AI compute youcan verify.

Open AI inference where useful computation secures the chain and proves the model behind every answer.

Block heightBlocks added to the chain since genesis.31,304

Network computeInference proofs per second across the network, averaged over the last 50 blocks.Implied tokens per second: proofs per second × 256, the tokens each proof covers. Implied by difficulty and proof rate, not measured throughput.245.89K proofs/s62.95M tok/s

Effective difficultyHow hard it is to win a block. Rises as more compute joins the network.162.22M

HoldersAddresses currently holding $TSC.20,093

Live from scan.tensorcash.org

01

Question

User request

What pet should I get?

TensorCash network

A user or application sends a question to the network.

01 / The problem

Same label. Different model.

How do you know what actually ran?

Honest miner

Model

Qwen3-8B

Verified

Question

What pet should I get?

Scores as ranges

Number from block hash0.55

cat40%

dog35%

monkey25%

Output

The answer is dog.

Verifier sweep

Theanswerisdog.

Block accepted. Reward paid.

Cheating miner

Model

Qwen3-8B

actually: a cheaper model

Question

What pet should I get?

Scores as ranges

Number from block hash0.55

cat60%

dog25%

monkey15%

Output

The answer is cat.

Verifier sweep

Theansweriscat.Qwen3-8B rerun: 0.55 → dog

Fake model detected. Block rejected.

Same number. Different word.

One word can match by chance. The scores behind it never do.

02 / How it works

One computation. Two layers.

The same work, read once as an answer and once as proof.

Layer 01: InferenceUser facing
PromptUser input data
GPUInference execution
AnswerGenerated output

Same computation

Layer 02: Verification + settlementProtocol level
ProofCryptographic validity
TSC blockConsensus validation
RewardProtocol settlement

03 / One network

TensorCash

Need compute

For AI builders, agents and enterprises.

  • Get the model you paid for, and check that it actually ran.
  • Every answer can carry a proof, usable as an audit trail for decisions made with AI.
Use inference

Have compute

For GPU owners, data centres and mining pools.

  • Get paid for answers, and compete for block rewards with the same work.
  • Put idle hardware to work, with no separate mining rig needed.
Start mining

06 / Scaling infrastructure

Compute is the beginning.

AI inference

Verification

TensorCash protocol

Built into the protocol: a fixed set of financial functions, with no general smart-contract VM, the attack surface behind most DeFi hacks.

Stablecoins, shares, bonds and funds can be issued directly on the chain, with legal documents, KYC and holder voting built in. The issuer posts a $TSC deposit, which is returned once the asset has paid enough fees to prove it is genuine.

07 / Ecosystem

Ecosystem

Marketplace

Ask Tilly

AI chat and agent platform with a built-in wallet, and an API for developers

Community

Bob Labs

Community mining pool and pay-per-token AI chat, with agents and an API, paid in $TSC

Built on TensorCash

Explore the ecosystem

Useful computebecomesVerifiable value.