Aethir (ATH) News

How do GPUs actually get paid for? Mostly by the customer who signs. On recent IREN 3-year contracts, prepayments covered 45-55% of GPU capex. Investment-grade debt at ~6% took the Microsoft deal to 96% funded, with a ~2-year payback. The biggest lenders agree. SpaceX just https://t.co/wcqMtCLJsZ

The five largest US cloud and AI infrastructure providers have committed $660 to $690 billion to CapEx in 2026, nearly double 2025 levels. That spend gets recovered somewhere, and the obvious place is the price of a GPU hour. Land, grid upgrades, and custom silicon all end up in https://t.co/GjYtBvUz7K

The @MiamiHerald is covering @AxeCompute's new Miami headquarters. The move to Miami follows Axe's major compute deal announcements in recent months.

Buying GPUs is the easy part. Getting a cluster live on the date a customer was promised is the real job. That's what Aethir is built around: compute that actually shows up, ready for training and inference.

The neocloud boom is the most exciting thing happening in AI infrastructure right now, and it's also where the gap between talking and building is widest Every week I see another team announce they're "becoming a cloud". Some raised for a different idea and found that GPU

Gartner sized the AI-optimized IaaS market at $42.2 billion for 2026, up 96.4% from $21.5 billion in 2025. Inside that number, inference accounts for $23 billion against $19 billion for training, and inference share climbs to 59% in 2027. The industry spent three years building https://t.co/RmteZzltos

One update on the SpaceX GPU lease I walked through last week. Fitch looked at the $40B commitment and left SpaceX at BBB+, because that compute was already in its forecast. It's treating the deal like any other equipment lease. The numbers underneath are the part that got me.

Jensen Huang said, "Agents think, reason, iterate, use tools, and as a result consume an enormous amount of tokens." He's describing the exact problem Aethir addresses by combining our decentralized GPU cloud spread across 200+ locations and our ACCELERATE neocloud program, https://t.co/Rz8ucDcHyP

A measurement study published in March put a number on something practitioners have felt for a while. Chain-of-thought prompting lifted accuracy sharply on hard tasks, and raised energy per answer by 120 to 150 times. Majority voting added 72% to 177% for much smaller gains. https://t.co/KaxfCvr0ZF

Contracted power isn't capacity. Rentable GPU hours are. That's why Aethir ACCELERATE builds against demand that already exists: 10 mid-sized sites, up to 20 MW across the US and Europe, built for B300 and GB300 clusters and live in months, not years.

Had a genuine holy shit moment with @bot this month I'd been helping my mum and dad sort visas for Egypt and then Jordan (they're adventurous folk). The visa site was pretty terrible so I had Grok Bot fill it all in, travel dates included Couple of weeks later it pings me

Every neocloud quotes capacity as one number. It’s actually three, and only the last one is rentable: 1. Contracted power 2. Deployed megawatts 3. Available GPU hours Here’s how to tell them apart before you sign🧵 https://t.co/TsxnacTjxH

Why did nobody tell me Claude scores every account My conspiracy theory of the week is they use that score to guess how smart you are, then turn your reasoning up or down depending on whether they think you'll notice Reasoning is the model thinking before it answers. It works

Industry-standard GPU lead time in 2026: 36–52 weeks. CoWoS packaging and HBM supply are fully allocated through 2027. Aethir enterprise provisioning: 48 hours for selected GPU types, from distributed capacity that already exists across 94 countries. The shortage is real. The https://t.co/4yKE4Pbmg1

No exposure to the Binance incident, but we're not taking chances. Our security team is auditing every system now. We appreciate how fast Binance is responding and its commitment to cover all user funds. Protecting the network and our community comes first.

Aethir's co-founder, @MRRydon, just published a quick, step-by-step walkthrough of how GPU finance leasing actually works. A great read for anyone who wants to understand the complexities of GPU compute financing today.

Axe Compute ($AGPU) now trades on @binance as a 20x perpetual and as AGPUB, a 1:1 backed tokenized stock. Two routes, 24/7, in USDT to tokenized GPU compute. AGPU has also been listed across multiple exchanges in various forms, including @Gate, @Aster_DEX , @HTX_Global, https://t.co/66v0lm7aKa

Great coverage of @AxeCompute's journey, strategy, and GPU compute contract portfolio by @Investingcom

Training a frontier model needs thousands of GPUs wired together with almost no tolerance for latency between them. Inference doesn’t. As IEEE Spectrum puts it, inference workloads don't require tightly interconnected GPUs, which is precisely why they suit distributed placement. https://t.co/WE7rDUB6kL