The listed price is not the traded price.
Public marketplaces show H100 at $3.25/hr.
Real year-long contracts clear at $2.10–2.70.
Why the gap?
Bilateral negotiations are confidential.
Nobody publishes what actually traded.
Price discovery needs transparency.
🔗 Join the Waitlist: https://t.co/LuaHr8A0uT
Compute is being contracted like capacity, not bought like hardware.
Oracle is selling Tencent access to ~100,000 advanced chips over 5 years, about $7B, with ~30% paid upfront.
Broadcom is lining up as much as $42B of financing against Anthropic’s multi-year compute commitment.
Nebius is buying inference tech so GPUs spend less time idle.
That is the real market:
power, chips, utilization, and who is willing to prepay for the output.
The scarce layer is still compute.
Three things every commodity market needs:
1. Fungible units
Gold: 1 oz = 1 oz anywhere
Compute: every contract is unique ❌
2. Price transparency
Oil: WTI, Brent published daily
Compute: real deals are confidential ❌
3. Tradeable infrastructure
Electricity: day-ahead markets exist
Compute: can't resell a signed contract ❌
That's the gap we're closing.
More on commodity formation: https://t.co/atMVIN7M2v
To everyone who’s been waiting with us — we hear you. 🤍
We know many of you have been on the Waitlist for a while, and we truly appreciate your patience.
We’re now moving deeper into ComputeFi, with discussions already underway for our first batch of idle compute partnerships.
And yes — your community points still matter. Holding an NFT will not only come with its own benefits, but also boost your community points.
We’re getting closer, and there’s more to share soon!
Thank you for staying with us!
We’re building this with you!
Compute capacity is shifting east.
East Asia: Semiconductor and data centre expansion
Southeast Asia: Power + connectivity advantages
India: Scale + cost efficiency
MENA: Capital meets ambition
Infrastructure is growing fast.
But pricing transparency? Still catching up.
Building a more transparent and accessible compute market is exactly what @RaxFinance is working toward.
A permanent framework is moving closer.
When the rules get clearer, capital usually gets more comfortable showing up.
The RWA ecosystem stands to benefit most if that shift continues.
Via @a16zcrypto
The debate on AI compute has moved.
From "will demand exist?"
To "do the unit economics work?"
GPU pricing, financing costs, utilization.
Contracted power, density, offtake.
A 1GW factory only works if it can be rented, powered, and paid back.
Some names are being marked down on returns.
Others are being marked up for contracted power, density, and long-term offtake.
The market is underwriting compute like real assets now, not like a theme.
Via @wallstengine
BlackRock: AI compute could be tokenized.
Translation:
compute is about to get a market.
The world’s largest asset manager just named the future.
We’re building the unit it needs.
1 cT = 1 GPU-hour.
Fungible. Redeemable. Tradable.
That’s the layer.
🛠️ Builder’s Note #01
An index without a pool is a number nobody can transact on.
A GPU cloud without an index is capacity nobody can price.
@RaxFinance is building all three:
✓ The machines — custodied in partner data centres
✓ The unit — cT, backed 1:1 by those machines
✓ The venue — where cT can actually clear
One place where GPU compute is:
Minted → Priced → Traded → Redeemed
A benchmark, once referenced, becomes a market standard.
🔗 Join the Waitlist:
https://t.co/0sXOt71OLm
💬 Drop your address below.
🎁 Surprise Unboxing Opportunities await.
A founder's dilemma:
"Signed 12-month GPU contract.
Raised $2M seed.
Built for 3 months.
Funding talks fell through.
Now: 9 months left on contract.
$27K/month burning.
Can't resell. Can't transfer.
Just... stuck."
This isn't rare. This is Tuesday in AI.
AI compute is not priced by hardware alone.
The economics depend on a combination of:
• GPU supply
• Workload demand
• Utilization
• Energy costs
• Capital efficiency
As AI infrastructure scales, understanding these variables becomes increasingly important.
Because the next compute opportunity may not be about owning more GPUs.
It may be about understanding where demand is heading.
Compute is not a commodity yet.
Why? No standard unit.
Every GPU purchase is different:
• Region differs
• Provider differs
• Price differs — $1.99 to $5.99 for the same H100
• Contracts are locked and hard to resell
@RaxFinance is building a standard:
1 cT = 1 GPU-hour
Backed 1:1 by real machines.
Trade it or redeem it for compute.
Same unit, anywhere in the pool.
Making compute fungible and tradable.
🔗 Join the Waitlist:
https://t.co/0sXOt71OLm
💬 Drop your address below.
🎁 Surprise Unboxing Opportunities await.
🎁 The GPU Price Prediction Bonus Challenge is live!
Starting August 20, we’ll have 12 prediction rounds running through August 31.
Think you can beat the GPU market?
📌 Check the full details here:
https://t.co/bUhQ39wXyN
Keep predicting. Keep improving.
AI companies have a compute problem:
Sign a 12-month GPU contract.
Need it for only 3 months.
→ $30K/month still gets burned.
40% of your reserved capacity sits idle.
You’re locked into the contract.
→ Capital stays tied up.
@RaxFinance is building a marketplace for tradable compute capacity.
Have 9 months of H100 capacity left?
Monetize the unused portion.
Need compute for a shorter period?
Access available capacity without committing to a long-term contract.
Compute capacity shouldn’t have to sit idle.
Coming Soon 👉 https://t.co/LuaHr8A0uT
2026: AI compute is becoming the next oil.
Just like crude oil has:
• Spot markets
• Futures contracts
• Regional benchmarks
@RaxFinance is building the same for AI compute:
• Collateralized lending (Vault)
• Trading marketplace
• Regional price index
Tokyo vs Mumbai: 2x price spread on GPUs
A trillion-dollar market is being born 🌱
Join the waitlist → https://t.co/0sXOt71OLm
AI companies have a compute problem:
Sign a 12-month GPU contract.
Need it for only 3 months.
→ $30K/month still gets burned.
40% of your reserved capacity sits idle.
You’re locked into the contract.
→ Capital stays tied up.
@RaxFinance is building a marketplace for tradable compute capacity.
Have 9 months of H100 capacity left?
Monetize the unused portion.
Need compute for a shorter period?
Access available capacity without committing to a long-term contract.
Compute capacity shouldn’t have to sit idle.
A 1GW AI factory is being described as a $50B -a-year rental machine.
Jensen’s CNBC numbers make the unit economics clearer:
• ~$50–60B to build
• ~$50B in annual rental revenue
• capital that can pay itself back on a very short cycle if utilization holds
That is why this is no longer just a GPU cycle.
It is power + chips + cooling + financing, packaged as a factory that sells compute by the year.
When the world’s largest compute company starts quoting AI factories in gigawatts and rental revenue, the asset is the factory itself.
Compute is becoming the scarce, cash-flowing layer underneath the entire AI buildout.
https://t.co/FTU9qT9Axv