Universal Basic Compute
Democratizing AI Power for Everyone
TLDR: Universal Basic Compute proposes that advanced computing power should be accessible to everyone as a fundamental right, not just tech giants—creating a system where AI resources are available to all students, educators, and communities.
Every Kid Should Have Access to Compute
Imagine it's finals week. Your roommate's fancy gaming laptop creates a virtual world in an AI-powered engine in seconds, while your five-year-old budget laptop blue screens trying to run the same thing. This isn't just annoying—this is now the new digital divide.
We talk about internet access as a necessity, but in 2025, having access to serious computing power is becoming just as essential. As coffee shops start training robots, and media companies pop up in home studios, the world will accelerate and grow at a breakneck speed. It should be a foregone conclusion that AI will transform everything from how we learn to how we work. Those without access to these computational resources risk being left behind.
Are we going to continue to pay Microsoft, Amazon, or Anthropic increasingly exorbitant fees for our access to AI? Or will we have access to computing the way we have access to clean water, gas, garbage collection or electricity from the city grid?
Surprisingly, a concept was proposed by Sam Altman, CEO of OpenAI: Universal Basic Compute (or UBC.) It's a concept that treats powerful computing resources like a public good—something everyone should have access to, regardless of their economic status.
The idea stems from the vision of Altman, but also Demis Hassabis, Emad Mostaque and other AI thought leaders - a post-AGI (Artificial General Intelligence) world where traditional labor will become obsolete, anything on a screen is automated in the near future, and we finally have robots doing our laundry and yard work. In this world, computational power is more valuable than any sort of currency.
Could we actually create a democratized computing network that creates needed processing power? As crazy as it sounds, we've actually done similar things before.
Public Networks
Long before I understood the politics of public funding, Big Bird and Terry Gross were quietly shaping my worldview. I absorbed spelling lessons from Muppets.
This public access revolution began in 1967 when President Johnson signed the Public Broadcasting Act, declaring broadcast spectrum a shared national resource. His "Great Society" vision provided the seed funding, but what made these networks sustainable was their innovative funding model: for every dollar of federal investment, they generated seven more through community support, foundations, and business sponsorships. Anyone who's lived through pledge week—with Mister Rogers temporarily interrupted for appeals for PBS tote bags—has experienced this community-powered approach firsthand.
The internet followed a similar path from public to private.
What now connects us to endless TikTok feeds began in 1969 as ARPANET—a Department of Defense project linking just four university computers. For years, this network remained the domain of researchers and academics until the National Science Foundation expanded access in the 1980s. Then came Tim Berners-Lee's HTML, transforming this government research project into the worldwide web. The internet wasn't born from corporate server farms but from a collection of networked computers with shared protocols—a public utility before becoming the largely privatized system we know today.
These technological revolutions demonstrate how transformative innovations often begin as public initiatives. Their success hinged on a crucial governance principle: distributed authority with shared standards. Local stations and network nodes maintained independence while following common technical specifications that ensured interoperability. This balance—local autonomy within a coordinated framework—mirrors the American federalist system and offers a compelling blueprint for how Universal Basic Compute might actually work.
Volunteer Computing to DePIN
Remember SETI@home? My Dad used to let his home computer screen saver search for ET!
Launched in May 1999, it let regular people donate their computers' idle processing power to search for alien signals. At its peak, it utilized more than 5 million personal computers worldwide—basically creating a supercomputer at a fraction of the cost. This proved people would voluntarily contribute computing resources to projects they found meaningful. The project operated for about 20 years before going into hibernation in March 2020.
SETI@home was the precursor to BOINC, Berkeley's shared computing initiative for research, which in turn led to Folding@home, a BOINC-powered shared processing initiative launched in 2000 that used collective computing to fold proteins. These volunteer computing networks proved the viability of shared computing.
Folding@home - Fighting disease with a world wide distributed super computer.
"Folding" in our community refers to running simulations of proteins, the molecular machines that perform most of the active processes we associate with life, from muscle contraction to sensing light and digesting food. We exploit the biological insight these simulations provide to inform drug discovery and other efforts to combat global health threats.
Decentralized Physical Infrastructure Networks, (also called DePIN,) are built on blockchain technology. They are similarly shared computing initiatives, but instead of sharing your computer for the joy of participation itself, your contributions are tracked and tokenized. One example in the space is Helium, a decentralized wireless network, where participants can purchase open-source hotspots and share their tokenized wifi with the world. Verizon and AT&T probably have nothing to worry about in the short term, but should these systems gain enough traction, they will prove a viable alternative to large telecom systems.
Helium - Own the Air
Helium allows anyone to build and own massive wireless networks.
When it comes to decentralized shared processing power networks, there are several notable players. Golem Network serves as a decentralized marketplace connecting providers of computational resources with requestors who need them, using GLM tokens as the medium of exchange. RenderToken focuses on GPU-based rendering services for 3D content creators.
Render Network
Try the leading decentralized GPU computing platform today at: https://rendernetwork.com/
Particularly fascinating is io.net, which has rapidly evolved since its founding in late 2023 to become a massive network with over 25,000 nodes. Io.net is now on a mission to assemble more than 1 million GPUs from independent data centers, crypto miners, and other projects like Filecoin and (the above mentioned) Render. Their goal is to create a decentralized alternative to traditional cloud providers, offering computing power at up to 90% lower cost.
io.net
The Internet of GPUs is finally here. Developed for AI startups, our decentralized GPU network delivers cost-efficient compute power to fuel innovation
What makes this approach interesting is how it aligns economic incentives with social goals. Instead of a centralized funding socialist initiative, it becomes a capitalist-driven distributed governance—an economic model humans have yet to truly experience. The transparency of blockchain ensures that every transaction between resource providers and users is visible and verifiable, creating trust without centralized control.
Admittedly, there is some speculation about the viability of these networks. Like many blockchain networks that have yet to find their validity, they are very early in their development. But the increasing validation of the Bitcoin network (as an example) may indicate a future for tokenized decentralized networks, which could be the underpinning for distributed artificial intelligence and computing power networks such as those discussed here.
The Computing Commons
Could Universal Basic Compute actually work?
Imagine an educational computing pool funded through donations and grants.
Just as public broadcasting stations conduct pledge drives, educational institutions could solicit contributions to a UBC cryptocurrency fund. These funds could be tokenized through any major cryptocurrency network with a shared wallet, governed by a collective of owners. The funds in these shared wallets could then purchase computing time from networks (like io.net,) making it available to students and educators. In addition to the fund, participants could "donate" idle processing power and earn tokens from the pool.
This vision aligns with the goals of Decentralized Physical Infrastructure Networks (DePINs), which offer a revolutionary approach by linking distributed computing resources into a cohesive global network of GPUs. These networks can be utilized for machine learning and other high-performance computing tasks, potentially serving as the infrastructure needed to deliver computational power to billions.
io.net has built its network on Solana, which appears to be a popular choice for DePIN projects. Considering the ease with which one can purchase Solana tokens, implementing a system where digital tokens equal processing tokens seems increasingly possible.
Web3 Infrastructure for Everyone
Fast. Decentralized. Scalable. Energy efficient. Solana can power thousands of transactions per second.
This approach could counter the centralization of AI by creating a commons-based alternative specifically serving educational and civic needs. Schools and communities could access the same AI tools as elite institutions. Community organizations could develop AI applications tailored to local needs without prohibitive costs. Considering the wild proliferation of open-source AI and platforms like Hugging Face, the knowledge and models will be just as available as the computing power.
Realistically, such systems would not be able to compete with centralized corporate AI players initially. There would be coordination challenges, technical complexity, and questions about long-term sustainability. The largest challenges would be cultural, because decentralization is a complicated subject for most non-technical people to understand. Any UBC initiative would need to start modestly, like the early ARPANET, perhaps serving specific educational needs in a few small communities before expanding.
However, public broadcasting faced similar skepticism, but if my early learning with Big Bird and the Count is any indication, it found mainstream success in early television development. It might be possible that these DePIN networks could find similar success as artificial intelligence becomes commonplace in everyday computing.
We are deciding, today, how AI networks should run and operate. AI will likely become as common as the electric grid or your water bill. The privatization of AI is something we should be greatly concerned about for a variety of reasons.
The decisions we make about computing infrastructure will shape how fairly its benefits are distributed. Universal Basic Compute offers a path away from centralized digital feudalism, and toward technological democracy—one where advanced computing power isn't a privilege for the few but a resource available to all.
Let's see how it turns out. If you find my ideas worth considering, please consider sharing with friends and subscribing.
Thanks for reading. We'll see you next time.
Nye Warburton is a creative technologist and educator. He lives in Savannah, GA. This essay was written and augmented with Perplexity, Anthropic Claude, and a variety of open source AI tools downloaded from Hugging Face. You can find out more about the author at https://nyewaburton.com.
