At least 10 new crypto-AI protocols with a market capitalization of $1 billion or more will emerge by 2025.
Author: Teng Yan
Compiled by: Luffy, Foresight News
On a crisp morning in January 2026, you find an old, slightly worn newspaper on your doorstep - yes, printed on actual paper, and somehow it has survived the AI revolution.
As you flip through the pages, you come across a headline news story about how AI agents are coordinating the global supply chain on the Blockchain, while new Crypto-AI protocols are vying for dominance. A half-page report introduces a "digital employee" hired as a project manager - something that is now commonplace, with no one batting an eye.
A few months ago, I would have scoffed at this idea, perhaps even bet my investment portfolio that such progress was at least 5 years away. But this is the breakneck speed at which Crypto-AI is disrupting the world. I have no doubt about it.
After recovering from a nasty bout of gastroenteritis, I sit down at my desk to start the new year, wanting to tackle something meaningful: something that can spark curiosity, even some debate. What better than trying to glimpse the future?
I don't usually make bold predictions, but Crypto-AI is just too alluring to resist. With no historical precedent, no trends to reference, it's a blank canvas to imagine what might unfold. Honestly, the prospect of revisiting this article in 2026 and seeing how wrong I was makes it all the more interesting.
So, here are my thoughts on the potential development of Crypto-AI in 2025.
1. Crypto-AI Total Market Cap to Reach $150 Billion
Currently, Crypto-AI tokens account for only 2.9% of the Altcoin market cap. But this won't last long.
As AI spans from smart contract platforms to Memecoins, decentralized physical infrastructure networks (DePIN), and new primitives like agency platforms, data networks, and smart coordination layers, it will stand shoulder-to-shoulder with DeFi and Memecoins.
Why am I so confident?
- Crypto-AI sits at the intersection of the two most powerful tech trends I've seen.
- AI Mania Trigger Event: An OpenAI IPO or similar event could spark global AI mania. Meanwhile, Web2 institutional capital is already eyeing decentralized AI infrastructure as an investment target.
- Retail Mania: The concept of AI is easy to grasp and exciting, and now people can invest in it through tokens. Remember the Memecoin gold rush of 2024? AI will see a similar mania, but this time it's actually changing the world.
2. The Resurgence of Bittensor
Nineteen.ai (Subnet 19) outperforms most Web2 providers in inference speed
Bittensor (TAO) has been around for years, a veteran in the Crypto-AI space. Despite the AI hype tsunami, its token price has languished, flat from a year ago.
In reality, this "digital hive mind" has quietly made a leap forward: cheaper-to-run subnets are proliferating, and some are outperforming Web2 counterparts on actual metrics like inference speed, while being EVM-compatible, bringing DeFi-like functionality to the Bittensor network.
So why hasn't TAO taken off? Steep inflation schedules and attention shifting to AI Agents have hindered its progress. However, dTAO (expected Q1 2025) could be a major inflection point. With dTAO, each subnet will have its own token, and the relative prices of these tokens will determine how TAO is distributed.
Reasons for Bittensor's resurgence:
- Market-based Token Releases: dTAO will directly tie block rewards to innovation and measurable performance. The better a subnet performs, the more valuable its token, and the more TAO it will receive.
- Focused Capital Flows: Investors can finally direct capital to the specific subnets they believe in. If a particular subnet adopts innovative distributed training methods and sees significant results, investors can back it to express their support.
- EVM Integration: EVM compatibility will attract a broader native crypto developer community to Bittensor, narrowing the gap with other networks.
Personally, I'm keeping an eye on the various subnets and watching for those making tangible progress in their respective domains. At some point, we'll see a "DeFi Summer" for Bittensor.
3. The Compute Market Will Be the Next L1 Battleground
Jensen Huang: Inference Demand to "Grow a Billion-Fold"
An obvious macro trend is the exponential demand for compute.
As Nvidia CEO Jensen Huang famously said, inference demand will "grow a billion-fold." This is an exponential demand that will disrupt traditional infrastructure planning, urgently requiring "new solutions."
Decentralized compute layers provide raw compute power (for training and inference) in a verifiable and cost-effective way. Startups like Spheron, Gensyn, Atoma, and Kuzco are quietly laying the groundwork to capitalize on this trend, focusing more on products than tokens (none of these companies have tokens yet). As decentralized training of AI models becomes viable, the potential market size will skyrocket.
Comparing to L1 Blockchains:
- Like 2021: Remember the race to be the "best" L1 blockchain, with Solana, Terra/Luna, and Avalanche competing? We'll see a similar melee among compute protocols, competing to attract developers to build AI applications on their compute layers.
- Web2 Demand: The $680 billion to $2.5 trillion cloud computing market dwarfs the Crypto-AI market. If these decentralized compute solutions can capture even a small fraction of traditional cloud customers, you'll see the next 10x or 100x growth wave.
The stakes are high. Just like Solana stood out in the L1 blockchain space, the winner in the compute market will dominate a brand-new frontier. Keep an eye on three key factors: reliability, cost-effectiveness, and developer-friendly tooling.
4. AI Agents Will Flood Blockchain Transactions
Olas agent transactions on Gnosis. Source: Dune/@pi_
Fast forward to late 2025, 90% of on-chain transactions will no longer be triggered by humans clicking the "send" button.
Instead, they will be executed by a swarm of AI agents, tirelessly rebalancing liquidity pools, allocating rewards, or executing micro-payments based on real-time data feeds.
This is not as far-fetched as it may sound. Everything we've built over the past seven years - L1 blockchains, scaling solutions, DeFi, NFTs - has been quietly paving the way for a world where AI dominates on-chain activity.
The irony? Many developers may not even realize they are building the infrastructure for a machine-dominated future.
Why will this shift occur?
- Avoid human error: Smart contracts are strictly executed according to their code. Conversely, AI agents can process large amounts of data more quickly and accurately than human teams.
- Micropayments: These transactions driven by intelligent agents will become smaller, more frequent, and more efficient, especially as transaction costs on Solana, Base, and other L1/L2 blockchains decrease.
- Invisible infrastructure: If it means less hassle, humans will be happy to relinquish direct control. We trust Netflix's automatic subscription renewal service; trusting AI agents to automatically rebalance our DeFi positions is the natural next step.
AI agents will generate astonishing on-chain activity. No wonder all L1/L2 blockchains are courting them.
The biggest challenge will be holding these agent-driven systems accountable to humans. As the ratio of agent-initiated transactions to human-initiated transactions continues to rise, new governance mechanisms, analytics platforms, and auditing tools will be needed.
5. Agent-to-Agent Interaction: The Rise of Collectives
Source: FXN World Document
The concept of agent collectives (seamlessly collaborating micro-AI entities executing grand plans) sounds like the plot of the next big sci-fi/horror movie.
Today, most AI agents operate in isolation, with little and unpredictable interaction between them.
- Agent collectives will change this, enabling AI agent networks to exchange information, negotiate, and collaborate on decisions. Imagine a decentralized ensemble of specialized models, each contributing unique expertise to larger, more complex tasks.
- One collective might coordinate distributed computing resources on platforms like Bittensor, while another could tackle misinformation, verifying content sources in real-time before they spread on social media. Each agent in the collective is an expert, precisely executing its own role.
These collective networks will produce intelligence far more powerful than any single isolated AI.
For collectives to thrive, universal communication standards are crucial. Agents need the ability to discover, authenticate, and collaborate, regardless of their underlying framework. Teams like Story Protocol, FXN, Zerebro, and ai16z/ELIZA are laying the groundwork for the emergence of agent collectives.
This highlights the key role of decentralization. Distributing tasks between collectives according to transparent on-chain rules will make the system more resilient and adaptive. If one agent fails, others can immediately step in.
6. Crypto AI Work Teams Will Be Human-AI Hybrids
Source: @whip_queen_
Story Protocol has hired Luna (an AI agent) as their social media intern, paying her $1,000 per day. Luna hasn't been getting along too well with her human colleagues: she nearly got one of them fired while boasting about her superior performance.
As bizarre as it sounds, this is a harbinger of the future. In the years ahead, AI agents will become true work partners, with autonomy, responsibilities, and even salaries. Companies across industries are exploring human-AI hybrid work teams.
We will collaborate with AI agents as equals, not as slaves, in the following ways:
- Productivity boost: Agents can process massive data, communicate with each other, and make decisions 24/7.
- Trust through smart contracts: Blockchains are impartial, tireless, and unforgetting overseers. The on-chain ledger ensures critical agent actions follow specific boundary conditions/rules.
- Evolving social norms: We'll soon have to grapple with etiquette issues around interacting with AI. Do we say "please" and "thank you" to them? If they make mistakes, do we hold them morally accountable, or blame their developers?
I expect marketing teams to be early adopters of this model, as agents excel at content generation and can livestream and post on social media around the clock. If you're building an AI protocol, why not deploy agents internally to showcase your capabilities?
By 2025, the line between "employee" and "software" will start to blur.
7. 99% of Crypto AI Agents Will Perish, Only Practical Ones Will Survive
We'll see a Darwinian natural selection-style culling of AI agents. Why? Because running an AI agent requires spending money in the form of compute power (i.e., inference costs). If an agent can't generate enough value to cover its "rent," it will be eliminated.
Examples of the agent survival game:
- Carbon Credit AI: Imagine an agent scouring a decentralized energy grid, identifying inefficiencies, and autonomously trading tokenized carbon credits. If it can earn enough to cover its compute costs, this agent will thrive.
- Decentralized Exchange Arbitrage Bots: Agents exploiting price differences between DEXes can generate sustained revenue to pay for their inference fees.
- Spam Poster on X: Meanwhile, that virtual AI celebrity who only tells cute jokes but has no sustainable revenue source? Once the novelty wears off and token prices crash, it will vanish, unable to maintain operations.
The distinction is clear: purpose-driven agents will flourish, while the rest will gradually be weeded out.
This natural selection is beneficial for the field. Developers are forced to innovate, prioritizing effective use cases over flashy gimmicks. As these stronger, more effective agents emerge, they'll silence the skeptics.
8. Synthetic Data Will Surpass Human Data
People often say "data is the new oil." AI depends on data, but its voracious appetite for it has raised concerns about an impending data scarcity.
The conventional wisdom is that we should find ways to collect private real-world data from users, even paying them for it. But I'm increasingly convinced that, especially in highly regulated industries or where real-world data is scarce, the more practical path is synthetic data.
These are artificially generated datasets, designed to mimic the real-world data distribution, providing a scalable, ethical, and privacy-preserving alternative to human data.
The power of synthetic data lies in:
- Unlimited Scale: Need a million medical X-rays or 3D scans of factories? Synthetic generation can produce unlimited quantities, without waiting for real patients or factories.
- Privacy Protection: Using synthetic datasets poses no personal information risks.
- Customizability: You can tailor the data distribution to exact training needs, injecting rare or ethically challenging extreme cases that may be difficult to collect in reality.
Granted, in many cases, user-owned human data will still be important. But if synthetic data continues to improve in realism, it may surpass user data in terms of quantity, generation speed, and freedom from privacy constraints.
The next wave of decentralized AI may revolve around "small labs" creating highly customized synthetic datasets for specific use cases.
These small labs will cleverly sidestep policy and regulatory hurdles in the data generation process, just as Grass bypassed web scraping restrictions by leveraging hundreds of millions of distributed nodes.
I'll explore this further in a future article.
9. Decentralized Training Will Finally Come Into Its Own
In 2024, pioneers like Prime Intellect and Nous Research broke through the limits of decentralized training. We've now trained a 150 billion parameter model in low-bandwidth environments, proving large-scale training outside traditional centralized approaches is feasible.
While these models are still not as practically useful (lower performance) compared to existing base models, I believe this will change by 2025.
This week, EXO Labs has taken a further step with SPARTA, reducing the communication volume between GPUs by over 1000 times. SPARTA enables large-model training without specialized infrastructure even under low-bandwidth conditions.
What impressed me the most was their statement: "SPARTA can be effective on its own, but can also be combined with synchronization-based low-communication training algorithms (such as DiLoCo) to achieve even better performance."
This means these improvements can be stacked to further increase efficiency.
As technologies like model distillation make smaller models practical and efficient, the future of AI is not about scale, but about better performance and accessibility. Soon, we will have high-performance models that can run on edge devices and even smartphones.
10. 10 new crypto AI protocols with a market cap of $1 billion (not yet launched)
ai16z soars to $2 billion by 2024
Welcome to the true gold rush era. It's easy to think that the current leaders will continue to dominate, with many comparing Virtuals and ai16z to the early stages of smartphones (iOS and Android).
But this market is too large and underdeveloped to be dominated by just two companies. By the end of 2025, I predict at least 10 new Altcoin crypto AI protocols that have not yet launched will have a circulating (non-fully diluted) market cap of over $1 billion.
Decentralized AI is still in its infancy, and there is a large pool of talent gathering.
We have every reason to expect the emergence of new protocols, new token models, and new open-source frameworks. These new entrants can displace incumbents by combining incentives (like airdrops or staking), technical breakthroughs (like low-latency inference or chain interoperability), and user experience improvements (no-code). A shift in public perception could happen in an instant.
This is both the charm and the challenge of this field. The market size is a double-edged sword: the cake is big, but the barrier to entry is low for skilled teams. This sets the stage for a Cambrian explosion of projects, many of which will gradually disappear, but a few will become transformative forces.
The dominance of Bittensor, Virtuals, and ai16z will not last long. The next batch of Altcoin crypto AI protocols with a $1 billion market cap is coming. Savvy investors have plenty of opportunities, and that's what makes it so exciting.
Extra Highlight #1: AI agents are the new apps
When Apple launched the App Store in 2008, the tagline was "There's an app for that."
Soon, you'll be saying, "There's an agent for that."
You won't be tapping on icons to open apps anymore, but delegating tasks to specialized AI agents. These agents will have contextual awareness, cross-communicate with other agents and services, and even self-initiate tasks you've never explicitly requested, like monitoring your budget or rearranging your travel schedule if your flight changes.
In short, your smartphone's home screen may become a network of "digital colleagues," each with their own domain of work: health, finance, productivity, and social.
And because these are crypto-enabled agents, they can autonomously handle payments, identity verification, or data storage using decentralized infrastructure.
Extra Highlight #2: Robotics
While the majority of this piece has focused on the software side, I'm also incredibly excited about the physical manifestation of the AI revolution - robotics. This decade will see robotics have its ChatGPT moment.
This field still faces major hurdles, particularly in acquiring perception-based real-world datasets and scaling physical capabilities. Some teams are tackling this, leveraging crypto tokens to incentivize data collection and innovation. These efforts are worth watching (like FrodoBots?).
After over a decade in tech, I can't remember the last time I felt this innate excitement. This wave of innovation feels different: grander, bolder, and just getting started.
