AxonDAO’s B200 and RTX Pro 6000 GPUs are now operational and deployed, with public rentals on Vast.ai coming soon.


The bottleneck for the next wave of scientific progress isn’t ideas. It isn’t funding. It isn’t even talent.


It’s compute.


Researchers working with large-scale biomedical datasets, AI builders training models on sensitive health data, DeSci teams trying to move beyond whitepapers and into real discovery; they all hit the same wall. Access to high-performance GPUs is expensive, fragmented, and rarely designed with scientific workloads in mind.


AxonDAO was built to solve this. And starting today, the infrastructure is live.


The Hardware


AxonDAO's premium GPU fleet is now operational in the data center, with public rentals launching soon through Vast.ai, one of the largest GPU cloud marketplaces in the world.


The fleet includes:


  • NVIDIA B200 GPUs (8x) — Blackwell architecture, 192 GB HBM3e, 8.0 TB/s memory bandwidth. Datacenter-class machines built for large-scale AI training, fine-tuning, and high-throughput inference. Foundation models, protein folding, genomics — if the workload demands serious memory and compute, this is the GPU.
  • RTX Pro 6000 GPUs (8x)** — Blackwell-class workstation GPUs with 96 GB GDDR7 and 125 TFLOPS single-precision. Built for inference, rendering, mid-scale training, and 3D/creative workloads: power without the full datacenter footprint.


These machines will be available to anyone: AI builders, research teams, scientific organizations, creative professionals — on a pay-per-hour basis, with no contracts or waitlists.


Once rentals go live, every hour of utilization generates revenue that flows back to the DAO.


Why Vast.ai First?


The answer is simple.


Vast.ai is where compute buyers already are. It's the fastest path to real utilization, real pricing signals, and real revenue — and all three are needed before anything else can scale responsibly.


The intended progression is straightforward:


  • Phase A (Now):Marketplace distribution through Vast.ai — utilization, demand signals, revenue.
  • Phase B (Next): Axon Hub - AxonDAO's own compute platform where users purchase GPU time directly with AXGT, with discounts and priority access for holders and stakers.
  • Phase C (Scale): AxonOS integration, direct enterprise and research clients, and deeper token utility across the full stack.


We start where the buyers already are, prove demand fast, then scale the full system.


Revenue-Backed, Not Just Mission-Backed


For our community and AXGT holders, this launch represents a structural shift in what the token is connected to.


Until now, AXGT’s value has been tied to AxonDAO’s mission, its community, and its long-term vision. Those things are still at our core, but starting today, AXGT is also connected to revenue-producing physical infrastructure. Once rentals launch, GPU utilization generates real income from real compute buyers.


AxonDAO intends to allocate a portion of GPU profits toward AXGT buybacks, creating a direct link between real-world utilization and token value.


When buybacks are executed, they will be documented publicly with on-chain transaction links.


The goal is to make value accrual credible and verifiable, not something the community has to take on faith.


The Axon Hub: What Comes After Vast


The longer-term plan moves AXGT from a governance and community token into a payment and pricing primitive for GPU compute.


When the Axon Hub launches, users will be able to pay for compute directly with AXGT. Token holders and stakers will receive discounts and tiered pricing. During periods of high demand, AXGT holders may get priority access and reservation rights.


This creates structural demand for AXGT from compute buyers - people who hold the token not because they’re speculating, but because it saves them money every month. That’s a fundamentally different kind of utility, and it’s what the Axon Hub is being built to deliver.


AxonOS: Turning Compute Into a Product


Raw GPU access is a commodity. What turns it into a product is the software layer on top - the workflows, the tools, the experience that makes researchers come back.


That’s what AxonOS is being built to be: an AI-native workflow platform that makes AxonDAO’s compute easier, safer, and more useful for the DeSci audience.


AxonOS is currently in beta on AxonDAO’s existing cluster (B100/V100). Integration with the new B200/RTX Pro fleet comes later as the platform matures.


Over time, AxonOS becomes the primary interface through which researchers and builders interact with AxonDAO’s compute — increasing retention, economic activity, and the token utility that flows from both.


Building Toward Enterprise-Grade Compute


AxonDAO recently announced the selection of Oracle Cloud Infrastructure (OCI) as the foundation for its next-generation compute platform.


This signals where the project is heading: secure, sovereign, and compliant GPU infrastructure built for regulated AI, healthcare, and life sciences applications. The kind of workloads where data sovereignty, mandatory encryption, and network controls aren’t optional.


With OCI, AxonDAO can serve enterprise and institutional clients directly, particularly those working with sensitive, human-origin data. This is a meaningful step beyond marketplace distribution and toward infrastructure that operates alongside academic institutions, healthcare systems, and research organizations.


Vast and OCI aren't competing strategies. Vast validates demand and generates revenue. OCI provides the backbone for the direct, compliance-grade platform that follows.


Get Started


AxonDAO’s B200 and RTX Pro 6000 GPUs are deployed and will be rentable on Vast.ai soon. As soon as they go live, we will publish a detailed quick start guide.


Learn more: https://axondao.io