@patrickbetdavid Plus, Noone is discussing the stockpiles of 5-20% HEU that #iran still has!! this level of enrichment represents roughly 90% of the effort required for weapons-grade material. With advanced IR-6 centrifuges, a 184.1 kg stock of 20% uranium can be rapidly converted. So sad
My thesis in founding Akash was that every tech company will eventually offer cloud hosting services, and they need a distribution point to aggregate demand.
That is the vision of Akash and thats what we were building towards.
$AKT is currency for compute.
Gemma 4 31B IT is now LIVE on Akash ML.
As a dense 31B instruction-tuned multimodal model built by @Google, Gemma 4 31B delivers state-of-the-art performance across text and image understanding, complex reasoning, and native function calling.
This open model supports a 256K context window, meaning it can process the equivalent of two full novels, an entire codebase, or hours of transcripts in a single prompt without losing context.
→ Deploy your first multimodal agent with limited trial access today: https://t.co/GcXkcxC8xG
@claudeai is so much dumber than it used to be, he cant even fix my @openclaw basic provider config without making 5 mistakes that I need to explicitly point him to! I know @AnthropicAI are all about optimization but dont overdo it 😵
@cosmos@0xMagmar@cosmoslabs_io@DefiantNews@grok what does that mean if I’m holding a crypto portfolio right now? Should I sell or buy to avoid market shifts due to the stated claim in the post ?
As we move from the structured frameworks of Web2 to the dynamic, decentralised landscape of Web3 AI, we summarise the key players from an AI tech stack perspective and share FLock’s role in the ecosystem. #machinelearning#DecentralisedAI#Web3AIStack
1/7
these types of partnerships signal hope and a brighter future for the cloud computing industry, as we move forward, we will see more and more advanced DePIN platforms to provide proper alternatives to the monopolized cloud market!
By recycling $400B in closed financial loops, the compute oligarchy creates an artificial AI bubble that traps builders in expensive, high-wastage infrastructure.
Akash and @RunOnFlux provide the decentralized solution: a permissionless cloud that eliminates vendor lock-in and democratizes high-performance compute.
Read more → https://t.co/q5oLHd6UBW
Akash saw its most significant technical expansion in Q4 '25, following the deployment of Mainnet 14 and the launch of AkashML.
Utilization currently sits at a healthy ~30%, which is a vital indicator of the network’s ability to scale and absorb massive high-end capacity even during industry-wide uncertainty.
Read the full Q4 report ↓
https://t.co/CFKuEBdPr2
The world’s most advanced data centers are sitting idle, not for lack of GPUs, but for lack of power.
The servers are ready, The racks are built….but the juice isn’t there.
And that’s exactly why decentralization wins. We need to scale where power exists, not just where corporations build.
Read more 👇
https://t.co/wH83KFc3XZ
OpenAI’s new $38B cloud deal with Amazon underscores the risk of AI centralization.
“If intelligence infrastructure mirrors today’s cloud oligopoly, societies risk ceding both data and decision-making to a few corporate platforms. Decentralized AI argues for the opposite: a world where models live across devices, governed by open protocols rather than proprietary APIs.”
Read more 👇
https://t.co/XM7zyR9QOt
Starlink at current speed is ideal for training, with 1Gbps it'll be able to serve as edge nodes.
That's the future we're headed with Homenode on Akash, which is perfect for Starlink.
https://t.co/CaoKelp80X