
In 2025, we made a mistake.
We didn’t just ship features. We didn’t just publish research. We didn’t just talk about AI agents.
We accidentally turned AI agents into an onchain business model.
What started as an experiment in crypto UX became something much bigger: a production-grade, revenue-generating, verifiable agent ecosystem — running fully onchain.
This is the story of how that happened.

For a long time, AI agents in crypto lived in demos, threads, and decks. In 2025, they started living onchain. By the end of the year, the Warden ecosystem had processed:
10,876,397 cumulative onchain transactions
11,736,663 inference proofs generated
$749,872.20 in total protocol revenue
$703,736.90 in cumulative agent revenue
These weren’t test transactions. They weren’t internal benchmarks.
They were real users, real capital, real execution — across EVM and Solana.
AI agents stopped being a narrative and started being infrastructure.
The real breakthrough wasn’t volume — it was who generated it.
In 2025, a growing set of production agents emerged, each solving a specific problem:
Swap Agents for execution and liquidity routing
Kaibot for automated strategies and evolution over time
Messari Agent for research-driven intelligence
WachAI for risk-aware trading
deBridge Agent for cross-chain execution
Portfolio and DCA agents for long-term allocation
Community-built agents, deployed permissionlessly
This wasn’t a single app. It was an ecosystem of autonomous actors — each composable, monetizable, and accountable.
Early in the year, we made a deliberate choice:
Instead of scaling a closed system, we opened the agent stack. Anyone can:
Build an AI agent
Deploy it onchain
Earn based on usage and value delivered
This shifted Warden from a product into a platform.
Builders weren’t just contributing code — they were launching businesses. Agents weren’t just tools — they were economic participants.
Crypto UX has been broken for years.
Too many dashboards. Too many tabs. Too much cognitive load.
In 2025, we leaned fully into a different idea: What if execution felt like conversation?
The AI Trading Terminal and Warden App reframed how users interact with crypto:
Intent → command
Command → execution
Execution → onchain proof
Instead of learning protocols, users described outcomes. Agents handled the complexity.
UX stopped being a bottleneck and became a multiplier.
Finance doesn’t have to be sterile.
With Betflix and the League of Traders, trading evolved into something new:
Swipe-based strategy discovery
Competitive tournaments
Social mechanics powered by agents
This wasn’t gamification for its own sake. It was a new distribution layer for agents — where strategies compete, evolve, and earn.
And we just launched our AI Trading Terminal. Did you try it? Signals and AI acting as your copilot for trading.
As AI agents started handling capital, one question became unavoidable:
How do you trust a black box?
The answer wasn’t trust, it was verification.
SPEX (Statistical Proof of Execution) became one of the most important pieces of infrastructure we shipped in 2025.
SPEX enables:
Verifiable AI inference
Resistance to manipulation
Cryptographic accountability for agent actions
Instead of asking users to trust AI outputs, SPEX lets them verify that:
an inference happened
it followed expected statistical properties
and it was executed honestly
SPEX isn’t a feature. It’s a firewall between AI and capital.
Partnerships in 2025 weren’t about announcements, they were about integration.
Across the year, Warden shipped alongside partners including:
Messari
Venice AI
Hyperlane
TEN
deBridge
Irys
BNB Chain
and others
Each partnership extended what agents could do: more data, more chains, more execution paths.
The strongest validation didn’t come from charts, it came from usage.
In 2025, Warden saw:
Millions of monthly active users
Hundreds of thousands of daily active users
Sustained engagement across multiple peaks
This wasn’t a one-off spike. It was repeated adoption — driven by agents that solved real problems.
By August 2025, something changed.
We stopped explaining what we were building and focused on why.
The Warden Manifesto articulated a belief that shaped everything afterward:
AI agents are actors, not assistants
Execution without accountability is dangerous
Verifiability is non-negotiable
That mindset now defines the protocol.
Throughout the year, everything was visible:
Open dashboards
Public roadmaps
Research articles
Monthly updates
Onchain metrics
There were no black boxes — operationally or philosophically.
The system spoke for itself.
We didn’t mean to do all of this.
But in 2025, we accidentally proved something important:
AI + crypto can actually ship. AI agents can earn. And onchain verification can scale.
This wasn’t the end.
It was the baseline.
And if 2025 was a mistake, 2026 is going to be worse.
So what are you waiting for?