of us in the telecom and networking world. So before we dive into what's next at Megaport, let's uh watch a quick video together uh from your website I thought might inform the audience you're on. What's the big idea at Megaport?

SPEAKER_01:

Megaport's spent the past decade building out over 850 data centers across 150 cities in over 25 different countries, and stitching together the world's largest network as a service platform so that in less than 60 seconds you can choose your data center, create a port, connect into our backbone at up to 100 gigs, and do that in less than 60 seconds. And there you have it, you're now live inside the Megaport backbone, and we can take that connection, and now you can choose to connect to cloud providers, private connections, megaport internet, our internet exchanges, or in our incredibly large marketplace. Let's do a cloud. We'll take AWS, choose a hosted connection, and you want to pick your country. You can pick any location around the world. Let's go into the United States, let's choose Ashburn. You have a choice between different colours which give you diversity zones. Let's choose a blue into Ashburn, and then you give it a name. Pick a speed, we'll go right up to 10 gig into Amazon Web Services. Choose your term. We could do month to month all the way up to 36 months and beyond. Choose your Amazon Web Services account ID. That's all you need to pick, and then click add. And now what you can see is in less than 60 seconds, we've connected to the Megaport backbone at 100 gig in that data center, and you now have a 10 gig connection across the US into that Amazon Web Services on-RAMP. Let's go back in and do a little bit more. Take that same port and now build out what we call Global WAN as a service. Click a private VXC and we can choose Australia, which is where my accent's from, and in fact, where Megaport was founded here in Brisbane. Let's choose the Brisbane location. We'll give it a name, pick a speed, jump through, and all you need to do is click next and add that connection. And there you have it in less than 60 seconds. We've built out a global WAN high-speed connection, private, all the way over to Brisbane, Australia, where the company was founded. There's so much more you can do with Megaport. Please take the time to check out our website or talk to one of our reps anywhere in the world.

SPEAKER_02:

Wow, Michael, it's an extraordinary accomplishment what you and the team have built. And you know, when you look at that network at video today, what makes you most proud of how far you've come?

SPEAKER_01:

Well, we crossed a thousand data centers. So it's funny listening to that video. I think I said 850, and I can't remember when I recorded that. So that might have been a couple of years ago or a year and a half ago. Don't know. But anyway, we must have added 150 data centers since that video was recorded. So we've been, I think you open, we've we've been on a real mission to scale the company globally and just land in as many data centers as we can to service as many customers as we can. Uh, so I think you know, being a being a being a uh a tech platform that's actually scaled to deliver global services is really it's it's quite an incredible feat. Doing that from Australia is always a sort of a rare thing. There's only a few, I think, companies that that have actually gone and done that. And so that's that's always really exciting. For me, it's just this sort of hyper-scale uh journey, you know, we're we're building out more and more and more. I think Alexis probably heard me. What's the strategy? The strategy is world domination and and and more. And then people say, Well, more of what? I'm like, more of everything. Uh the only thing actually get more of is Alexis's. There's only one of them. Uh and you know what? I became famous uh inside Megaport. I walk down the street now, and people say, Hang on, are you that guy? And I'm like, Yeah, that's me. And they're like, the guy that hired Alexis. I'm like, Oh, yeah, yeah, that's that I'm that guy.

SPEAKER_02:

And uh Alexis, I mean, you you make network engineering sound fun and uh in interesting and entertaining. You take it out of the textbook and you know, you put it in audio and video and written word. Um, what's the secret to keeping, you know, both tech, you know, tech content accurate and watchable and you know, entertaining?

SPEAKER_00:

Yeah, I've made this comment multiple times and no one seems to believe me. I use myself as a filter. I I really do think I am the dumbest person in every room, and I have to simplify things the way that I do so they make sense in my own brain. Like there's no secret formula, there's no every you know, couple words you need another hook, or there's frame like it's what am I trying to get across? What do I need to personally remember about this topic? And how do I put it in my own words so that I remember it later on? Like I am just the filter. Um, and it seems the way that my brain breaks it down, other people also enjoy. Um, so that's been cool too.

SPEAKER_02:

Yeah, and Alexis, I mean, the platform that we we we saw on the video is really built on instant connectivity, which sounds almost um magical. Uh what's the wildest or most unexpected use case you've seen a customer pull off with this platform?

SPEAKER_00:

Um the most unexpected one. This this is the one that comes to mind when you say that is we actually had one of the popular space companies that was um, they had a launch going in, it was like a time-sensitive launch the next week. And I got a DM on TikTok that said, Hey, Alexis, I saw all of your your content about Megapor. I'm actually an engineer at X company. And we have this problem where we need to transfer data between two of our data centers. We need this done next week, and our telco can't get a link up. Can you help? Um, and we were we were able to help. So that was pretty cool, especially because it was marrying almost my space roots. And also it came from TikTok. So that was fun.

SPEAKER_02:

Oh, that's such a such a fun story. And, you know, Michael, you you've you realized the the vision of software-defined networking that we've been talking about in those textbooks, gosh, for a couple of decades, SDN, you know, we've been reading about this architecture. Um, so you've done it, you pulled it off. Now what? Is what's next? Is it AI-defined, intent-based networking, autonomous networking? What's the next big thing that you're uh thinking about?

SPEAKER_01:

Yeah, it's funny because Megaport is if you look at Megaport, we're about two and a half thousand network devices in 26 countries, in over a thousand data centers. And connecting those is over 6,000 pieces of fiber that run a sub-sea, round the world, terrestrial, you name it, all diverse. And then we have you know three, you know, 2,900 customers on top of that with about 30,000 plus connections. And every single connection is changing every second, and that's what's so astounding. Now, behind that, we don't have a single human that touches the network. So it's funny when we hire network engineers, they don't know what to do when they come here because they look to log on to a device. You actually can't log on to a megaport device anywhere. The way you make a change is by actually coding software to upload, update the code that then runs the entire platform. So we built that sort of 13 years ago, and that would be the largest autonomous network there is, there'd be nothing larger than that, totally automated. So we kind of, I don't know, we've been living in the automated network uh for a long, long period of time. It's it's kind of the only reason the company can even exist and be successful because if you want to deliver something in seconds, you have to automate it. Uh, even if you try to compete with Megaport and you had infinite money and you could hire thousands of people to have them waiting at every data center to code every single thing, we'd still beat you on time because of the automation. So for us, it's getting more locations and getting much more speed. So if you look at we 10x the speed of the network in the last year or so, we took it from 10 gig to 100 gig. To do that, we had to build 400 gig backbones globally, which is a pretty, pretty um difficult thing to do when you talk about the scale of the network. And then the next piece was add as many products that our customers need to help solve problems, whatever problems they have, that we can sort of fall into the same thesis. Needs to be automated, needs to be disrupted, needs to solve a problem, needs to be simple and easy to use. Uh, and then if you can do that, you can do it at mass scale. So every product that we've added continues to just fall into that category. And so we recently we added a whole heap of security features. You can terminate IPsec tunnels anywhere. We launched an internet product, which is the first automated internet product on the planet that you can deliver at 100 gig from data centers.

SPEAKER_02:

Wow.

SPEAKER_01:

We didn't know how that would go, but that thing exploded. So it's just sort of like if you keep listening to your customers, they'll keep telling you the problems they've got, and then you can solve them with software engineering and then deploy it across the backbone. So, yeah, there's a huge list, and I think Alexis brings most of these things now. She comes back, hey, all these customers want this. You're like, all right, well, let's go and build something cool. That's amazing.

SPEAKER_02:

And Alexis, I mean, you you probably hear all sorts of myths about the cloud, about networking that you would wish would just vanish already. Uh, what's one that you, you know, one myth that you think um is misunderstood and you know, with megaports should really disappear about networking?

SPEAKER_00:

Man, you put me on the spot on that one. Let me think. Um, I think a big myth for most people is that connecting to the cloud is easy, right? Most leaders, when they tell their team to go into the cloud, they don't expect networking to be a hurdle.

SPEAKER_03:

Great point.

SPEAKER_00:

Because to a lot of, I mean, I'm talking, I'm not talking about network engineers. I'm talking about like leadership higher up. We're talking the people who define the strategy and digital transformation, all of that. They don't, they expect shifting into the cloud to be the hard part. They don't expect connecting to it to be the hard part. And a big problem that Megaport solves is that we're flexible, right? So what we've seen with a lot of our customers is that they define these cloud strategies and then things change. And if their networking isn't supportive of the changes, maybe they're locked into longer contracts with their telcos, or their telcos can't get the links up fast enough for them to make the changes that they need, they're kind of stuck, right? And networking becomes the long pole in the tent where they can't make these changes. So that's really where we come in and we're able to be that flexible backbone they can use to continue to shift as their cloud strategy changes, when new workloads come out, when new AI features come out, when there's new things they'd like to adopt. Um, and we can really be there to support them.

SPEAKER_02:

Fantastic. Yeah, that's a great such a great point. And on a leadership level, Michael, I mean, you make this all look so easy and seamless, but uh behind the scenes, you know, it is is there a decision or decisions that looked risky at the time, but now have paid off big time?

SPEAKER_01:

Sorry, it broke up. Say the last bit it was um behind the scenes something.

SPEAKER_02:

Yeah, where were there decisions that had to be made that looked risky at the time, whether it was one, five, ten years ago, but really are paying off today, big time?

SPEAKER_01:

I think, yeah, there's so many decisions that I think at the time probably look risky. The goal is to try and make big bets constantly and lots and lots of bets. And if you if you sort of have a culture where you're innovating and trying to place consistent bets that aren't betting the company, clearly, but at least constantly having bets in certain areas, you'll follow the ones that are successful. And I think I said that like a great example was that internet product. We didn't know whether that was gonna be something that would, I mean, internet, everyone needs internet, but you know, you think, oh, well, everyone's got internet and everyone's in that space. And are we gonna solve a problem that's that important? And when we launched it, it just it exploded. And actually, it's funny, the team, the technical team internally didn't think that it was gonna be something that was that was going to be as as fast as adopted. Then there's a lot of decisions where you go, well, um, it's pretty obvious. Like we built an AI exchange, and that was listening to our customers who had all these different neo-clouds, so these new cloud providers appearing out of nowhere, with the same issues that cloud had trying to get access to their customers. And so we we built that AI exchange really quickly, and that got awesome adoption as well. So some of the some of the decisions are easy. But one of my favorite sayings is it's Bezos is sort of saying it's the one-way and two-way doors. And so you we often think about a decision inside Megaport is is this a two-way door or is it a one-way door? Uh 99% of the time, it's really a two-way door. And then there are some one-way doors. Like a good example of a one-way door might be a big acquisition that you wanted to go down. Um, a small acquisition is less of a problem because it's not gonna cause too much pain. But a big one, if it's not successful, it's not something you can undo or walk backwards. So with the one-ways, just take the time, really think about it, de-risk as much as you can, and then at some point you're gonna have to swallow hard and move forward. Well said.

SPEAKER_02:

And Alexis, you know, as a as a network engineer, um, you know, IP networks often are kind of stitched together with bubblegum and duct tape, kind of. And, you know, you got so many clouds, regions, partners. What's the trick behind making it all seem and feel seamless instead of, you know, try to stick stitched together?

SPEAKER_00:

I mean, the short answer to that question is having a partner like Megaport where you can use them globally instead of working with regional telcos and also have everything in one platform. Um, the longer form answer is that we're hosting an event next Thursday on October 23rd. And I didn't come here intending to pitch this, but you teed me up personally. Um it's actually the first cloud networking summit that Megaport's hosting, and we're bringing together 32 different speakers from all of the different CSPs, from our partner network, from our technical community, from our customers. And it's meant to be a no pitch, there's no product pitches, right? It's all engineers talking to engineers about cloud networking. Um, there's three different levels. We've got the fundamentals track for people who are brand new, the advanced track for engineers who already have cloud connections stood up, but maybe they need some fine-tuning, and then an executive track where we brought in a lot of leadership to talk about where we think things are moving uh in the future, especially connecting to AI and all of that. Um, so as far as how to simplify it, I think everyone should go. It's next Thursday, October 23rd at 9.15.

SPEAKER_02:

And is this online, virtual, in-person both?

SPEAKER_00:

Or can we online and everything will be recorded? So if you're in a not fun time zone uh that's not Eastern, you can always catch the recording.

SPEAKER_02:

That's phenomenal. Um, I I I'm gonna sign up and I'll I'll provide the link um in the post here. Um and speaking of opportunities, Alexa, I mean, everyone's talking about AI and and edge workloads now. The edge is where all the excitement is at. What are customers expecting from edge networking these days? And what are your capabilities at the edge as it extends further and further to you know small customers?

SPEAKER_00:

Evan, I'm gonna be honest, I don't know how to answer that one.

SPEAKER_02:

I'm thinking, you know, I'm thinking, Michael, I mean, the edge is gonna be everywhere, right? I mean, it's gonna be in every CVS and every retail and every hospital. Um, how deep into the customers do you see yourself going?

SPEAKER_01:

Well, yeah, good question. So if you look at SASE or SD WAN, you know, SD WAN's been around for a bit. Megaport with we are, if you think about us, we're a massive private backbone with thousands of pops. And so what if you if you think about it, when when you're at the edge and you're connecting back to something, uh the problem with the edge is it's typically across the internet all the way back to wherever you're trying to connect to. So something that we built out, which is an actual product that we have, we have a thousand, sorry, we have thousands of um compute, but we have a hundred physical um uh compute stacks distributed all around the world, very close to clouds, and in in in each city around the globe, all the cities that I think you know make sense. And inside that stack, you can spin up edge devices. So you can spin up SD WAN platforms, you can spin up um uh all of the you know, firewall platforms, Cisco Palo Aldo 40s, um, you know, versus, you know, there's a huge list. Meraki's just launched. So you can spin up all of these edge networking devices or firewall devices, and then connect using an IPsec tunnel from or the tunnels that they have through that space from your local, uh, let's say you're in you know LA, you would connect from your site in LA and directly land in the LA pop. And then when you're in the LA pop, you're on the global backbone. And you can do that from every location around the world. So that's that's sort of the edge from a network. What you're gonna see, what we are seeing now, which is really fascinating, I guess, from the edge in AI, and that is the inference rollout. And so if you look at AI being two things, one is like training the models, giant data centers with millions of uh GPUs, and then they sort of billions of dollars and then they spit out a new model. When you access the model and type something into Chat GPT or whatever it is, that's your inference. It's actually where AI is being monetized and the speed at which that needs to be responding, particularly if you think about AI voice. So if you think of it, it needs to be less than 150 millisecond delay, otherwise the human ear just starts to pick up that you're a robot. You need to deliver that at the edge, and then you need to have something far quicker than a traditional GPU to just process it. And so companies like Groc, GROQ, are a really interesting company that have just scaled so fast and building basically GPU-like services, but they call them LPUs, language uh language processing unit, not a graphical processing unit, the new thing, and that sits at the edge and and processes extremely fast, like something like I don't know, five times faster than what the traditional GPU will do, and it does it locally. So that that's a really interesting trend that we're seeing that we support. We actually connect a lot of the backbones to deliver that to as many locations as possible. That's the AI exchange that we're talking about. So they're the two kind of edges that when you when you that I would say are sort of interesting to us.

SPEAKER_02:

Wow. That's amazing. Alexis, what what's it like to have a CEO like Michael with such deep technical chops? Uh it's almost like you've I've never seen anything like it.

SPEAKER_00:

It's fantastic, right? He grew up as an engineer.

SPEAKER_02:

Right, right, right, obviously. And so it's it it must be really uh uh fun to collaborate.

SPEAKER_00:

And also it puts a lot of trust in the the leadership and the direction of the company to have a CEO that actually understands what's going on, right? You're not just leading the company from a business perspective and looking at the dollars and cents, but you are you do have a basic understanding of the technology, why customers are using it, how to explain it. You're able to have an in-depth conversation. And especially as an employee, like, I mean, it makes me trust you more. Frankly.

SPEAKER_02:

Speaking of trust, I mean, you built amazing trust through your content and through your presence on social media for those aspiring content creators, influencers, um, Gen Zers uh like yourself. Um, any tips, tricks, best practices? What's really worked for you in building across all these platforms, gosh, 90,000 LinkedIn followers, which is uh unheard of. What's been your secret uh to success?

SPEAKER_00:

I think, and this sounds very generic, and I know everyone's gonna roll their eyes, but not being afraid to be yourself on the internet. And when you're looking, when you're trying to do that, it's really easy for people to get wrapped around the axle and feel like content needs to be this extra thing. It's this extra video. I need to come up with extra ideas. But really, the secret is finding things that you're already doing in your day-to-day that you can repurpose as content online because then it's also authentic of things you're actually working on. So, for example, how I started was customers were asking me questions. I was a Cisco SE making technical presentations or writing emails to answer their questions. And so just repurposing that content into videos so it's useful for other people as well. Um, now my role is a little bit different. I get help from the marketing and product team on what to make videos on, and I do some more research um on Reddit and you know, see what people are saying. I read my comment section to figure out what people are interested in. Um, it's taken on a whole new life, but I do miss getting asked those questions from my dedicated customers because you're never going to beat that principle if one customer has a question, more due to.

SPEAKER_02:

So wonderful. Just documenting everything that's going on around you in a fun and educational way. Well, thank you so much, Alexis. Thanks, Michael, for the chat and the update. And as always, it's just thrilling to watch your progress onwards and upwards.

SPEAKER_01:

Awesome. Appreciate you having us.

SPEAKER_02:

Thank you. And thanks everyone for listening, watching, sharing this episode, and asking Alexis some very deep technical questions uh stu to try to stump her. So we'll we'll see you online. Take care. Bye bye. Ciao.