Demand-side activity on Akash grew as new leases increased 44% QoQ and active leases increased 11% QoQ, hitting all-time highs.
Akash providers earned a record $176,000 (+8% QoQ) in revenue in Q2 as 27,000 new leases were opened throughout the quarter.
Akash’s initiatives to grow GPU supply paid off as it onboarded a daily average of 184 GPUs (+91% QoQ) to its capacity, with 33 more (+55% QoQ) being used in Q2.
Long-term AKT stakers benefited from supply and price increases, seeing rewards grow by 22x as they accumulated the emitted AKT throughout the year.
A variety of notable projects announced integrations with Akash, including Brev (acquired by NVIDIA), Venice, Prime Intellect, Passage, NeuralAI, tao.bot, FLock, Waku, and FluxEdge.
Primer
Akash (AKT) is a decentralized cloud computing marketplace that facilitates the buying and selling of compute resources. It is an open-source, permissionless protocol that provides an alternative to today’s centralized cloud services (i.e., AWS, Azure, and Google Cloud). Akash aims to leverage the global amount of underutilized server capacity, which can range from 5% to over 30%.
The Akash marketplace functions via a reverse auction, giving users the ability to name a price and describe the resources they want for deployments. Akash’s decentralized network of compute providers runs its open-source software and competes to provide resources, often at a fraction of the cost of big cloud providers. Specifically, Akash hosts containers where users can run any cloud-native application (e.g., gaming servers, blockchain nodes, and websites). Akash offers extensive cloud management services like Kubernetes, which can be used for hosting and managing containers.
Akash is a Tendermint-based blockchain built using the Cosmos SDK. Marketplace activity (requests, bids, lease details, etc.) is stored onchain and payments are settled with Akash’s native token (AKT). For a full primer on Akash, refer to our Initiation of Coverage report.
Akash’s marketplace uses a reverse auction, in which users propose a bid that describes the resources they’d like to use for a deployment. When accepted, a lease is opened onchain managing the activity of this relationship.
New leases on Akash show the number of new deals created between users and providers for renting resources on the network. Akash had a record quarter, with users opening over 27,000 leases on the network, an increase of 44% QoQ. Part of the growth in new leases may be attributed to the types of projects integrating with Akash throughout the quarter. Passage, Brev, FluxEdge, and Prime Intellect are all platforms that connect creators and developers to compute resources, using Akash as the backend.
Additionally, Waku and FLock have encouraged their respective network operators to deploy validator nodes to their networks utilizing computing resources on Akash. With the integrations of every new project and its user base, there is the potential to further increase lease creations and the number of active leases on the Akash Network.
Active leases on Akash are leases that currently utilize and manage resources on the network. Leases that close are no longer considered active. Active leases grew 11% QoQ, while revenue grew 8% QoQ. The number of leases does not necessarily need to correlate with revenue, given that revenue measures the USD value of what was spent on leases. For this reason, the chart above shows that revenue often does not move in tandem with leases, especially as leases peaked at the start of the quarter.
The short-lived spike in daily active leases in April 2024 was likely a result of Spheron Network using Akash as a deployment tool for small workloads that ultimately routed traffic back to its own machines. Because the workload size was so small per deployment, revenue did not see the same type of short-lived upward growth as leases. This dynamic demonstrates the importance of lease size and the impact on revenue, which is why Akash was (and still is) focused on onboarding high-performance GPUs to the network. Leasing deals with a focus on GPU compute tend to be more profitable for providers than leasing other network resources. Akash’s push to add support for more GPUs should continue being profitable for providers, especially as AI increases chip demand.
All Resource Compute
GPUs, CPUs, storage, and RAM are the principal resources that can be leased on Akash. In Q2, each resource saw capacity grow more than usage (with the exception of RAM, which shrank less). This led utilization in each category to also fall QoQ. A notable critique of decentralized resource networks is their lagging demand. While many have proven to scale the supply side successfully, many lack user-friendly tools and interfaces that simplify network usage. Akash is addressing user onboarding in various ways, most clearly seen in Overclock Labs’ focus on building out Akash Chat.
Akash Chat is a generative AI chat application like ChatGPT, except it is free to use. Within the user interface, it offers users the option to explore how to deploy their own AI models, directing them to the Akash Console. In Q2, Overclock Labs integrated Cloudmos Deploy into the Akash Console, making it a convenient platform for deploying one-click templates, creating custom deployments, viewing provider status, and more. Additionally, a number of teams (i.e., Passage, Brev, FluxEdge, and Prime Intellect) have built their own development platforms or applications that connect their users with Akash’s resources without interacting directly with the console or other more technical deployment options. With sub-25% compute utilization and sub-15% storage utilization, the Akash community is likely to continue to focus on building around the network’s usability.
GPU Compute
The Akash community prioritized onboarding GPU supply at the start of 2024 when AKT holders approved a proposal to allocate $5 million to GPU provider incentives. Year-to-date, Akash has grown GPU capacity by 147% and has seen growth in GPU usage. In Q2, Akash hit an all-time high average usage of 93 GPUs, up 55% QoQ. Akash’s various partnerships (as outlined earlier) should further contribute to the growth of GPU demand on Akash.
Active Providers
As a permissionless network, resource providers can join the Akash Network from anywhere in the world. This geographic diversity protects the network against regional disruptions, such as natural disasters, power outages, or political instability. It can also enable Akash to provide lower latency and better performance to users around the world, by processing tasks closer to the end users, reducing the time it takes for data to travel across the network. The number of average active providers on Akash grew 2% QoQ, reaching a record-high daily average of 74 active providers. Most Akash providers are concentrated in North America and Europe. In Q2, new resource providers joined from Australia and Latin America (Mexico and Brazil), expanding Akash’s global provider presence.
Token Analysis
Market Cap
AKT’s market cap peaked in the first half of the year, reaching $1.4 billion before ending Q1 at $1.1 billion. It ended Q2 at $867 million, falling 23% QoQ. Simultaneously, the protocol emitted roughly 11 million AKT, growing the supply by 5% QoQ. In addition to the general market price volatility that categorized crypto in Q2, sell pressure from various proposals, such as the Praetor acquisition ($550,000) and GPU Provider Incentives ($5 million), could have also contributed to the 26% fall in AKT price.
Staking
Staked AKT has stayed relatively steady throughout the year. However, rewards have jumped by almost 24x in USD terms YoY, vastly increasing the profitability of AKT staking. Although the end-of-quarter price of AKT fell in Q2, the average price of AKT throughout the quarter increased by 11%. Both the amount of AKT emitted and the price of AKT continued to grow QoQ throughout the year, driving the return on staking rewards. USD growth with a stagnant total network stake has led to more profitable staking opportunities for long-term AKT stakers.
Qualitative Analysis
Key Events
Brev.dev (acq. by NVIDIA)
Brev is a platform for AI/ML development that enables users to build, train, and deploy machine learning models on cloud infrastructure. It officially announced support for Akash GPUs in April 2024, as one of Akash’s closest partners. It has since been acquired by NVIDIA (mid-July), technically making NVIDIA one of Akash’s primary partners now. Because AKT is required to deploy workloads on Akash, Brev had to acquire and hold AKT on its balance sheet. With the acquisition, NVIDIA now holds AKT on its balance sheet.
Training Foundation Models
A foundation model is a large-scale machine learning (ML) model pre-trained on a vast amount of data. It functions as a starting point for developing other ML models (like LLMs) that power specific applications (like ChatGPT) at a more affordable price than if an application developer were to build an ML model from scratch. The first successful training of a foundational model using Akash’s infrastructure was completed at the start of Q2 2024. Thumper.ai provisioned a training cluster for Akash-Thumper-V1 (AT-1) using a series of resources leased from Akash’s decentralized cloud computing marketplace. The successful training of AT-1 on Akash showcases the potential for decentralized, permissionless computing resources to train large-scale AI models at efficient prices.
Integrations and Deployments
Passage (virtual world builder): Enables creators to launch their virtual worlds directly on Akash GPUs.
Brev (platform for AI/ML development ): Announced support for Akash GPUs, enabling users to build, train, and deploy machine learning models on Akash.
NeuralAI (3D-based AI asset generation Bittensor subnet): Runs TAO miners and inference workloads on Akash based on its 3D-focused AI models.
Waku (privacy-oriented communication protocols): Encourages network operators to run Waku nodes on Akash.
Venice (private, uncensored, generative AI application): Uses Akash GPUs to process inference requests.
tao.bot (Telegram bot for Bittensor): Integrated with the Akash Chat API to power its Interact platform, which enables users to access Bittensor subnets from Telegram.
Cicero (first LLM pre-trained on Latin): Used GPUs from the Akash network to pre-train its large-language model.
FluxEdge (resource marketplace): Adding Akash resources to its marketplace that will route traffic through the Akash Network.
Prime Intellect (decentralized AI development platform): Aggregates and manages global GPU resources from various sources, including Akash.
Unified Masked Diffusion (Academic Research): Used a cluster of NVIDIA A100 GPUs from the Akash Network to conduct the research.
Developments
Core development on Akash is approved through token-weighted governance and goes through a structured proposal process. The community can also simply create tools that anyone can use to access Akash. Below are a few notable examples of these tools.
Akash Chat: Overclock Labs shipped Akash Chat, an open-source AI chat application running on Akash. It hosts six open-source LLMs (that users can choose from), including Llama 3.1 405B, Llama 3, Mistral-7B, Nous Hermes 2 Mixtral, Mixtral, and Dolphin-Mixtral.
Akash Chat API: Overclock Labs delivered the Akash Chat API to expand the utility of the Akash Chat application.
Akash Console 2.0: In late May 2024, Overclock Labs integrated Cloudmos Deploy into the Akash Console. This integration makes the Akash Console a convenient platform for deploying one-click templates, creating custom deployments, viewing provider status, and more.
Praetor Acquisition (governance): AKT holders approved funding of $550,000 for Overclock Labs to acquire Praetor. Afterward, Akash integrated with Praetor to open source the Praetor Akash Provider platform, which accelerates the resource provider onboarding process.
Mainnet 11 Update (governance): AKT holders approved updating the implementation details around deployment deposits on Akash. Now, only one depositor will be allowed per deployment, and depositors will be able to reuse unspent funds.
Mainnet 12 Update (governance): AKT holders approved the resolution of an issue regarding fee allowances.
Deployment tool (governance): AKT holders approved $60,000 in funding to develop a deployment tool that will enable developers to make deployments from GitHub, GitLab, and Bitbucket to the Akash Network.
Closing Summary
Akash reached all-time highs for various metrics throughout the quarter. On the demand side, revenue grew to $176,000 (+8% QoQ), and users opened 27,000 new leases (+44% QoQ). Also, Akash’s initiatives to grow GPU supply paid off as it onboarded a daily average of 184 GPUs (+91% QoQ) in Q2. The average number of resource providers also grew to a high of 74, as Akash’s global provider footprint expanded to Brazil, Mexico, and Australia. Long-term AKT stakers benefited directly from supply and price increases, seeing rewards grow by 22x as they accumulated emitted AKT throughout the year.
Much of this growth was likely driven by many different projects adopting Akash. AI development platforms like Brev (acquired by NVIDIA) and Prime Intellect expand users' access to Akash resources. Additionally, the generative AI application Venice uses Akash GPUs to run its LLM. From a development side, Akash continues to update its protocol and improve the experience for both providers (Praetor acquisition) and users (Akash Console). With its AI-forward initiatives and community growth plans, Akash is primed to take advantage of the synergies created between crypto and AI.
This report was commissioned by Overclock Labs. All content was produced independently by the author(s) and does not necessarily reflect the opinions of Messari, Inc. or the organization that requested the report. The commissioning organization may have input on the content of the report, but Messari maintains editorial control over the final report to retain data accuracy and objectivity. Author(s) may hold cryptocurrencies named in this report. This report is meant for informational purposes only. It is not meant to serve as investment advice. You should conduct your own research and consult an independent financial, tax, or legal advisor before making any investment decisions. Past performance of any asset is not indicative of future results. Please see our Terms of Service for more information.
No part of this report may be (a) copied, photocopied, duplicated in any form by any means or (b) redistributed without the prior written consent of Messari®.
