DePIN and AI: Bridging the Future
The following is a written synopsis of a recent virtual panel discussion on DePIN and AI hosted by the Boston Blockchain Association featuring myself, Daniel Keller from RunOnFlux, and Greg Osuri from Akash.
Consider the below post the cliff notes to our conversion. If you would prefer you can watch the entire virtual panel here:
Introduction
In recent years, the integration of Decentralized Physical Infrastructure Networks (DePIN) and Artificial Intelligence (AI) has emerged as a transformative force in the tech world. These two technologies, when combined, promise to reshape industries by decentralizing resources and making AI more accessible, efficient, and resilient.
What is DePIN?
DePIN refers to networks of physical infrastructure — compute power, storage, and connectivity — that are distributed and decentralized. Instead of relying on centralized entities like Amazon Web Services (AWS) or Google Cloud, DePIN enables individuals and organizations to contribute their computing power to a shared, decentralized network. This approach not only provides resilience against failures and central authority but also creates economic incentives for participants. As noted during a Boston Blockchain Association (BBA) session, DePIN involves leveraging resources such as servers, GPUs, and even IoT devices like smart home systems to power decentralized networks. (DePIN 101)
The Intersection of DePIN and AI
AI’s rapid growth, particularly with models like ChatGPT and Llama, is putting unprecedented pressure on compute resources. Training and running advanced AI models requires massive amounts of processing power, typically provided by high-performance GPUs. However, the demand for GPUs is outstripping supply, leading to increased costs and bottlenecks.
This is where DePIN comes into play. By decentralizing access to compute resources, DePIN can aggregate the idle computing power available in homes, offices, and data centers around the world. This aggregated power can be used to support the computational needs of AI systems, providing an alternative to centralized cloud providers.
As discussed by experts from Akash and Flux during the BBA session, DePIN is not just an enhancement to AI — it is a necessity for AI to reach its global potential. Greg Osuri, the founder of Akash Network, emphasized that without decentralized compute power, AI’s growth will be constrained by the limited resources of centralized data centers. DePIN can unlock a new era of distributed computing, enabling AI to scale in ways that were previously unimaginable.
Why DePIN is Essential for AI
One of the key challenges in AI today is the cost and availability of GPUs. High-end GPUs like Nvidia’s H100, essential for training large AI models, are in short supply and expensive. Osuri noted that Akash was the first network to offer on-demand H100s at a much lower cost than traditional cloud providers. This affordability makes it possible for smaller organizations and even individual developers to access the computational power they need to experiment with and develop AI systems. Greg Osuri emphasized that “DePIN has become a necessity for AI” stressing the critical role of decentralized infrastructure.
Additionally, Dan Keller from Flux highlighted the risk of AI’s future being dominated by a few large corporations that control both the input and output of AI models stating, “It’s all about the resources and control.” This centralization poses significant risks, including the potential for biased AI systems and limited access to cutting-edge AI technology. By decentralizing the infrastructure needed to run AI, DePIN ensures that the development and operation of AI systems remain open, transparent, and accessible to everyone.
The Challenges Ahead
Despite the promise of DePIN and AI, there are still challenges to overcome. One of the major hurdles is ensuring the interoperability of different decentralized networks. As Keller pointed out, no single DePIN project can cover all use cases — AI will need to run on a combination of decentralized networks like Akash, Flux, and others. Achieving seamless integration between these platforms will be critical for the future success of DePIN in supporting AI.
Another challenge is regulatory oversight. As AI becomes more integrated into society, there will be increasing pressure from governments to regulate its use. The decentralized nature of DePIN provides some protection against overreach by centralized authorities, but it also presents new regulatory challenges that will need to be addressed. Keller predicts that AI regulation will become a key battleground in the coming years, and DePIN will play a crucial role in ensuring that AI remains open and decentralized. There was a strong discussion on the importance of what Dan termed a “Truth Model” in AI, with the consensus being that open-source and decentralized systems are the only way to achieve true transparency and fairness.
The Road Ahead
Looking to the future, the relationship between DePIN and AI will only deepen. Over the next five to ten years, we can expect DePIN to become a core part of the AI ecosystem, providing the compute resources needed to train and run increasingly complex AI models. As Osuri stated, AI cannot reach its full potential without the decentralized infrastructure that DePIN provides.
In conclusion, DePIN and AI together represent a powerful force for decentralization and innovation. By distributing compute power across a global network, DePIN democratizes access to AI technology, ensuring that the benefits of AI are available to all, not just a select few. As these technologies continue to evolve, they will unlock new possibilities for industries and individuals alike, making the future of AI more open, resilient, and scalable than ever before.
