DePIN Week 😎

You asked for DePIN, we delivered DePIN!

We have three X Spaces this week speaking to the gigabrains in the DePIN space. The first with Messari has been and gone, so if you missed it, make sure you don’t miss the next two:

Today @ 5:00 PM EST: DePIN Derby with Irys, 375ai, DAWN and Heurist.

Tomorrow @ 5:00 PM EST: All Things Akash with Founder Greg Osuri and Messari Analyst Micah Casella 

SET YOUR REMINDERS!!!

Market Metrics

Total Crypto Market Cap: down 4.7% to $2.208T
Total AI Sector Market Cap: down 3.8% to $24.4B

Top Movers (24 hours):

📈 Dojo Protocol (DOAI): up 12.2% to $1.65
📈 Arbius (AIUS): up 6.5% to $36.69
📈 DeepBrain Chain (DBC): up 5.9% to $0.001975

Akash Network: Who’s Using It?

It’s DePIN week at Onchain Outpost, and since we’ve already given you a primer on DePIN (which you can find here), we thought it was time to address a question that many people have:

Well, it turns out there’s actually quite a few, with more joining the ranks every day.

Given there are seemingly more DePIN projects than grains of sand (we would know), we’ve decided to focus in on one you’ve likely already heard of: Akash Network.

For those who don’t know, Akash is a decentralized marketplace for computing resources. On one side, people with idle computing hardware provide their resources to the marketplace to put it to work to earn, and on the other side, users can access affordable computing power for tasks like AI model inference, among a bunch of other cloud computing tasks similar to the ones offered by AWS & GCP.

For many decentralized marketplaces, token incentives have driven a lot of resources on the supply side but the demand side has been lacking. However, it appears the tides may be turning for Akash in particular.

A recent Messari report by Micah Casella has revealed that Akash Network is seeing significant growth in demand for its resources across a wide range of projects and use cases. The report shows QoQ growth has been very strong, with new leases up 44%, active leases up 11% and revenue for providers at a new record of $176,000 for the quarter. This piqued our interest, so we decided to take a closer look at Akash and the projects driving the surge in demand 👇

Brev - Flexible ML Development

Brev dev is an ML development platform that lets developers build, train, and deploy machine learning models in the cloud. Brev differentiates itself by being flexible, allowing developers to start small before scaling up to larger GPU clusters as the demand grows, which is a common challenge in ML project scaling.

In April this year, Brev integrated support for Akash GPUs, allowing Brev users to tap into Akash's decentralized GPU power for their AI projects, providing permissionless access to cost-effective compute. In mid-July, Brev was acquiring by NVIDIA, signaling the giant’s interest in improving access to AI development tools. Interestingly, this acquisition makes NVIDIA an indirect partner of Akash.

Passage - 3D Spaces Made Easy

Passage lets anyone create 3D digital experiences. Think of it as a drag-and-drop editor for building virtual spaces – no coding required. Whether you're hosting an event, setting up a virtual office, or creating a members-only area, Passage lets you do it in 3D. It's like upgrading from static web pages to interactive, immersive “game-like” environments.

Passage has integrated with Akash to enable creators to launch their virtual worlds directly on Akash GPUs. This integration provides creators with access to decentralized, cost-effective computing resources, drastically lowering the barriers to entry for developing virtual environments.

Side note - getting Metaverse vibes here

Venice.ai - Privacy-First Chat Bot

Venice.ai, created by Erik Voorhees, is an privacy-first AI chat platform that enables uncensored content creation. It stores chat history locally in your browser and encrypts prompts before sending them out. This means Venice only knows your IP and email, while compute providers only see conversation history - neither gets the full picture. Venice also offers uncensored AI interactions, setting it apart from mainstream AI tools which have built-in bias and certain topics that cannot be answered.

Venice.ai uses Akash Network's decentralized cloud to process its AI inference requests. Encrypted user prompts go through a proxy server to Akash-powered resources, separating the user from the compute provider. You can check out Venice for free here.

NeuralAI - 3D Asset Generation on Bittensor

NeuralAI claims to be revolutionizing 3D asset creation with AI. Their platform transforms text and images into high-quality 3D models, making the process faster and more accessible. NeuralAI's mission is to level the playing field between indie game developers and AAA studios by providing free, efficient tools for 3D asset generation.

Operating as a Bittensor subnet, NeuralAI runs its TAO miners and inference workloads on Akash, utilizing the network's decentralized GPU power to train and operate its 3D-focused AI models.

Other Notable Mentions

Compute Aggregators

• FluxEdge: Integrating Akash resources into their marketplace, expanding decentralized compute options.

• Prime Intellect: Tapping into Akash's GPU resources to broaden their global compute offerings.

• Spheron Network: Aggregates Akash resources with its suite of dev tools, enabling anyone to deploy a web app that runs onchain.

AI Model Training

• Nous Research: Leveraged Akash-sponsored compute to train their Hermes 2 Θ model.

• ThumperAI: Pioneered the first successful foundation model training on Akash with their AT-1 model.

• Cicero: Utilized Akash GPUs to pre-train the first Latin-focused LLM.

Miscellaneous

• Waku: Encouraging node operators to leverage Akash for their privacy-focused communication protocol.

• FLock: Promoting Akash for deploying validator nodes in their decentralized data-privacy infrastructure.

• tao.bot: Powering Bittensor subnet access via Telegram using Akash Chat API.

• Fractal.fun: Uses Akash IPFS gateways for fetching NFT collection data.

X Post of the Day

👀 We tend to agree