Decentralised AI is breaking through as a practical and scalable alternative to traditional AI. Tokenised crypto incentives are solving the GPU shortage, and DePINS are on the rise. X402 introduced agentic blockchain-based web payments last year. Data privacy is getting prioritised with techniques like federated learning. Here’s to an exciting 2026!
ChatGPT launched in late 2022. It’s now 2026. It might feel like only yesterday, but three years have passed since the breakout year of generative AI. It didn’t take long before the dominating centralised corporations faced searing opprobrium, and decentralised AI (DeAI) ventures like FLock emerged as a fairer alternative in 2024.
Last year, AI was marked by the move from pilots and experimentation to reinventing entire workflows at the infrastructure level. In 2025 it also dawned on the business, public and academic worlds alike that privacy-preserving DeAI is not just more secure and transparent but more powerful and accurate. It’s transitioned from speculative hype and whitepapers to a functional on-chain economy with practical adoption by public bodies like the United Nations Development Programme (UNDP) and the NHS.
It’s been riveting to watch it all pan out. This time last year, our annual round-up of industry news recounted how the world woke up to centralised AI’s pitfalls. The warning signs haven’t slowed since.
The good news? The alternative is finally breaking through in a real way: we are part of a mission to cultivate an equitable, community-driven ecosystem with emphasis on transparency, scalability, lower costs and extra powerful models.
[ 👋 Hi there! If you’re here to find out more about FLock.io, follow us on X, read our docs and sign up to AI Arena on train.flock.io. We just launched FOMO, completing the full-cycle DeAI platform.]
This blog talks through the state of the DeAI industry as it stands in January 2026: trends and key moments. See our FLock 2025 year in review here.
Tokenised incentives are solving the compute shortage
AI is increasingly embedding itself into the crypto economy. It’s felt at every layer of the stack, from smart contracts and security to consumer DeFI frontends.
The marriage between AI and blockchain has matured, driven by tokenised incentives. These on-chain tokens solve bottlenecks in AI development such as compute shortages and ward off bad behaviour.
Unlike in big tech, with DeAI there’s no single corporation with deep pockets and a huge data centre. Fortunately, there’s no need. Why? Because tapping into a global network of expertise, idle computing power and diverse data is actually vastly more convenient, efficient and innovative. Crypto incentives ensure a steady stream of high-quality contributions.
At FLock, by issuing tokens we lower the cost floor for AI inference. Until FOMO launched last week, the standard prices were set by cloud providers (like AWS and Alibaba Cloud) and gateways (like OpenRouter and SiliconFlow). Their prices are dictated by the cost of compute, since they cannot price inference below hosting costs without making a loss.
Ocean Protocol, however, faced backlash in October after allegedly dumping $120 million in Fetch.ai tokens, prompting public accusations and legal threats. This happened after Ocean left the Artificial Superintelligence Alliance to regain its independence. The exit fractured the crypto AI supergroup, which had been formed as an open-source alternative to corporate AI.
X402 introduced agentic web payments
The internet was designed without a native payment layer, creating decades of friction. Yet, a hidden piece of infrastructure has been waiting for its moment: the HTTP 402 “Payment Required” status code. In May 2025, Coinbase launched x402: the on-chain payment gateway for AI and APIs. This revolutionary protocol finally turned that long-unused code into a powerful engine for instantaneous, blockchain-based micropayments across the web.
It’s so great that FLock couldn’t resist adopting it. We are creating a unified gateway for on-chain Agents to pay for the brains (AI Models) that drive them with X402. Read this blog to learn more.
DePINs have grown as GPU demand surges
The AI industry has an insatiable appetite for compute as inference demand has skyrocketed over the past year. September’s UK-US tech prosperity deal revealed plans for several enormous data centres to keep up, but centralised servers compromise security and privacy.
For tackling the global AI GPU shortage, DePINs (Decentralised Physical Infrastructure Networks) are becoming the go-to solution. Networks like io.net and Akash allow users to lease idle GPU power, with tokens providing an incentive mechanism that can reduce costs by up to 70% compared to traditional cloud providers.
In 2025, DePINs transitioned from a niche sector to an infrastructure revolution, with the industry projected to grow from a $50 billion+ market cap in 2024 toward a potential $3.5 trillion by 2028. Startups on such networks surged by 40% in October 2025, driven by soaring energy costs and the need to bypass centralised, energy-intensive cloud servers.
The relentless demand for compute, combined with the advancing capabilities of open-source models, provides decentralised compute networks with new revenue streams, according to a Messari report.
Data privacy is getting prioritised with techniques like federated learning
Data privacy is non-negotiable in the public sector, healthcare, finance and beyond. The problem with traditional AI is that it sends data to a centralised server for processing, presenting significant security risks due to the single point of failure for all collected data.
Federated Learning (FL) focuses on keeping data local. It’s a more privacy-preserving, decentralised alternative to traditional ML, allowing collaborative model training without exchanging raw data. Developed in 2017, it’s in its proof-of-concept stage. It involves training statistical models over remote devices or siloed data centres, such as mobile phones or hospitals, while keeping raw data localised.
With AIGEN Sciences and Professor Jaewoo Kang, co-creator of BioBERT, FLock applied privacy-preserving AI to drug discovery, enabling hospitals and research institutions to collaborate securely without sharing sensitive data.
We are sometimes asked: “Why do you need a blockchain for your technology? Is this traditional AI wrapped in a crypto buzzword?”. The answer is: definitely not. Blockchain is essential to our delivery of collaborative but secure models, and it has little to do with crypto cliches. It’s so paramount that we’re named after it: FLock stands for FL on Blockchain. Read this blog to find out how blockchain acts as a core security and decentralisation tool to solve systemic flaws in traditional FL.
AI as infrastructure is the latest wave in business
There have so far been two AI waves. The first wave strived towards efficiency and cost-cutting by treating AI as an assistant. In the second wave, businesses undergoing a new kind of digital transformation are reinventing entire processes. See our blog on it here.
By the end of last year, most organisations were still only experimenting with or piloting AI, rather than capturing enterprise-level value, according to a McKinsey report. But the few high performers intend to use AI to redesign and orchestrate workflows.
As The Upwork Research Institute found: “introducing new technologies into outdated work models and systems is failing to unlock the full expected productivity value of AI.”
The conclusion is clear. Organisations still piloting clumsy, optional tools are only hampering their productivity. Employees getting burnt out from AI even has a name now: AI fatigue. As many as 80% of employees report that AI has increased their workload, because companies are treating AI as an extra rather than inextricable from the business process.
The way forward is to reimagine the core business infrastructure and embed autonomous agents inside.
When you use AI as an assistant (like spellcheck, a plug-in suggesting your next sentence, or ChatGPT for brainstorming ideas), it’s layered on top. If you drop it, the underlying work process stays the same. When AI instead becomes the orchestrator of the workflow, it becomes process infrastructure that directs the flow of labour. An example is a taxi driver using Google Maps to suggest a faster yet optional route, versus Uber deciding everything from which driver gets the job to monitoring the journey.
DeAI has proved itself as a force to be reckoned with – here’s to an exciting 2026!
Nearly two years ago, Stability AI CEO Emad Mostaque stepped down to “pursue decentralised AI”. Now, he’s founded Intelligent Internet focusing on exactly that. He’s not alone in the mission to break the monopoly of Big Tech over AI and the profit-hungry black boxes that come with it.
The DeAI ecosystem is rich and expansive, so we can’t name all projects (although, in late 2024 we did put together a taxonomy that classified the existing protocols). But here’s a small handful: Akash Network (AKT), Ocean Protocol (OCEAN), io.net and FLock.io (that’s us!).
Here’s to hoping our next DeAI industry snapshot will recount even greater steps in the right direction by this mission FLock is a part of.
