{"id":2711,"date":"2026-03-11T16:13:43","date_gmt":"2026-03-11T17:13:43","guid":{"rendered":"http:\/\/gogetmuscle.com\/?p=2711"},"modified":"2026-03-11T17:49:12","modified_gmt":"2026-03-11T17:49:12","slug":"ai-at-the-edge-how-armada-is-taking-compute-everywhere-the-cloud-cant-go-dan-wright-ceo-of-armada","status":"publish","type":"post","link":"http:\/\/gogetmuscle.com\/index.php\/2026\/03\/11\/ai-at-the-edge-how-armada-is-taking-compute-everywhere-the-cloud-cant-go-dan-wright-ceo-of-armada\/","title":{"rendered":"AI at the Edge: How Armada is Taking Compute Everywhere the Cloud Can\u2019t Go | Dan Wright (CEO of Armada)"},"content":{"rendered":"
The GTM Podcast is available on any major directory, including:<\/p>\n
Dan Wright<\/a>\u00a0(Co-founder & CEO of Armada) joins GTMnow to unpack what it actually takes to bring AI infrastructure to the places the cloud was never built to reach.<\/p>\n The cloud covers about 30% of the world. The other 70% (think: oil rigs, the Arctic tundra, military ships, remote mines) is where some of the most critical decisions happen, making latency a life-or-death and billions of dollars difference. Armada is building the infrastructure for that part of the world: modular, ruggedized AI data centers that go to the data, instead of the other way around.<\/p>\n From the first offshore edge computing deployment with the US Navy, to cutting avalanche response times in Alaska from over a day to real time, to sovereign AI installations in Saudi Arabia with Aramco and Microsoft, Armada is redefining what operating at the edge even means.<\/p>\n 00:00<\/a>\u00a0\u2013 Why the cloud only reaches 30% of the world and what Armada actually builds<\/p>\n 03:11<\/a>\u00a0\u2013 From Saudi Arabia to the Arctic: recent deployments that redefine the edge<\/p>\n 04:12<\/a>\u00a0\u2013 AI\u2019s physics problem: why distance from data breaks everything<\/p>\n 05:31<\/a>\u00a0\u2013 How Starlink turned every remote location into a potential AI cluster<\/p>\n 07:14<\/a>\u00a0\u2013 What distributed intelligence means and why it\u2019s Armada\u2019s founding principle<\/p>\n 08:14<\/a>\u00a0\u2013 The global AI race: sovereign compute as a national security strategy<\/p>\n 09:00<\/a>\u00a0\u2013 The Genesis Mission, the White House, and Davos: why sovereign AI is mainstream<\/p>\n 12:49<\/a>\u00a0\u2013 GTM for a hardware company when the demo involves shipping a 40-foot container<\/p>\n 15:18<\/a>\u00a0\u2013 Why Armada\u2019s customers do the selling for them (and do a better job than some reps)<\/p>\n 17:35<\/a>\u00a0\u2013 The Microsoft partnership and extending Azure to the places it can\u2019t reach on its own<\/p>\n 18:51<\/a>\u00a0\u2013 Category creation: the lesson Dan learned about specificity and business value<\/p>\n 21:02<\/a>\u00a0\u2013 What\u2019s next for Armada, SpaceX, and why Dan thinks we\u2019re two years from the moon<\/p>\n 25:34<\/a>\u00a0\u2013 The founder advice Dan wishes more people took seriously<\/p>\n 1. The cloud was built for a world before AI. 2. Connectivity unlocked the edge, but without infrastructure it\u2019s useless. 3. Sovereign AI is here and the US and China are in a global AI race. 4. Category creation is about problems, not technology. 5. Write your manifesto before you do anything else. If you run go-to-market, you already know the problem: your data lives everywhere. Spreadsheets, CRMs, sales calls, ad platforms\u2026 yet you\u2019re still guessing what to do next.<\/p>\n HockeyStack is the AI platform for modern GTM teams. It unifies all your sales and marketing data into a single system of action. Built-in AI agents help teams prospect the right accounts, improve conversions, close and expand deals, and scale what works. That\u2019s why teams like RingCentral, Outreach, ActiveCampaign, and Fortune 100 companies rely on HockeyStack to eliminate wasted spend, take better decisions, and make space to think.<\/p>\n Learn more at\u00a0hockeystack.com<\/a><\/p>\n Outbound sales today is still built around manual work: writing emails, building lists, and managing sequences across disconnected tools.<\/p>\n Nooks is breaking the mold. With their new AI Sequencing product, Nooks is an agent workspace where AI works alongside reps, helping them understand accounts, prioritize the right prospects, and generate context-rich outreach based on actual first-party interactions.<\/p>\n No more wasting time in multiple tools \u2013 it\u2019s one unified workspace for all your outbound channels. Arm your sales team with intelligent outbound. Learn more at nooks.ai<\/a>.<\/p>\n 00:00<\/p>\n Dan Wright<\/strong>: We call ourselves the hyperscaler for the edge. Rather than having the data sale to some very far away data center, we can actually bring these data centers to the data. Saudi Arabia. The state of Alaska in the Arctic, US Navy everywhere beyond the cloud providers networks where there\u2019s data.<\/p>\n 00:16<\/p>\n Sophie Buonassisi<\/strong>: Dan Wright is the CEO of Armada deploying AI data centers to the most extreme places on Earth. When an AI model is too far from a data source, what starts to actually fracture?<\/p>\n 00:27<\/p>\n Dan Wright<\/strong>: It just doesn\u2019t work. Oil rigs, mines, ships. The way infrastructure has been is that it was built for a world before.<\/p>\n 00:35<\/p>\n Sophie Buonassisi<\/strong>: AI, and you\u2019re operating in extreme remote environments. And previously these were data desert. How has satellite connectivity like Starlink helped enable this edge first approach?<\/p>\n 00:45<\/p>\n Dan Wright<\/strong>: Starlink totally changed the game and now everywhere is a potential AI cluster, a potential AI hub.<\/p>\n 00:51<\/p>\n Sophie Buonassisi<\/strong>: We\u2019re now in the age of AI sovereignty. It\u2019s no longer just about who has the best model. It\u2019s more about who owns the infrastructure stock.<\/p>\n 01:00<\/p>\n Dan Wright<\/strong>: It is a global AI race and the countries that move the fastest. Those are going to be the big winners. There\u2019s going to be data centers in space. We are going to go to the moon. We are going to go to Mars.<\/p>\n 01:10<\/p>\n Sophie Buonassisi<\/strong>: How many years away do you think we are from seeing that on the moon?<\/p>\n 01:13<\/p>\n Dan Wright<\/strong>: I would say.<\/p>\n 01:15<\/p>\n Unknown<\/strong>:<\/p>\n 01:26<\/p>\n Sophie Buonassisi<\/strong>: Dan welcome to GTM now.<\/p>\n 01:27<\/p>\n Dan Wright<\/strong>: Thanks for having me on.<\/p>\n 01:28<\/p>\n Sophie Buonassisi<\/strong>: Absolutely. It\u2019s a pleasure to have you here.<\/p>\n 01:30<\/p>\n Dan Wright<\/strong>: Great to be here.<\/p>\n 01:31<\/p>\n Sophie Buonassisi<\/strong>: And I want to start off with what you\u2019re building<\/p>\n 01:33<\/p>\n Sophie Buonassisi<\/strong>: because the cloud is king. Narrative has hit a bit of a wall now in extreme environments, even a few milliseconds of latency, the delay it takes to send data to<\/p>\n 01:44<\/p>\n Sophie Buonassisi<\/strong>: the data center is<\/p>\n 01:45<\/p>\n Sophie Buonassisi<\/strong>: can really be a deal breaker. And Armada seems to be that solution to the problem.<\/p>\n 01:51<\/p>\n Sophie Buonassisi<\/strong>: Yeah. So for anyone unfamiliar with Armada, how do you explain what you\u2019re building?<\/p>\n 01:55<\/p>\n Dan Wright<\/strong>: Yeah, we call ourselves the Hyperscaler for the edge. And what I mean by that is we are building a new type of cloud company<\/p>\n 02:01<\/p>\n Dan Wright<\/strong>: the cloud, if you think about what it is, it\u2019s these massive hyperscale data centers,<\/p>\n 02:06<\/p>\n Dan Wright<\/strong>: they\u2019re great, but they only cover about 30% of the world. And so we\u2019re building the infrastructure for the rest of the world, which is really needed because a lot of the biggest problems happen there in that area of the world.<\/p>\n 02:20<\/p>\n Dan Wright<\/strong>: We call that area the remote edge. Everywhere beyond the cloud providers, networks where there\u2019s data.<\/p>\n 02:25<\/p>\n Sophie Buonassisi<\/strong>: I\u2019ve heard Armada mentioned in so many different contexts<\/p>\n 02:28<\/p>\n Sophie Buonassisi<\/strong>: from extreme environments, everything from Alaskan drone imagery through to the military. Is there a particular extreme environment that you feel like<\/p>\n 02:35<\/p>\n Sophie Buonassisi<\/strong>: showcases this edge first approach best?<\/p>\n 02:38<\/p>\n Dan Wright<\/strong>: What\u2019s exciting is that the edge is now constantly being redefined with Armada. When we started the company,<\/p>\n 02:45<\/p>\n Dan Wright<\/strong>: people thought of the edge as the edge of the cloud providers networks. When we say the edge, we mean the edge in terms of wherever the data is. And then when I say wherever, I mean,<\/p>\n 02:55<\/p>\n Dan Wright<\/strong>: literally anywhere. The reason that we named<\/p>\n 02:57<\/p>\n Dan Wright<\/strong>: company Armada is that rather than having the data at Sail to some very faraway data center, we can actually bring these cutting edge modular AI factories, AI data centers to the data, and just a few recent examples.<\/p>\n 03:11<\/p>\n Dan Wright<\/strong>: We were the first ones to do that in the middle of the summer in Saudi Arabia with Aramco in partnership with Microsoft.<\/p>\n 03:19<\/p>\n Dan Wright<\/strong>: Extending all the things that,<\/p>\n 03:21<\/p>\n Dan Wright<\/strong>: you know, Aramco likes about Azure to the edge, but then also bringing cutting edge models,<\/p>\n 03:25<\/p>\n Dan Wright<\/strong>: from, you know,<\/p>\n 03:26<\/p>\n Dan Wright<\/strong>: the leading new model providers, anything they want to this remote edge, doing the same thing with the state of Alaska in the Arctic, focused on emergency response, helping them cut latency to process data from,<\/p>\n 03:41<\/p>\n Dan Wright<\/strong>: drones for responding to avalanches and floods from over a day to real time to save lives.<\/p>\n 03:47<\/p>\n Dan Wright<\/strong>: And then we also did the first offshore edge computing ever with the US Navy.<\/p>\n 03:52<\/p>\n Dan Wright<\/strong>: taking all of the data from drones, taking these really powerful AI models, bringing them to the middle of the ocean, where you can go on a ship, you can put it in a weapons bay, you can port for, deploy it in a C-17 C-130.<\/p>\n 04:04<\/p>\n Dan Wright<\/strong>: those are a few recent examples, but it seems like every week,<\/p>\n 04:06<\/p>\n Dan Wright<\/strong>: you know, we\u2019re constantly redefining the edge.<\/p>\n 04:09<\/p>\n Sophie Buonassisi<\/strong>: Always pushing the boundaries. Yes. And speaking boundaries.<\/p>\n 04:12<\/p>\n Sophie Buonassisi<\/strong>: AIS often talked about is this a serial thing? But in reality it\u2019s governed by the laws of physics. And when an AI model is too far from a data source, what starts to actually fracture?<\/p>\n 04:23<\/p>\n Dan Wright<\/strong>: It just doesn\u2019t work.<\/p>\n 04:24<\/p>\n Dan Wright<\/strong>: the problem with the way infrastructure has been is that it was built for a world before I. Right? All of the cloud providers built for a world, you know, 25 years ago when we,<\/p>\n 04:36<\/p>\n Dan Wright<\/strong>: didn\u2019t have AI in any meaningful sense, we certainly didn\u2019t have these really large, powerful models that we have today.<\/p>\n 04:43<\/p>\n Dan Wright<\/strong>: And on top of that, the data was in more centralized, you know, urban locations. But that\u2019s not true anymore either, because now all the data is on oil rigs, mines, you know, ships, you name it, where they\u2019re sensors all over these things. And so the problem is, if you don\u2019t have something like Armada,<\/p>\n 05:03<\/p>\n Dan Wright<\/strong>: you can\u2019t actually use AI where you need it most, where the data is.<\/p>\n 05:07<\/p>\n Dan Wright<\/strong>: that\u2019s where we come in, is we can take<\/p>\n 05:09<\/p>\n Dan Wright<\/strong>: these really powerful models, as you said, you know, they\u2019re not magic. There\u2019s physics involved. So you actually need the infrastructure to run them. We take that full stack to the data and then we solve, you know, really important problems like saving lives or,<\/p>\n 05:22<\/p>\n Dan Wright<\/strong>: preventing,<\/p>\n 05:22<\/p>\n Dan Wright<\/strong>: catastrophic events at oil rigs, mines, you know, enabling split second decisions on battlefields in a way that can really,<\/p>\n 05:30<\/p>\n Dan Wright<\/strong>: solve important problems.<\/p>\n 05:31<\/p>\n Sophie Buonassisi<\/strong>: And you\u2019re operating in extreme remote environments. And previously, these were kind of data deserts. Yeah. How has<\/p>\n 05:38<\/p>\n Sophie Buonassisi<\/strong>: satellite connectivity like Starlink? Yeah. Helped enable this edge first approach.<\/p>\n 05:43<\/p>\n Dan Wright<\/strong>: first of all, I like that term data desert. I think I might steal that from you if that\u2019s okay.<\/p>\n 05:46<\/p>\n Sophie Buonassisi<\/strong>: I like my yes.<\/p>\n 05:48<\/p>\n Dan Wright<\/strong>: so we have seen that that is one of the big trends that is happening. That was always a blocker to AI at the edge. If you don\u2019t have the internet, then how are you going to get these really powerful AI models to the edge? Right. You have to have them at least to be able to push them to the edge.<\/p>\n 06:05<\/p>\n Dan Wright<\/strong>: Then you can run them what\u2019s called air gapped, you can run them disconnected from the internet, but you have to have the internet to push them there. And usually what people want to do is run them and then send the insights or the metadata<\/p>\n 06:16<\/p>\n Dan Wright<\/strong>: back to the cloud so they can benefit at all their other sites as well.<\/p>\n 06:19<\/p>\n Dan Wright<\/strong>: So Starlink totally changed the game, and most people don\u2019t know. They only launched in public beta in November 2020. So it\u2019s a relatively new product.<\/p>\n 06:28<\/p>\n Dan Wright<\/strong>: Now. They\u2019re in, I think, 155 countries right around there,<\/p>\n 06:31<\/p>\n Dan Wright<\/strong>: and they\u2019re rolling out in new countries every week,<\/p>\n 06:34<\/p>\n Dan Wright<\/strong>: and the service is getting better. When it first came out, it was primarily a consumer product.<\/p>\n 06:38<\/p>\n Dan Wright<\/strong>: Then it went into enterprise a couple of years later. And then there was Star Shield for the government. Initially it was being used as a backup, and now the service has gotten so good that it\u2019s being used as a primary source, increasingly<\/p>\n 06:49<\/p>\n Dan Wright<\/strong>: the internet.<\/p>\n 06:50<\/p>\n Dan Wright<\/strong>: And so what that means is if I get a call like I did the other day and somebody says, hey, can you deploy one of your galleons, which is what we call our modular factories in Antarctica?<\/p>\n 06:59<\/p>\n Dan Wright<\/strong>: I could say, yeah, we can do that because Starlink\u2019s in Antarctica, we don\u2019t even need fiber in the ground.<\/p>\n 07:05<\/p>\n Dan Wright<\/strong>: And that\u2019s important because<\/p>\n 07:06<\/p>\n Dan Wright<\/strong>: everywhere is a potential, AI cluster, a potential AI hub, because you have Starlink and you have reliable internet everywhere.<\/p>\n 07:14<\/p>\n Sophie Buonassisi<\/strong>: what is the term distributed intelligence need to you?<\/p>\n 07:17<\/p>\n Dan Wright<\/strong>: Yeah. I mean, distributed intelligence is the<\/p>\n 07:20<\/p>\n Dan Wright<\/strong>: founding principle of armada. And<\/p>\n 07:22<\/p>\n Dan Wright<\/strong>: what that is, is<\/p>\n 07:23<\/p>\n Dan Wright<\/strong>: the world<\/p>\n 07:24<\/p>\n Dan Wright<\/strong>: moving from a situation where you have these really large hyperscale data centers that just train these big, you know, AI models, right?<\/p>\n 07:33<\/p>\n Dan Wright<\/strong>: To a world of inference, a world where you can use those models anywhere and you can have it distributed the same way that,<\/p>\n 07:42<\/p>\n Dan Wright<\/strong>: energy is distributed.<\/p>\n 07:44<\/p>\n Dan Wright<\/strong>: Right? Because there\u2019s lots of stranded energy around the world that can be used to power these high data centers. That\u2019s also been a big block or too I,<\/p>\n 07:51<\/p>\n Dan Wright<\/strong>: I talked about the connectivity, but then energy as well.<\/p>\n 07:54<\/p>\n Dan Wright<\/strong>: But also data is distributed because of the rise of IoT devices and the fact that cameras and sensors and drones have gotten so much less expensive, about ten less expensive over the last five years, now, it only makes sense that you would have distributed intelligence, because, again, that\u2019s where the power lives, and that\u2019s where the data lives.<\/p>\n 08:12<\/p>\n
\nDiscussed in this episode<\/h2>\n
\n
\nEpisode highlights<\/h2>\n
\nKey takeaways<\/strong><\/h3>\n
<\/strong>Every major cloud provider built their infrastructure 25 years ago, before large models existed, before sensors were on everything, before data lived on oil rigs or in the Arctic. The world has changed but the infrastructure has not. That mismatch is a massive market opportunity.<\/p>\n
<\/strong>Starlink changed the game for connectivity. But without compute, all you have is faster access to a problem you still can\u2019t solve. Every time Starlink goes live in a new country, Armada is the first mover on infrastructure. Connectivity + compute = the edge finally works.<\/p>\n
<\/strong>We\u2019re in a global AI race. And infrastructure is the battleground. For companies like Aramco and the US Navy, data sovereignty means the data cannot leave the site \u2013 by law, by security protocol, by necessity. You can\u2019t send classified sensor data from an oil rig or a Navy ship to a data center thousands of miles away and wait for a response. You need compute where the data lives (that\u2019s what Armada builds).<\/p>\n
<\/strong>Nobody wakes up wanting to buy infrastructure. Dan\u2019s approach from day one was to get specific about the problems (avalanche response in Alaska, split-second decisions on a battlefield, catastrophic failure prevention on an oil rig) and let the technology follow.<\/p>\n
<\/strong>Before a single hire, before an office, before raising a dollar. Armada had a manifesto \u2013 this is why we\u2019re building, here\u2019s where the world is going, here\u2019s why it matters. Dan calls it an unfair advantage. When you can articulate exactly why your company needs to exist, recruiting gets easier, fundraising gets easier, partnerships get easier. You\u2019re not selling a product. You\u2019re building a movement. A 10-slide pitch deck might get you to Seed, but it won\u2019t build a generational company.<\/p>\n
\nThank you to our sponsor: <\/strong><\/h3>\n
HockeyStack<\/strong><\/h3>\n
Nooks<\/strong><\/h3>\n
\nFollow Dan Wright<\/strong><\/h3>\n
\n
\nFollow Sophie Buonassisi (Host)<\/strong><\/h3>\n
\n
\nWhere to Find GTMnow<\/strong><\/h3>\n
\n
<\/a><\/li>\n
<\/a><\/li>\n
\nGTM 182 Episode Transcript<\/h2>\n