London, January 2026 — FLock.io, a decentralised AI model development platform founded in 2022, has launched a new product called Flock Open Model Offering (FOMO), which it describes as a launchpad for users to publish AI models and receive cryptocurrency-based rewards when those models are used.
The company said FOMO is designed to make it easier for developers to deploy user-created models and integrate them into applications through an API. FLock.io framed the product as an alternative route to distribution and monetisation compared with centralised AI providers, where access and pricing are typically controlled by a small number of large platforms.
A token-driven approach to model pricing and rewards
FLock.io said FOMO is built around what it calls a “franchise” economic model, intended to lower API usage costs relative to traditional AI infrastructure.
Under this approach, revenue generated from the use of a particular model is used automatically to buy back and “burn” tokens linked to that model, reducing the token supply over time. The company said users can also stake those model-specific tokens to qualify for discounted API pricing.
FLock.io said the model is designed to encourage early usage by allowing model owners and token holders to subsidise inference costs at the beginning, while shifting more value toward participants who support a model as adoption grows. The company argues this structure differs from conventional cloud pricing, where costs are closely tied to compute usage and infrastructure overhead.
In practice, the cost reductions implied by the model would depend on factors such as inference demand, token market dynamics, the scale of the model, and how subsidies are structured. Like other token-based systems, the mechanism also introduces volatility risk, since token prices and liquidity can change rapidly.
How FOMO fits into FLock.io’s broader platform
FLock.io said FOMO completes its broader “full-cycle” ecosystem by connecting several components:
AI Arena, which the company describes as a training environment for machine learning models
FL Alliance, presented as a computation support layer
FOMO, positioned as the demand and distribution layer where models are launched and accessed via API
The company said this structure is intended to link model development, compute resources, and deployment into a single workflow, allowing models to move from training to real-world usage within the same network.
Company metrics and partnerships
FLock.io said it launched its mainnet in 2024 and has since seen increased platform activity. It reported that 10,007 fully trained machine learning models have been created on the network, and that more than 56,000 users hold its native FLOCK token.
The company also said it has been appointed as an AI strategic partner of the United Nations Development Programme (UNDP) and works with organisations including Alibaba, the UK’s NHS, Aigen Sciences, CIMG Inc., and Moorfields Eye Hospital. The announcement did not provide details on the scope, duration, or commercial terms of these relationships.
Positioning and longer-term goals
FLock.io’s CEO and founder, Jiahao Sun, said the company aims to support broader access to AI development and deployment through decentralised infrastructure and distributed ownership. The company argues that decentralised structures may improve transparency and reduce reliance on single providers, particularly for teams that want alternative governance or distribution models.
At the same time, decentralised AI platforms face open questions around performance consistency, security, compliance, and the reliability of token-based incentives over time. Adoption will likely depend on whether developers and enterprises see clear advantages in pricing, model quality, and integration compared with established cloud providers.
