Artificial intelligence is the most transformative technology of our time, but it’s also one of the most centralized. Big tech companies control the data, the models, and the infrastructure. Ashton Hettiarachi, Founder and Architect of OpenxAI and Openmesh Network, believes AI should be decentralised and democratised for all. His vision - An open-source, permissionless AI ecosystem, where anyone, anywhere, can build, own, and interact with intelligent systems without corporate gatekeepers. In conversation with HackerNoon, Ashton unpacks the vision behind OpenxAI, explains why decentralization is essential to the future of AI, and how we can unlock a more open and equitable intelligence ecosystem. What inspired you to start OpenxAI, and what gap did you see in the AI ecosystem that others weren’t addressing? AI is the most advanced technology that humans have ever invented. However, the power of AI is not distributed equally. Today’s AI does not follow democracy. It is a monopoly. AI is no longer just a tool. It is influence and power. AI is political. It is a modern version of the ark of the covenant. And right now, that power is concentrated in the hands of trillion-dollar corporations. These companies control the AI narrative, talent, infrastructure, rules, and even the regulations. They decide who gets access and at what cost. That is not just a monopoly. It is a chokehold on the future of intelligence itself. I started OpenxAI because I believe AI should be free, open, and accessible to everyone, like the internet. The gap was clear. If AI is so important, it cannot be centralized. It also needs to be decentralized, permissionless, and community-owned. OpenxAI is my contribution to that vision. It is a protocol that makes intelligence as open and unstoppable as the internet. I started OpenxAI because I believe AI should be free, open, and accessible to everyone, like the internet. The gap was clear. If AI is so important, it cannot be centralized. It also needs to be decentralized, permissionless, and community-owned. OpenxAI is my contribution to that vision. It is a protocol that makes intelligence as open and unstoppable as the internet. I started OpenxAI because I believe AI should be free, open, and accessible to everyone, like the internet. The gap was clear. If AI is so important, it cannot be centralized. It also needs to be decentralized, permissionless, and community-owned. OpenxAI is my contribution to that vision. It is a protocol that makes intelligence as open and unstoppable as the internet. OpenxAI Network is the first peer-to-peer permissionless AI protocol. What Bitcoin did for money, OpenxAI does for intelligence. With OpenxAI, anyone anywhere in the world can launch an AI company in under five minutes without needing a bank account, corporate approval, or millions of dollars in funding. That is the kind of change the AI ecosystem desperately needs How does the permissionless, peer-to-peer AI model network actually work, and how do you ensure trust and security within it? The same way IPFS or Filecoin allows anyone to store files permissionlessly and share them publicly, OpenxAI applies those principles to AI apps and agents. Anyone can host an AI agent or application, own the entire infrastructure and application layer, and share it in a peer-to-peer manner. Trust and security are built into the protocol. More importantly, the only person who can alter, update, or delete is the user. No one, not even a government or a core team member of OpenxAI, can change that. There are no middlemen or centralized choke points. Builders and consumers interact directly in a peer-to-peer manner. We also have an on-chain layer where builders or creators can write specific smart contract logic. They do not need to rely on traditional SaaS technologies. Payments, subscriptions, and security are all handled on-chain. This eliminates the need for third-party wallets or Web2 components, making the entire process seamless and decentralized. The “Skyscanner for Compute & GPUs” metaphor is compelling. What are the hardest challenges in standardizing and pricing compute fairly across providers? The compute industry is a lot like the hotel industry. Prices fluctuate depending on the region, the time of the month, the season, or even the minute. What we have done over the last four years is work with some of the largest data centers in the world, including Equinix, High Velocity, and Vultr, to build an universal API that aggregates pricing across a variety of servers. This allows us to scan and match pricing in a more efficient way. We call it Google Search for GPUs or Google Search for Compute. When a user wants to host an application or rent a compute environment, our search tool matches their requirements with the available supply from over 25,000 servers in our network. This ensures they get a compelling and affordable price. By working directly with wholesale providers, we bypass the retail pricing of AWS, Google Cloud, and Microsoft. This approach has allowed us to reduce costs by up to 80 percent, and in some cases, even 90 percent. OpenxAI claims over 80 percent cost savings and major energy and emission reductions. What drives those numbers, and how are they benchmarked? We work directly with wholesale providers like Equinix and Vultr, as well as innovative data centers that run on renewable energy. This allows us to offer significant cost savings while also giving users the option to prioritize green energy. By aggregating underutilized GPUs worldwide, we bypass the bloated overhead of centralized cloud providers. This reduces energy usage by up to 40 percent and cuts carbon emissions by 35 percent. These numbers are benchmarked against traditional providers like AWS and Google Cloud. Decentralization is not just cheaper. It is greener. Decentralization often introduces latency or reliability issues. How is OpenxAI tackling this trade-off? This is a great question. If you look at past attempts in decentralized compute projects like Golem, they faced challenges with reliability and latency. Golem started with the grand vision of letting anyone be part of the network. It was fully decentralized and fully peer to peer, but it introduced so many challenges that it compromised the core utility and the user experience. For example, if someone rented out their laptop but had a poor internet connection, the end user suffered. We chose a different route. OpenxAI is built on a reliable, large network of data centers. We work with world-class providers such as Equinix, High Velocity, and Vultr, which guarantee 99.99 percent uptime. These are the same providers trusted by S&P 500 companies, AWS, and IBM. This means OpenxAI users can enjoy the same reliability and latency as AWS. For optimal performance, we leverage fiber connectivity between continents using Equinix Fabric, which ensures high-speed, low-latency performance. While we plan to add more independent compute providers in the future, we prioritized reliability and low latency first to ensure the network delivers a high-quality experience from day one. Data privacy and ownership are big promises. How do you technically ensure users retain control over their models and data? In the Web2 world, when you rent a server from AWS or create an account with Hugging Face, your entire account is usually tied to your email or phone number. Many security experts have highlighted this as extremely vulnerable to hijacks, SIM swaps, phishing, and social engineering attacks. With OpenxAI, when a user rents a server, the encryption key is attached directly to their Web3 wallet. Instead of relying on outdated methods, we leverage the core of Web3: cryptography and distributed ledgers. All your data, applications, infrastructure, and backpack are connected to your Web3 wallet. This adds an additional layer of security on top of industry standards like SSH keys, which are trusted by S&P 500 companies. In our design, the only person who can access the server credentials is the user. For someone to compromise your data, they would need to breach your Web3 wallet or the Ethereum network itself, which is nearly impossible. As long as users protect their private keys, their data and their entire AI stack are as secure as Ethereum. How does tokenized monetization work for developers and model owners on OpenxAI? Launching an AI company today is like launching an internet company in the 1990s. Let’s look at how AI developers or entrepreneurs monetize their AI apps, services, or agents. Once they build the application, they need to create some form of SaaS system to track each user’s consumption and generate billing. For collecting and processing payments, the developer or startup has to sign up with a credit card processor like Visa, Mastercard, Stripe, or PayPal, which also take a cut of the revenue. To actually receive payments, the developer needs a bank account, either personal or corporate. If it is corporate, they must set up a legal entity and hire lawyers. Beyond the technology build, most AI founders go through a painful process that takes weeks, if not months, just to go from zero to one. We also have to acknowledge that not everyone has access to banking. Large parts of Africa, Asia, and the Middle East remain unbanked. AI is still trapped on 20th-century rails. Beyond the technology build, most AI founders go through a painful process that takes weeks, if not months, just to go from zero to one. We also have to acknowledge that not everyone has access to banking. Large parts of Africa, Asia, and the Middle East remain unbanked. AI is still trapped on 20th-century rails. Beyond the technology build, most AI founders go through a painful process that takes weeks, if not months, just to go from zero to one. We also have to acknowledge that not everyone has access to banking. Large parts of Africa, Asia, and the Middle East remain unbanked. AI is still trapped on 20th-century rails. OpenxAI has invented a completely new approach. Centralized cloud → Worldwide decentralized GPU and compute Expensive infrastructure → 80 percent cheaper than AWS or Google Bank accounts → Wallets IP and data privacy → Cryptographically wrapped compute and storage Company → ERC asset for ownership and revenue Credit card → On-chain subscriptions, usage fees, and micropayments Everything is on-chain. Everything is permissionless. Developers can use smart contracts to design business logic for their AI applications. You can configure how much compute equals how much revenue, set subscription fees, and embed all of this into the on-chain layer. Consumers can pay creators directly with zero middlemen, and developers can accept any form of crypto payment, stablecoins, ETH, or even memecoins. This eliminates the need for banks, credit cards, and intermediaries, empowering developers to launch their own AI services from their bedroom in a matter of minutes while tokenizing and monetizing their AI models, agents, and services seamlessly. Zero middlemen. Pure peer-to-peer AI. With no KYC, no region locks, and borderless access, how do you navigate regulatory or compliance risks? We believe in democracy, equality, and fairness. OpenxAI is an open-source, permissionless protocol, similar to Bitcoin. There is no company, no CEO, no foundation, and no board. It is a living and breathing digital organism, like the internet, IPFS, or Ethereum. Our network and technology are open and permissionless, meaning anyone can access them and there is no gatekeeping. We do not collect user data and we do not have the ability to shut down services. We are a platform and a network. Technically, anyone building on the network must follow the laws of their own jurisdiction. But unlike most AI platforms, we cannot enforce actions. Centralized systems can ban or remove someone from their ecosystem. With OpenxAI, even the creators cannot do that. If someone abuses the network, it becomes a law enforcement issue, not a protocol or OpenxAI issue. Regulators will need to catch up to this kind of innovation, just as they did with Bitcoin. Trying to regulate at the protocol level stifles innovation and is not the right approach. Imagine if the internet protocol itself had been regulated at the base level. Our world would look very different today. In a market with both centralized AI giants and open-source players, how does OpenxAI differentiate long-term? OpenxAI is not just open-source. It is permissionless and decentralized. That means anyone, anywhere in the world can host and deploy AI agents or services. A developer in Nairobi, a student in Mumbai, or a founder in São Paulo can launch their own ChatGPT-style business from a bedroom without needing a bank account, venture funding, or corporate approval. We are unlocking global AI entrepreneurship by opening the doors to regions and communities that have been excluded from the AI economy. The breakthrough is ownership. For the first time, anyone can actually own AI models. By tokenizing them, they can be traded, monetized, and shared directly, with no middlemen. This is not a feature. It is a new economic model. It will unleash a wave of AI entrepreneurship across the globe unlike anything we have seen before. And we are not predicting it, we are building it. For non-technical users, how do you make the decentralized infrastructure approachable and easy to adopt? We designed OpenxAI Studio with simplicity in mind. For non-technical users, the experience is as easy as signing up for Gmail. The web based console, X Studio allows anyone to launch AI services or agents without needing technical expertise. For advanced users, there are tools to leverage the full capabilities of the network. The goal is to make decentralized infrastructure feel as familiar as using platforms like Hugging Face. Give it a try and let us know if it's too hard. What are the biggest milestones or features on your near-term roadmap? Our biggest milestone is the launch of Studio v5 this October. Studio v5 is not just another product release. It is one of the most powerful Web3 AI platforms ever built. For the first time, both non-technical creators and professional developers will have the tools to build, deploy, and monetize AI services directly on chain. October will mark a turning point for OpenxAI, a month of breakthroughs and innovation that will set the standard for permissionless AI. Some announcements we cannot reveal yet, but the world will see them very soon. In November we expand beyond Ethereum to chains like BNB and Solana. With this expansion we are introducing cross chain AI provisioning, a completely new capability that allows builders to deploy an AI application on one chain while managing revenue or infrastructure on another. This opens the door to a multi chain future for AI that has never existed before. Entirely new use cases will emerge, many of which the industry has not even imagined yet. The start is great, no doubt. Looking three years ahead, how will you measure OpenxAI’s success? Three years is hard to predict, but within one year we expect to see significant adoption. Thousands of AI entrepreneurs will use our network to launch serious applications and businesses, running powerful AI services from their bedrooms. We anticipate more than 10,000 deployments in the first year alone. These creators will be able to rent, sell, and share intelligence itself, knowledge, datasets, and services, with no middleman and no borders. What the internet did for content creators, OpenxAI will do for AI creators. Just as YouTube turned anyone with a camera into a broadcaster and just as Spotify gave musicians a global stage, OpenxAI gives AI creators the same freedom. But this time, the opportunity is even larger. Imagine a student in Kenya running a ChatGPT-style service from their bedroom, offering it to the world, collecting payments, and expanding without banks or corporations. That is true power. With on chain monetization, developers can design new business models that were never possible before, micro subscriptions for individual agents, pay per query knowledge markets, community owned datasets, tokenized AI models that generate yield, and AI services that can be shared and traded like assets. This is not just about building applications. It is about unleashing a new creative and economic class, where intelligence becomes a shared global resource instead of a corporate monopoly. We are opening the door for millions, and eventually billions, of people to participate in the AI economy. It will be as transformative as the printing press, as the personal computer, and as the internet itself. That is true power. With on chain monetization, developers can design new business models that were never possible before, micro subscriptions for individual agents, pay per query knowledge markets, community owned datasets, tokenized AI models that generate yield, and AI services that can be shared and traded like assets. This is not just about building applications. It is about unleashing a new creative and economic class, where intelligence becomes a shared global resource instead of a corporate monopoly. We are opening the door for millions, and eventually billions, of people to participate in the AI economy. It will be as transformative as the printing press, as the personal computer, and as the internet itself.