Cloud Decentralization

“We believe compute should be a public utility. Compute should be as free as the air we breathe and water we drink, and that’s the future we’re all building towards. What that means is restoration [of] sovereignty: we need to have control over our data, we need to have control of our compute, and we need to empower communities rather than corporations to drive this technological progress.” — 00:08:48

“In a case like Akash, we have been live for about 4 years and there hasn’t been a single time the network went down or slowed down. It’s always available, it’s always live, and that’s possible because it’s open source — people are looking at the source code all the time, people are pushing updates all the time — and it’s decentralized: there’s not a single system that controls the entire network.” — 00:10:14

“[Personal computing] created the internet on the foundation of decentralization, connecting these individual devices in a hive network, sharing intelligence and removing the need for a centralized coordinated system to communicate… historically, decentralization is what really spun the innovation and the accessibility that we have today. But eventually somehow the centralization came, with large hyperscalers… close to 70% of all the IT spend is spent by four companies.” — 00:02:15

“It’s really about removing the risk of a system failure — removing systemic failure risk. Today you have centralized systems like Amazon and Google of the world: if the company collapses, the product exits, no matter how important it is to you or to the society. If the company dies, the product dies.” — 00:12:28

“Internet was created in the ’60s to combat the centralization of data centers… because we had large mainframes and people were like, this is not good, because one mainframe gets attacked, we’re done. The same thing is happening — we have 20 hyperscalers. It’s not good, because they get attacked, you’re done. So the best thing you can do is, number one, distribute AI as much as possible. AI should be with everyone, not with anyone.” — 00:40:36

“Instead of having 20 data centers, how about 10,000 data centers, where it’s impossible for adversary to compromise our AI infrastructure in the future… one data center goes down, who cares, we have 10,000 of them, even more.” — 00:43:26

“A lot of places don’t have energy now… right now it’s about 10 years you’ve got to wait before you can get an interconnection request approved… enterprises are in a very big trickle where you can’t really build out massive data centers… and you have to eventually go distributed.” — 04:49:31

“To build one gigawatt AI cloud, we need around 8,300 of these [GB200 NVL72] racks. But here’s the key: they don’t have to live in one facility. They can live everywhere… This is hyperscale rebuilt from the ground up to be permissionless, flexible, and resilient. The cloud of the future isn’t a place. It’s a planet-wide mesh of compute, and Star Cluster is how we build it.” — 04:34:16

“Today, Akash runs everything from websites to AI inference, but tomorrow it’ll help train and host the models that define the new era. Because the cloud of the future shouldn’t be owned by one company — it should be owned by all of us.” — 00:02:16

“Infrastructure is destiny, but only if it’s in the right hands — of everyone. Let’s build a cloud no one can shut down. Let’s build the intelligence layer of a free civilization.” — 00:20:25

“We call it Akash Starcluster: a product roadmap that brings hyperscale performance to a globally distributed network. Phase one launches in 2025 — data center grade star nodes. These are high-density racks running in existing colo spaces, powered by renewable energy, operated by independent node keepers. They form the foundation, the raw power layer.” — 00:00:01

“They don’t have to live in one facility. They can live everywhere — distributed across regions, climates, and use cases. This is hyperscale rebuilt from the ground up to be permissionless, flexible, and resilient. The cloud of the future isn’t a place. It’s a planet-wide mesh of compute, and Star Cluster is how we build it.” — 00:10:11

“This is a hyperscaler rebuilt from the ground up to be permissionless, flexible, and resilient. The cloud of the future isn’t a place — it’s a planetwide mesh of compute, and Starcluster is how we build it.” — 00:02:09

“This is the architecture of the future: a planetary AI mesh, alive with local energy and local compute. Redundant by design because there’s no single points of failure. Open to anyone, anywhere… The future isn’t hyperscale, it’s hyperdistributed.” — 00:08:12

“In this context, massive centralized data centers are obsolete. They’re rigid, expensive, and confined to one geographic location. Even worse, they create single points of failure. If one power grid goes down or is overheated, an entire segment of the country is plunged into a technological dark age.” — 00:02:11

“In case of a national conflict tomorrow, international conflict, what do you think is going to get targeted? These little sitting ducks, large targets… These are massive data centers. We all know where they are. There are about 500 of them in America… When someone takes out your infrastructure, they’re going to target this infrastructure.” — 00:38:23

“AWS went down four times, took almost all the internet with it. It causes such a disruption every time a data center goes down that it’s almost like decentralization sells for itself.” — 00:05:50