$TasQ is now live on Robinhood Chain via @flapdotsh.
100% fair launch. Every token started in the pool.
Official CA:
0xdc453b1405CfE25e2afBA8E7F4272b0a00437777
10% Team buy vested on Sablier:
https://t.co/4scLvChOnj
Tokenomics/Vesting schedule: https://t.co/19MWgnX0iZ
TasQ × @googlecloud ☁️
We've been accepted into the Google for Startups Cloud Program and received $2,000 grant in Google Cloud compute credits.
On top of that, we also received:
→ 12 months of Google Workspace Business Plus
→ $200 in Google Skills training credits
→ Access to Google AI Studio, partner perks and technical resources
The credits go straight into getting more GPUs up on our dApp, which means more capacity and more room for revenue.
Thank you @googlecloud 🙏
TasQ's first major workload is live on the network 🟣
A 7B-parameter model is being fine-tuned on TasQ right now, for a project our BD team brought in.
Details stay private while it runs. Once it's done, we'll publish a full workload report, shared with the client's permission.
First of many.
Another burn!
We just burned another 0.3% robinhood:0xdc453b1405cfe25e2afba8e7f4272b0a00437777 Supply funded from our project revenue.
Total burned - 0.9% (over $9,000)
Tx:
https://t.co/GRxIJX1aRW
Revenue has started flowing in on TasQ.
50%+ of it goes to buying back and burning robinhood:0xdc453b1405cfe25e2afba8e7f4272b0a00437777. More usage, more burns. That's the flywheel.
We're setting up a public @Dune dashboard so anyone can track revenue and burns live.
https://t.co/a3MS0wGDgT
Why we use @privy_io 🔑
To join TasQ you need to hold $TasQ. Privy makes proving that simple:
→ Sign in with your own wallet
→ No seed phrase to get started
→ Self-custodial wallets: your keys stay yours
→ One signature proves you hold $TasQ. No gas, no transaction.
→ Secure Transactions
Easy to join, without giving up control.
Test 1 of 4 is live 🧾
A real AI job on the TasQ network, with a signed receipt:
→ Qwen2.5-1.5B-Instruct on an NVIDIA L4
→ 126 generations · 15.1 min of real compute
→ Every input and output hashed, receipt signed by the operator
Download the outputs and verify it yourself, we will also be publishing a detailed Nvidia attestation on our website - in - EAT Format
This one ran in standard mode. Test 2 goes out later today, and tests 3 and 4 run tomorrow on hardware GPUs with our privacy system enabled.
https://t.co/INflfuf4yu
We just burned another 2 Million tokens worth over $2000 from our buybacks funded from revenue ($TasQ token tax + compute revenue)
Total Burned: 0.5%
Total Vested - (3months initially): 10%
Burn TX:
https://t.co/VPbeteKJPr
We're now onboarding projects onto TasQ private AI compute 🔌
Know a project that needs private, verifiable compute? Refer it.
→ Send them our way
→ They onboard
→ You earn (referral rewards coming soon)
Infrastructure update 🛠️
Patch notes:
→ Faster scheduling: smarter job to GPU matching → Attestation service: keys only go to proven enclaves → Holder verification: Privy check live in the app → Stability fixes: better node uptime and monitoring
And tonight, as requested by the community 👇
🔴 Live GPU attestation test A real GPU proves its enclave, end to end.
We'll share the results after.
https://t.co/a3MS0wGDgT
A quick word from Newman Wang, our lead tech advisor 🎙️
He's happy to see the adoption and growth since day one, and he's here to help us grow further with his connections and deep experience.
Thank you, Newman 💜
https://t.co/tBiAzgAVRW
A lot of you asked how TasQ actually keeps data private. Here's the simple version:
→ Locked before it leaves you Your prompts, model and data are encrypted on your side.
→ Opened only inside the chip The job runs in a secure enclave on the GPU. The key is only released after the chip proves it's running approved code. The machine owner sees noise.
→ Proof out, never data Results come back encrypted to you. The ledger keeps the proof, never your data.
Full details in the yellow paper: https://t.co/Gskw4Sh0lf
Thank you everyone for tuning in to the $TasQ launch! 💜
What's next (7 days):
→ Ordering 2 real GPUs → Expanding the compute suite → Holder access in the app → Ledger node onboarding
We're just getting started.
How does TasQ compare?
Most compute today asks you to trust whoever runs it. TasQ lets you check instead.
-> Your data stays sealed while it runs
-> Every job is verified
-> The ledger keeps the proof, never your data
-> Anyone can lend a GPU
Web2 clouds sell privacy as a paid add-on. Web3 markets opened up the supply. TasQ puts both together and adds proof.
https://t.co/a3MS0wGDgT
A lot of you have asked how privacy actually works in compute infrastructure. Here's how TasQ does it:
->Your job runs sealed inside secure hardware, so the machine only sees encrypted data.
->Several independent machines, or a zero-knowledge proof, confirm the result is right.
->The ledger records the proof, never your prompts, models or data
Private compute you can verify, not just trust. Watch how it works 👇
https://t.co/a3MS0wGDgT
tldr of the cinematic video:
Your AI shouldn't need trust.
TasQ is a network of shared compute where your prompts, models and data stay sealed, and every result is proven correct.
Lend. Compute. Verify.
Read the yellow paper: https://t.co/w34QmKpYdo
TasQ in one line:
run AI on hardware you don't own, without trusting whoever owns it.
Lend compute. Run privately. Verify every result.
Yellow paper: https://t.co/w34QmKpYdo
@Concept_felipe Yes, currently we have outsourced models (wrapped in our privacy layer) public to be rented. We will prepare a report and execution demo shortly and publish.
We are also preparing to build our own suite of hardware , and will have 2-6 in-house GPUs in the month of october.