Smelt

Private beta opens soon. Apply for early access to the Colossal Capital ecosystem.

status: in development · ml infrastructure for the cc ecosystem

Raw data goes in. Trained models come out.

Smelt is the ML foundry for the Colossal Capital ecosystem — distributed training, model serving, data pipelines, and live 3D rendering for NFT collections, market models, and physics simulations. One foundry, every model the platform runs on.

How it works
# smelt foundry train → market models, portfolio agents, nft intelligence serve → live predictions across delt, zenith, twf render → on-demand 3d assets, reproducible from seed
What gets forged

Six model families in production rotation

Infrastructure

Four capabilities, one foundry

distributed compute
Training

Train models at production scale for the products that need them.

model serving
Serving

Live inference for ecosystem products — predictions where the work already happens.

data processing
Pipelines

Prepare training and live data without a separate foundry per product.

3d rendering
Rendering

On-demand 3D for NFT art and simulation visuals — reproducible from a seed.

Who it feeds

Every prediction in the ecosystem traces back here

Tendryl streams training data in — market feeds, chain events, financials. Delt takes predictions, signals, and portfolio recommendations into the trading app. TWF mints sea turtles live from this foundry. Aether serves rendered physics NFTs. Zenith displays live market predictions in the trading terminal, and ideight art-directs the visual pipeline the farm executes.
Smelt is internal infrastructure for the CC ecosystem — there's no standalone subscription. Platform access follows your Colossal Capital tier via the keys portal.
Developed in Partnership with Ackward Roots Inc

Product newsletter

Updates on Smelt

Product news, rollout notes, and when beta access opens. No waitlist queue — just the list.

Product updates only. Unsubscribe anytime.

Early access

Smelt is in active development. Leave an email and we'll write when cluster access opens.