The world's premier decentralized compute marketplace • $AKT Deploy on Akash Console today → akt.fyi/4pDDwVa
🌎
- Tweets 7,389
- Following 291
- Followers 128,470
- Likes 17,719
Pinned Tweet
Homenode is a ROI accelerator for local AI workstations.
It gives you the distribution channel to put idle compute to work.
Open beta is now public, link below.
10
24
165
20,564
Hackathons are some of the best feedback loops in the ecosystem. Novel use cases get built on Akash, and new users hit the marketplace for the first time.
Watching someone use it cold is the most direct UI/UX feedback there is, and it informs how the end product gets refined.
SF Tech Week was amazing!
we brought together some incredible builders for a day of hacking on cyber defense, one of the most critical challenges as AI becomes more capable and autonomous
and the projects? some seriously cool stuff:
🛡️ an agent that catches breaking API changes and auto-patches affected code.
🔎 one that queries 19 million auth logs in plain English to hunt threats.
👁️ AI that detects and redacts leaked secrets in screen recordings.
none of this would've been complete without our incredible mentors and judges @gregosuri @benjaminshafii @henryhengzj @pedroslopez @DaghanAltas @BryceNeil__ and more.
10
22
142
6,490
Congrats to the winners of the Akash Track during the Cyber Defense Hackathon hosted last week.
Well attended and many got to learn about Akash Network for the first time.
An agent that buys its own compute needs a price it can reason about, and an order book gives it one.
@akashnet Greg Osuri @gregosuri @tokensandai
"And providers, in real time, bid for your workload and they compete. It's a reverse auction, so you end up getting the best price."
Cyberdefense Hackathon at the AWS Builder Loft (@AWSstartups).
nitter.siddz.com/AgentGlassInc/status/2…
1
3
12
1,116
AkashML now averages 26M requests a week, double the 13M it averaged from mid-July to early August.
260M requests in three months, served on Akash GPUs.
As open models close in on the frontier and agents take on more work, demand keeps climbing.
11
50
246
13,559
1.46 trillion tokens served on OpenRouter in under 90 days.
Open models running on decentralized GPUs from the Akash Network is proving to be very competitive.
openrouter.ai/provider/akash…
1
4
24
1,509
Deploying on Akash should feel simple. Console now looks the part with an incredible UI update.
Redesigned pages, clearer layout, easier to manage your deployments from start to finish.
Here’s a quick tour.
12
41
199
12,590
But you have to drive it yourself to feel the experience.
console.akash.network
First time top-ups get a 10% bonus credit match, up to $1,000.
4
20
1,658
AI is making attackers faster. It can make defenders faster too.
Developers, security engineers, and agent builders welcomed to our hackathon to build the next generation of AI security on Akash's decentralized GPU compute.
Pull up to the Cyberdefense Hackathon in SF today and build alongside the @akashnet team. Let’s hack. 🧑💻
Register- luma.com/cyberhack?tk=stddCp
7
36
176
6,674
Akash Network retweeted
AkashML is the launch partner for @Architect_Fi’s Liquid Inference.
Every prompt is auctioned, and providers compete on price. The request routes to the lowest offer that meets the buyer’s rules.
OpenAI and Anthropic SDKs are compatible with one API key, where the price is locked before the first token.
Decentralized GPUs are now in a live inference market.
Introducing Liquid Inference by Architect, a new LLM model router where inference providers compete dynamically to offer lowest-cost tokens. Providers onboard in minutes, not weeks. Buyers have full price visibility and receive 20% of referred fees in free inference. More info:
Liquid Inference can be used in place of any OpenAI or Anthropic API-compatible backend. Behind it, our team has leveraged our experience trading and building financial exchanges to create real-time two-sided price discovery for inference. Benefits for buyers and providers:
INFERENCE BUYERS
• Fully compatible with agentic coding tools: Claude Code, Codex, OpenCode, Cursor, Pi, Cline, and many more. Full multi-modal support.
• Hundreds of open- and closed-weight models. Create routing rule presets or use our Auto routing algorithms.
• Providers compete dynamically on price for every prompt, offering the lowest marginal cost.
• Sign up for free with email, first 500 users get $20 of free inference.
• Receive 20% of referred fees as free inference, 10% for second-level referrals.
INFERENCE PROVIDERS
• Onboard through the Liquid Inference app. We verify new providers in minutes, not weeks.
• All prompts are OpenAI API standard. Simple REST/Websocket API to register models and quotes.
• Update quotes dynamically based on your own costs. Monetize GPU capacity only when you want.
• Standardize payouts via Stripe, full itemized records of all jobs performed.
4
26
92
6,583
Subsidized APIs hide the cost of intelligence. That bill is coming due.
Chips are scarce. Energy is scarce. Subsidies will shrink. The market reshuffles around whatever delivers intelligence most efficiently.
Akash is where idle GPUs surface. Owners earn. Builders pay less.
This is partly due to OpenAI halving the API-equivalent value of their Pro 200 plan last week. But even before that, they were still less generous than Anthropic. (4/6)
7
38
159
6,730
Power. Permits. Politics. New datacenters are getting harder to build. Meanwhile always-on agents and bots are about to multiply token demand.
The compute has to come from somewhere. It may as well come from your house, and pay you for it.
Akash Homenode in a nutshell. 🔗👇
Coming up on the next DeAI News Journal, @gregosuri from @akashnet wants you to imagine running a token factory at home... just like growing your own food. He says it’s cleaner, cheaper, and you can even sell your excess. Could local AI offer privacy and savings?
12
37
185
12,440
Now looking for 4090s / 5090s / Pro 6000s to join our fleet.
Help bring the cost of inference down for others while bringing in passive earnings to offset your own hardware.
homenode.akash.network
3
15
69
4,534
A chatbot costs tokens once per interaction.
An agent costs tokens until the work is done.
Most people haven't made that switch yet. As they do, demand for inference will grow exponentially.
AkashML is an agent-ready endpoint with the supply to meet it.
Artificial Intelligence (AI) agents use 5–50x more compute tokens than a typical chatbot, and demand keeps climbing.
Compute infrastructure (electricity, data centers, chips, memory) can take years to permit and build, creating bottlenecks. Providers with existing, compute infrastructure stand to benefit.
Read more on The Stack: grayscale.com/the-stack/why-…
10
43
213
17,579
New load balancer went live for AkashML. It keeps capacity saturated across all available GPUs hosting the model.
Yesterday: AkashML served 25.6B tokens as a dominant provider for gpt-oss-120b.
Traffic on OpenRouter is earned. The more performant endpoint serves the tokens.
13
48
248
12,750
Anyone can run the same GPUs. Fewer can make them produce like this.
The latest upgrade is @akashnetAI's load balancer. Custom built, routing every request on live metrics from the inference engines. Concurrency doubled in some cases. Errors down 90%.
akashml.com
10
37
2,728
GPUs are not the final bottleneck for AI.
Watch Greg Osuri's full Fintech TV talk on where the real constraint is, and how Akash is positioning for an energy-scarce environment.
terminal.fintech.tv/fintecht…
6
41
203
18,968
The ability to run open models locally have given enterprises and individuals a reason to own GPUs capable of near frontier intelligence.
.@gregosuri is on @FINTECHTVglobal to speak on Akash unlocking both commercial and residential compute to meet the rising AI demand.
The chip shortage is structural, not a production issue, and tons of GPUs are sitting idle.
@gregosuri, Founder of @akashnet, tells @RemyBlaireNews the network connects that idle compute to a grid, turning it into a liquid, tradable market.
8
44
235
14,390
Routing every request to a frontier model is a margin killer. The fix is lower cost inference.
Host your own models on rented GPUs at a predictable rate, or use an endpoint like AkashML at better unit economics to preserve your profits.
Akash does both.
TRACKED CHANGES: Harvey’s gross margin fell from about 50% to -50% by June as agent token use spiked twentyfold on rented OpenAI and Anthropic models, Bloomberg reports.
8
27
169
13,483
