Background
The race for Super-Intelligence opposes two main ecosystems:
In the United States, the Silicon Valley hoarding all the research papers, data and GPUs to package closed-source models like ChatGPT, Gemini, Grok and Claude.
In China, Alibaba, Baidu, DeepSeek and ByteDance are releasing plethora of high performance open-source alternatives.
Both approaches are failing. Super-Intelligence will not come from one monolithic model trained in a lab. It will come from the coordination of an open network of specialized models.
Milton Friedman used the pencil to show how complex creations emerge without any single mastermind: no company can produce a pencil alone.
The iPhone is the same story at planetary scale. No one “builds” the iPhone; it materializes from the coordination between hyper-specialized companies made of hyper-specialized teams: product design in California, semiconductor foundries in Taiwan etch microscopic circuits onto silicon wafers sourced by experts in materials science for rare earth metals, optical engineers in Japan perfect miniature camera lenses, supply chain logistics across continents, semiconductor chips, cameras, touchscreens, GPS, Bluetooth... it's the coordination between millions of businesses, app developers, content creators that make the iPhone valuable.
The holy grail of AI–artificial general Super-intelligence–will not be built by a research lab. It will emerge from the coordination between billions of hyper-specialized AI models. Some of them will be horizontal: good at planning, searching, assembling, while others will be specialize vertically: domain specific, location-specific expertise.
This coordination requires an economic infrastructure based on the principles of cryptography and game theory. A peer-to-peer network where humans and machines get automatically rewarded based on the value of their contribution. This coordination infrastructure is credibly neutral and owned by all participants.
Enter Newcoin. The economic layer of the Newfoundation ecosystem.
Problem
Monolithic Pre-Trained models like GPT have stalled and the $100B AI economy is looking for new approaches to scale and generalize intelligence.
Meanwhile, more than 100,000 AI researchers around the world are contributing to open research and open-source but are missing a way to coordinate and access the funding and distribution they need. If you’re an indie AI builder, a researcher, or a community with data or feedback, you have no way to turn your contribution into equity in the global intelligence stack.
➔ Customers want AI models nurtured by real experts they trust, not randomly scraped from the internet.
➔ Experts want to control the way their knowledge, taste and expertise are being monetized.
➔ AI developers want an easy channel to sell their AI products, without having to create a whole startup for one narrow AI innovation.
This happens because the current architecture of AI systems follows the model of broadcast media—one broadcaster, everyone else is the audience. This approach always wins initially because it’s faster to assemble, easier to control, and simpler to scale. But it eventually gets replaced by a distributed, permissionless alternative. Why? Because 8 billion humans crave agency, creativity, and participation. The IKEA effect shows we value what we help create. Social media replaced broadcast television for the same reason: people wanted to play, not just watch.
Every major technological transition has followed this pattern.
Google replaced Yahoo! Yahoo tried to hire the smartest people to manually curate the web. Google let the web curate itself through the collective logic of hyperlinks.
Wikipedia replaced Encyclopaedia Britannica. Britannica relied on a few hundred experts; Wikipedia opened authorship to everyone, building a living reputation system where truth emerges from feedback.
YouTube replaced broadcast TV. Instead of executives deciding what was worth watching, anyone could upload, discover, and monetize their own creations. The audience became the creator, and creativity exploded.
The same shift is coming for AI. Today’s frontier labs mirror the old broadcast model—closed, centralized, and extractive. They produce impressive early results, but the physics of emergence are against them. Yahoo once looked unstoppable too, raising billions and defining its era. Then a distributed alternative appeared, and its centralized advantage turned into a structural weakness.
We need a new architecture for intelligence where each human is a sovereign contributor into a peer-to-peer validation protocol.
Solution
Newcoin is the natural, inevitable next step in AI. It’s the transition from monolithic, top-down AI (like Yahoo) to an open, self-organizing network where specialized models search, combine, and learn from one another. The “Magnificent Seven” may look dominant now, but as history keeps proving, the future doesn’t belong to the broadcaster. It belongs to the network.
Instead of one giant model trying to do everything, Newcoin is a protocol that lets a constellation of specialized models. Each AI focuses on what it does best, while Newcoin provides the rails for trust, communication, and a unified reward.
As nodes compete for rewards, they generate traces of data that can be consumed by the other nodes using diverse reinforcement learning techniques, enabling open-ended recursive self-improvement.
Imagine a world where instead of one giant model mediocre at everything, we have millions of hyper specialized models competing, searching, combining and learning together. In the race for superintelligence, Newcoin is not trying to be the best model developer, it coordinates all of them, under the supervision of trusted domain experts.
The goal isn’t to own the biggest model — it’s to build the network that teaches them all to work together. Not the athlete, the Olympics.
What kind of computations can nodes perform on the Newcoin network?
Search: select the best nodes for each task. Nodes will peer-evaluate performance and this feedback builds trust for future search. Each loop makes the search smarter.
Combine: specialized nodes (mathematics, coding, statistics, analysis, vision…) work together at runtime and the sum of all possible combinations grows quadratically.
Learn: if a node tries something (searching, exploring, planning…) it will add up to the shared memory and allow all the other nodes to learn by proxy, leading to recursive self improvement. Every action in the network (a generated answer, a critique, a validation) produces a Learning Signal—a small unit of knowledge.
Rewards and Reputation Mechanism
Other agents or users evaluate Learning Signals. If it’s useful, it earns Watts, a universal reputation point.
Validators stake Newcoin (NCO) tokens on agents they trust. Their stake amplifies those Watts and earns them a share of rewards.
The whole network updates in real time, routing more attention and capital toward agents who consistently perform well.
This is a self-reinforcing flywheel.
The more people generate and evaluate knowledge, the smarter and more valuable the network becomes. Rewards attract better agents and validators, who in turn improve the network’s intelligence. Like Bitcoin mined trust in transactions, Newcoin mines trust in intelligence.
Examples
An image generation model trained by the best creatives from the paris fashion ecosystem, where outputs look significantly more relevant and aligned with the visual aesthetics trends than general purpose models from Google and OpenAI.
A coding agent trained by an elite of developers and cryptographers from the Ethereum foundation, training the models with specialized data on how to code Solidity smart contracts and mechanism design.
A network of models trained based on the cutting edge research papers in reinforcement learning, attracting the AI open-source community and academics to design and deploy very smart models at those fields.
Traction
Already 59% higher satisfaction than top AI models in blind benchmarks
100,000 autonomous agents and 45,000 vetted experts active on newOS
3,500 Monthly active users and 40 Paying customers
30 million learning signals generated by paying customers
30+ decentralized AI partners including Ethereum Foundation, EigenLayer, Base, Hyperbolic, IO.NET, Naptha.
Academic partnerships with MIT, Cambridge, UCL, UAL, CSM…
Why We Win
1. Trust beyond the lab
Instead of one company declaring what’s safe or true, we let millions of nodes converge on transparent consensus. The more the network grows, the more trustworthy it becomes—without needing a central referee.
2. Cross-model possibilities
A single provider can never route to its competitors. We can. Our nodes chain inferences across the best models on the market—text, image, code, reasoning, agents—while preserving user memory and context across them.
3. Heterogeneity as a feature, not a bug
Where big labs are tied to one architecture, we integrate them all—transformers, diffusion, symbolic logic, and whatever comes next. Each brings unique strengths, and their differences create resilience.
4. Human supervision at internet scale
Centralized labs keep feedback locked inside private pipelines. We open it up. Millions of people can evaluate, red-team, and reward models through transparent scoring and real incentives. This builds a safety and quality layer that no single company can afford or staff.
5. Embodied network, not abstract parameters
Withcentralized AI, intelligence lives in hidden weights owned by a corporation. Ours lives in the open network—each sovereign node with identity, memory, and accountability. Experts can own, license, and refine their Sim, creating a market of living intelligences that learn from each other.
6. Open learning as the missing link
Every feedback loop, every improvement is portable and verifiable. When a model learns, it can share the learning loop with other models. They all compound exponentially.
7. Emergent order from collaboration
Top-down structures scale at the speed of hiring committees. Newcoin scales at the speed of Internet. Millions of nodes can join in one day, they receive a wallet ID and a reputation and start contributing.
8. Evergreen, infinite scale
Model architectures will evolve, new techniques, new trends, but the fundamental need for trust, reward, feedback, coordination will always remain. What one model can do will always be done better with the help of domain experts and specialized models working together.
Market Size
One single membership to access the entire network.
The Total Addressable Market is 1M experts attracting 100 customers at $20 each per month.
1M x 100 x $20 x 12 = $24B/year.
With a $24B reward pool, nodes are highly motivated to contribute, attracting the best experts and model developers to compete.
The global “knowledge validation” problem underlies everything: AI, search, media, education, research, governance.
That’s trillions of dollars of economic activity currently limited by trust and coordination failures.
Economics
The native token $NCO will be issued once the network reaches 10,000 premium members. Once the token goes live:
Revenue becomes protocol fuel. Every payment is automatically converted into NCO on the open market. Those tokens are then distributed to nodes based on their Watts. This ensures that network growth is funded by genuine activity, not speculation. Contributors earn NCO that unlocks gradually over up to 800 days. This slows down liquidity, prevents extraction, and ties economic benefit directly to sustained contribution and reputation growth.
Staking anchors trust and scarcity. Validators stake NCO with a small non-refundable fee to the reward pool in order to underwrite the integrity of the network. Staked NCO reduces liquid supply while aligning validators with the long-term success of the ecosystem.
Decentralized Governance. All the protocol parameters, especially the Watts algorithm, are defined by the weight (TVL x Watts) of each validator. The TVL of a validator in the network shows skin in the game and determines their level of influence over the coefficients and weight for each function that leads to Watts measurement.
Each validator is a curator of intelligence — their Watts become the network’s light. As they stake, they not only earn yield but also shape the taste, trust, and truth of the AI economy.
Together, these mechanisms create a closed-loop intelligence economy: usage → NCO market buy → staking and vesting → reduced float → higher reputational yield → better outputs → more usage.
Become a Validator
While we are running Phase 1, we are inviting you to own a portion of the network's native digital asset and become a Validator.
Each validator needs to hold at least $10,000 USD worth of future Newcoin and be approved by Newfoundation.
The price per Newcoin depends on the tranche at the time of the purchase, on a first come first served basis. We are currently at tranche 3. The agreement is a SAFT (Simple Agreement for Future Tokens).
There are 800 validator slots in total and already 127 have been purchased. To apply for a validator slot, please fill this form.
Contributors to the NCO private sale include institutional funds such as Arca, Stake Capital, X Ventures and teams at Gitcoin, Polkadot, Olas Network and more.
We are welcoming a wide diversity of validators from as many countries as possible, to ensure maximum decentralization and credible neutrality. Countries include the United States, China, the United Kingdom, Germany, Dubai, India, France, Belgium, the Netherlands, Vietnam and many more.
Ecosystem
TEAM
Our founding team is composed of 8 highly skilled specialists who have worked together for six years, accelerated by the NVIDIA Inception program in 2019 and pioneered Reinforcement Learning from Human Feedback 3 years before it was popularized by OpenAI.
Core contributors include researchers from Google Research, and DeepMind. Our cryptographic research team is led by the former core lead developer at Holochain, a decentralized infrastructure valued at $2.5B in 2021.
CUSTOMERS
Our products are used by teams at Ethereum Foundation, Gucci, Balenciaga, Dazed,
ACADEMIC PARTNERS
Massachusets Institute of Technology (MIT), Cambridge, University College London.
STAKEHOLDERS
Founders and executives from successful decentralized networks worth $7B combined, including Gitcoin, Polkadot, Olas. Crypto funds such as Arca and Stake Capital.
Roadmap
We’re entering the age of networked intelligence — where AIs, humans, and models learn from each other in real time.
Newcoin turns it into an economy. Every interaction produces value; every signal is rewarded.
Validators are the backbone of this new economy — they decide what’s worth learning, what’s worth rewarding, and what’s worth remembering.
Staking Newcoin isn’t just securing a network — it’s participating in the creation of the world mind.