Demand for Cloud GPUs is Soaring Due to Surge in Web3 Compute, says Akash Network CEO Greg Osuri
In Brief
Akash Network CEO Greg Osuri shared his vision for GPU incentives program and how the company is easing cloud computing access for Web3.
A batch of Nvidia chips are now accessible on Akash Network, a cloud computing platform built on blockchain. By offering access to idle chips within this peer-to-peer marketplace, Akash Network now provides Nvidia H100s and A6000s GPU access over the cloud, at competitive prices compared to other cloud service providers.
This aligns with Akash Network’s efforts to address increasing demand and signifies the commencement of its incentives pilot program, which includes a $5 million investment aimed at granting developers access to high-performance GPUs.
In a conversation with Metaverse Post — Greg Osuri, the CEO of Overclock Labs and Akash Network, shared insights into the motivations behind launching the program and broader developments for cloud computing in fostering Web3.
The increasing demand for GPU cloud computing of AI and Web3 usecases is swiftly rising and concentrating within major tech companies. The role of players such as Akash Network is to provide a variety of high-performance GPUs to alleviate this pressure for developers in organizations of various scales. Given the company’s blockchain foundation, its peer-to-peer GPU marketplace facilitates secure and cost-effective access to sought-after resources.
The company provides access to a set of 128 Nvidia H100s and A6000s at a significantly reduced cost compared to other services providers, priced at $1.49/hr and $0.49/hr, respectively.
“Sky-high price points from hyper scalers have forced startups out of the innovation race, and even larger players are on years-long waitlists to gain access to the resources they need for their AI workloads,” Greg Osuri told Mpost.
Osuri believes that the incentives pilot program, introduced with a $5 million allocation to subsidize high-performance GPUs, will facilitate access to more computing resources for developers on Akash. Providers who contribute GPUs to a homogeneous configuration pool will receive periodic earnings, potentially up to $750,000. For developers, the program offers opportunities for cost-optimized, high-quality, and reliable compute.
The incentives are expected to bring more than 1,000 A100s, and the program is projected to run for 120 days. Providers will be incentivized through three methods, including the committed pool, liquidity mining pool, and R&D Pool.
The committed pool will offer high-quality computing resources by involving professional providers of compute, typically tier-2 or higher data center operators, committed to supplying computing power to Akash Network for at least one year. Akash Network proposes to allocate 40% of the Pilot budget, approximately $3,500,000.
This distribution includes $1,800,000 for 83 H100 chipsets, $800,000 for 54 A100 80 GB chipsets, and $900,000 for 105 RTX A6000 chipsets. In the liquidity mining pool, providers will contribute GPUs to a homogeneous configuration pool, which allocates a fixed number of tokens per epoch distributed pro-rata to all participants.
Earnings for each provider will be distributed periodically, with a cadence allowing Akash Network sufficient time to verify the accuracy, uptime, and other critical metrics. Akash Network plans to allocate 40% of the Pilot budget, approximately $750,000, to any Nvidia GPU with at least 24GB VRAM.
Additionally, 20% of the program budget, totalling $750,000, will be allocated to the R&D Pool. This segment will be exclusively dedicated to GPUs launched in 2024 or later, covering both Nvidia and AMD GPUs.
“Tenants can come to Akash Network with a budget in mind and secure access to GPUs by placing a bid on the network. From there, providers can review the bid and offer their proposed rate to provide access to the resources agreed upon,” said Greg Osuri.
The Overclock Labs team behind the Akash Network doesn’t influence the bidding process due to the decentralized, peer-to-peer nature of the network. This setup addresses concerns related to centralization and censorship.
“By offering access to the most sought-after chips on the market, Akash Network gives startups a chance to compete next to the big tech players. By giving permissionless access to compute resources, including Nvidia A100s and H100s from a range of providers – from independent to hyperscale – alternative networks are uniquely positioned to mitigate inefficiencies,” added Akash Network’s Osuri.
Akash Network’s Long-Term Goals
Greg Osuri envisions that with the latest batch of 128 Nvidia A100s and H100s on the network, Akash Network will become one of the few alternative solutions in the market that continuously evolves to meet consumer demand.
“As high-density GPUs remain in high demand, we’ve had to expand our provider resource pool to ensure developers on Akash can continue to have cost-optimized access to high-density GPUs for training and fine-tuning AI models and scale AI inference apps and virtual world renderings,” he told Mpost.
The program enhances Akash Network’s flexibility and positions us to be the first to market with new GPUs as AI models and GPU architectures evolve.
Akash Network operates similarly to “Airbnb” for chips, allowing those with idle chip capacity to rent compute resources on the network. With over 75 providers offering GPUs, CPUs, and various compute forms in the United States, Europe, and Asia, Akash is scaling its operations by bringing more compute capacity providers onboard and encouraging community contributions.
While sharing insights into the broader developments and advancements anticipated from Akash Network Osuri said, “Our team is dedicated to providing cost-effective resources to encourage the enterprise adoption of blockchain, as seen through our partnership with health tech platform Solve.Care, and significantly reducing cloud costs for companies like the metaverse platform Passage.”
Disclaimer
In line with the Trust Project guidelines, please note that the information provided on this page is not intended to be and should not be interpreted as legal, tax, investment, financial, or any other form of advice. It is important to only invest what you can afford to lose and to seek independent financial advice if you have any doubts. For further information, we suggest referring to the terms and conditions as well as the help and support pages provided by the issuer or advertiser. MetaversePost is committed to accurate, unbiased reporting, but market conditions are subject to change without notice.About The Author
Alisa is a reporter for the Metaverse Post. She focuses on investments, AI, metaverse, and everything related to Web3. Alisa has a degree in Business of Art and expertise in Art & Tech. She has developed her passion for journalism through writing for VCs, notable crypto projects, and scientific writing. You can contact her at alisa@mpost.io
More articles
Alisa is a reporter for the Metaverse Post. She focuses on investments, AI, metaverse, and everything related to Web3. Alisa has a degree in Business of Art and expertise in Art & Tech. She has developed her passion for journalism through writing for VCs, notable crypto projects, and scientific writing. You can contact her at alisa@mpost.io
