Powering the Future of AI. Building Decentralised Cloud Compute Infrastructure. https://xitter.cf/t.co/0q8024pYbg https://xitter.cf/t.co/9TF9m3zDSE
Global
Joined May 2022
- Tweets11.1K
- Following202
- Followers68K
- Likes17K
Pinned Tweet
Major ATH tokenomics update: we’re introducing daily ATH burns, funded by a share of network fees.
Every burn goes to a public burn address, where tokens are permanently removed from supply. The mechanism goes live in November.
We’re also replacing the fixed platform fee with a floating network fee, set by a published formula that reads market conditions. When the market weakens, the fee falls and the share that gets burned rises.
Cloud Hosts are protected by design: the fee can never exceed the value of Cloud Host rewards.
The underlying host emissions schedule remains unchanged. The update changes the fee mechanism applied around those rewards.
A dashboard tracking daily, monthly and cumulative ATH burned will be available in November.
Burns that follow real network activity, and you can check every one of them.
Read the full announcement below👇
Aethir retweeted
Amazon, Microsoft, and Google held 63% of a $143.4 billion quarterly cloud market in Q2 2026, per Synergy Research.
When three companies control most of the capacity, they also set most of the terms: Which GPUs you can get, where, on what contract length, and at what price when demand spikes. Buyers end up negotiating from the weak side of the table.
Aethir spreads supply across independent GPU operators in 200+ locations and 94 countries, so pricing reflects a pool of suppliers competing to serve demand.
AI compute works better as a market with more than three sellers.
With our upcoming Aethir ACCELERATE data centers, AI compute clients will have an additional, flexible source for B300 and GB300 workloads across the US and the EU.
How do datacenters get built and how GPU neoclouds actually get paid?
@MRRydon breaks down the economics behind AI compute 👇
A lot of people ask me how the GPU neocloud business works. So I thought I would provide a little bit of an example. This is how GPUs actually get paid for in this market using IREN's public numbers as a reference.
Step 1: the customer signs a multi-year take-or-pay contract and prepays. On IREN's recent 3-year contracts, prepayments covered 45% to 55% of the GPU cost.
Step 2: a lender funds most of the rest, secured by the contracted GPUs. For the Microsoft contract that was investment-grade financing at around 6%, which together with prepayments covers 96% of the GPU capex. For a customer without an investment-grade rating, it was 9% fixed.
Step 3: the data centre is a separate, long-life asset. IREN says it has funded construction with cash so far.
Recent 3-year contracts bring in more than $20M of revenue per MW and pay back the GPUs in around two years. At that payback, 9% instead of 6% on the loan barely changes the economics.
So the customer's prepayment and credit do most of the heavy lifting. Who the customer is matters far more than what the market thinks of AI on any given day, and I think that's the thing to follow.
Every efficiency gain in AI has been followed by more consumption, not less.
Cheaper tokens did not shrink inference budgets. They widened what teams were willing to attempt: longer contexts, more retries, agents that call a model thirty times where a person would have asked once.
Jevons described the pattern in 1865, watching more efficient steam engines burn through more coal rather than less.
Anyone planning capacity on the assumption that efficiency reduces demand has the direction of the effect backwards.
Aethir retweeted
Aethir is the first decentralized GPU cloud running NVIDIA B300 at production scale.
Blackwell Ultra performance, no CapEx, no multi-year lock-in.
Our upcoming Aethir ACCELERATOR data centers will also include B300 and GB300 compute for AI clients across the US and Europe.
Frontier hardware stopped being exclusive to hyperscalers.
Even the late states diverge.
Roughly 9,042 MW held approval to energize in ERCOT.
Observed non-simultaneous peak draw: about 3,883 MW.
A permission to draw power tells you nothing about the load actually on the system.
The test only works if you apply it to yourself.
ACCELERATE is site control plus design across data center build sites, up to 20 MW.
What is rentable today is the existing Aethir decentralized GPU cloud network.
Full breakdown: aethir.com/blog-posts/aethir…
Aethir retweeted
The GPU cloud market has a second rack-scale supplier in production.
AMD Helios is shipping, and the first neocloud order just landed.
What it changes🧵
Cloud Hosts are protected by design.
The fee can never exceed the value of Cloud Host rewards, and that ceiling is written into the formula.
The underlying host emissions schedule remains unchanged. The update changes the fee mechanism applied around those protected rewards.
All of it is checkable.
We’ll publish the formula and the public burn address, while the first burn date and a dashboard showing daily, monthly, and cumulative ATH burned and every transaction will be live in November.
ATH contract supply stays fixed at 42B, as it always has been.
Track every burn personally.
Full breakdown: aethir.com/blog-posts/aethir…
190 GW of AI compute capacity has been announced since 2024.
~16 GW scheduled for 2026.
~5 GW under construction.
Announced, financed, permitted, energized, commissioned and rentable are six different states, and the industry uses one word for all of them.
Explore our new blog to learn how to read an AI capacity announcement👇
Aethir retweeted
Gartner expects worldwide AI spending to rise 47% in 2026, to roughly $2.6 trillion, with the AI platforms and models market up 63% on its own.
Demand climbing at that slope outruns anything a construction schedule can deliver, which is why capacity keeps getting rationed by quota, region, and waitlist.
Aethir scales differently: 430,000 GPU containers across 200+ locations in 94 countries, added by connecting supply that already exists and already has power.
With Aethir ACCELERATE, we are also building our own mid-sized data centers to support compute-heavy AI workloads that need access to premium-quality, affordable compute, without the multi-year data center buildout process.
A curve that steep needs an elastic answer, not a bigger campus.
Congratulations on this amazing milestone to @AxeCompute, the team that's building premium-quality enterprise AI compute services.
Last week Axe Compute rang the @NasdaqExchange Opening Bell in Times Square.
A milestone for the team, and a marker on the way to something bigger. The building is just beginning, and we’re here to define the next generation of AI infrastructure.
We’re just getting started.
$AGPU #AIInfrastructure
Aethir retweeted
Aethir ACCELERATE Presents AI ACCESS Spotlight. Featuring @Sogni_Protocol
Creative AI has a bottleneck, and it’s the meter. Every new variation costs credits. Expensive hardware and heavy censorship add more friction.
So creators settle for work they’re not happy with, or never finish at all.
Here’s how Sogni is fixing that🧵