Armada: The Hyperscaler for the Edge
Our investment in Armada
Today we’re thrilled to announce our partnership with Dan Wright, Jon Runyan, and the entire Armada team as they build the infrastructure stack for AI in the physical, distributed world.
Armada Co-Founders Dan Wright and Jon Runyan.
Image credit: Armada
Inference Moves to the Edge
The first wave of the AI buildout has been about training frontier models — gigawatt-scale data center campuses, frontier labs, and centralized clusters in a handful of locations chosen for cheap power and permissive permitting. The next wave is an inference story, and inference behaves differently. It is latency-sensitive, it is mission-critical, and increasingly it is sovereign: workloads that cannot afford to round-trip to a distant hyperscaler region, and data that cannot legally or safely leave the site, the company, or the country where it was generated.
The numbers behind that shift are striking. Roughly three-quarters of the world’s data is now generated at the edge, yet almost none of it is processed there. Deloitte predicts that inference-specific workloads will account for roughly two-thirds of all compute (up from a third in 2023 and half in 2025). Twenty years ago, cloud computing created the hyperscalers. In the last five years, the neoclouds emerged to serve AI training. The edge is the next platform shift, and Armada is the answer.
The Only Full-Stack Edge Platform
Most companies attack a slice of this problem — connectivity, or hardware, or compute, or software. Armada built all of it as one product. The Armada Edge Platform brings together Atlas, an operating system for edge assets; Galleon, a family of ruggedized, mobile, hyperconverged data centers, to the megawatt-scale Leviathan. The result is a turnkey path from raw land and power to operational AI compute in weeks rather than years — deployable on an oil rig, at a mine, aboard a Navy vessel, or across the remote stretches of Alaska.
Leviathan: Armada’s megawatt-scale modular data centers built for high-density AI training and inference workloads, sovereign neo-cloud, and multi-tenant compute environments. Image credit: Armada
Today’s announcement raises the ceiling again. Alongside its Series B, Armada today unveiled a global framework agreement with Johnson Controls for Galleon Forge One — a dedicated modular data center factory in Arizona, spanning up to 400,000 square feet and expected to create more than 500 jobs, with continuous production of Leviathan beginning this summer. As Dan framed it, the AI race will be won not by one-off projects but by those who can manufacture and continuously deploy AI infrastructure “with speed, scale and sovereignty.” Armada isn’t running pilots. It is building the industrial base.
The Okta Mafia, Again
At Garuda, we start with founders. Armada is led by a genuinely seasoned bench — operators who have built and scaled at Okta, VMware, Amazon, Microsoft, and AuditBoard — and that operational maturity shows up in how quickly the company has gone from stealth to scale.
Arpan also had the privilege of working alongside Jon Runyan, Armada’s co-founder and COO, during his years at Okta. Jon is exactly the kind of operator you want building hard, physical, high-stakes infrastructure: rigorous, calm under pressure, and relentless about doing things the right way. He’s an operator’s operator.
It also makes Armada our latest investment in what we’ve come to think of as the “Okta Mafia” — the remarkable group of operators who came through Okta and are now building category-defining companies of their own. Jon at Armada, Alex Bovee at C1, Alex Salazar at Arcade.dev, and Jay Srinivasan at Stitchflow.
Let’s Go!
We’re proud to back Armada alongside an incredible group of investors including Overmatch, BlackRock, 8090 Industries, Founders Fund, Lux Capital, Felicis, Shield Capital, Marlinspike, Silent Ventures, and new strategic investors Johnson Controls, Mitsui, and Singtel Innov8.
The world is waking up to the fact that latency-sensitive, mission-critical AI compute is moving to the edge. Armada is building this future for the United States and for our allies. We can’t wait to see what Dan, Jon, and the team build next. Let’s go!
For more on Armada, visit armada.ai and read the full announcement of the Series B and Galleon Forge One.
Garuda Ventures is a first check fund that partners with relentless, ambitious founders. Subscribe below to get the latest from us, and visit garuda.vc to learn more.