How Targon Supply Portal Could Reshape How Compute Gets Monetized

Inside almost every datacenter on Earth sits a quiet inefficiency: GPU and CPU capacity that earns nothing between workloads. Manifold Labs’ new Targon Supply Portal (TSP) converts that dormant capacity into revenue by routing it into Bittensor’s Subnet 4.

The portal is the unified onboarding surface for compute suppliers plugging into Targon’s confidential compute marketplace. It lowers the technical barrier for crypto-native operators and removes the crypto barrier entirely for traditional datacenters. 

Either way, the hardware ends up serving Targon’s enterprise workloads.

How to Use the Targon Supply Portal

The flow is intentionally short; most of the friction is front-loaded and picking the right path and confirming hardware fit after which the portal carries the rest of the onboarding:

1. Visit Targon Supply Portal, and click on ‘Become a Supplier’ to view onboarding path,

2. Pick your onboarding path from the options available. These are:

a. Permissionless: Connect directly to the Bittensor blockchain, register your hotkey, and start earning $SN4 rewards. No middleman, no intake call.

b. Targon Managed: Manifold handles every blockchain-side operation and pays you out weekly in fiat or a currency of your choosing.

Built for datacenters that want revenue without touching a wallet.

Both routes feed the same marketplace. Just through different operational mode.

2. Confirm your hardware and ensure you’re connecting to the portal through the supported confidential computing setups:

a. AMD EPYC 9xx4 series (Genoa/Bergamo) with SEV-SNP, for CPU workloads.

b. Intel TDX with NVIDIA Hopper GPUs (H100, H200).

c. Intel TDX with NVIDIA Blackwell GPUs (B200), in multi-GPU configurations.

3. Know what runs on the machine

Each configuration runs the Targon Virtual Machine (TVM), which deploys confidential VMs onto bare-metal hardware and provides end-to-end attestation across the host and its GPUs. Workloads stay encrypted and verifiable even on infrastructure the customer does not own. This is the layer that makes enterprise demand possible.

4. Clear the prerequisites and ensure:

a. BIOS access on the host (This is non-negotiable as operators without it should expect setup to stall).

b. A supported hardware configuration.

c. Bittensor wallet familiarity (permissionless path) or a completed hardware profile (Managed path).

5. Onboard to unlock node monitoring, performance tracking, and earnings visibility all live inside the same interface.

Conclusion

Decentralized compute has had a persistent on-ramp problem: the operators with the most idle hardware are often the least equipped to navigate token mechanics, and the ones fluent in token mechanics rarely run datacenters at scale. The Supply Portal collapses that gap into a single interface.

The deeper bet underneath it is that confidential, end-to-end attested compute is the format enterprises will eventually demand for any AI workload touching sensitive data. If that’s right, the suppliers walking through this portal today are not just earning yield on cold hardware. They are positioning on the supply side of a market that has not finished forming.