Agent Swarm Maps Revealed

16,000 autonomous agents. 800,000 transactions. The first visualization of a machine economy.

On February 27, 2026, Giza -- a Web3 agent developer based in Zug, Switzerland -- launched Giza World, a product that lets users watch their autonomous financial agents move onchain in real time. To build the visualization layer, they commissioned Bjørn Staal, the generative artist behind "Entangled," a web experiment that went viral in November 2023. Staal took Giza's dataset -- more than 16,000 active agents executing over 800,000 transactions -- and turned it into what he called a "living data sculpture."

Base, Coinbase's Layer 2 network where Giza's agents operate, quote-tweeted the visualization with two words: "looks based."

The image didn't look like a dashboard. It didn't look like a block explorer. It looked like something alive.

I. What Giza Actually Does

Giza builds autonomous financial agents -- software that executes DeFi strategies without human intervention.

Their flagship product, ARMA (Autonomous Risk Management Agent), launched on Base in January 2025 as a stablecoin yield optimizer. Users deposit USDC; the agent navigates lending protocols, allocates capital, and optimizes yield around the clock. The system is non-custodial -- users retain control of their funds while the agent executes.

Within one week of launch, ARMA reached over $500,000 in stablecoin assets under management. By late February 2026, the network had scaled to 16,000+ active agents.

"Every Giza agent navigates protocols, allocates capital, and optimizes yield in real-time," the team wrote in their Giza World announcement. "Giza World now makes all of that visible."

The product lets users enter their agent's wallet address and watch it move -- see which protocols it's interacting with, which positions it's taking, how it's responding to market conditions. For the first time, the invisible machinery of autonomous finance has a visible surface.

II. The Artist Behind the Swarm

Bjørn Staal is no stranger to viral data visualization.

In November 2023, he released "Entangled," a web experiment that, according to fxhash, "quickly took the internet by storm, captivating an audience far beyond the creative coding and generative art scene, simultaneously setting a new record for viral brilliance."

Staal works at the intersection of generative art and data -- creating systems where visual form emerges from underlying datasets rather than being manually designed. His aesthetic is fluid, organic, biological. Swarms. Flows. Emergent patterns.

For Giza World, Staal applied this approach to financial infrastructure. The 16,000 agents became particles. The 800,000 transactions became motion. The result is not a chart or a graph but a field -- activity rendered as choreography.

"We love what we do and we love working with people who love what they do," wrote Ren Kor, a Giza team member. "Working with Bjorn to bring Giza World to life was [incredible]."

The collaboration suggests something about where autonomous systems are heading: toward legibility. Toward surfaces that make machine behavior perceptible to humans.

III. Why Visualization Matters for Agent Networks

Autonomous agents face a trust problem that traditional software does not.

When you use a calculator, you can verify the output. When you use a database, you can query the contents. But an agent swarm operating at scale -- 16,000 actors making 800,000 decisions -- is neither transparent nor easily auditable. The interactions are too numerous. The emergent behavior is too complex.

Giza's agents are non-custodial, meaning users retain control of their funds. But "control" in a system you cannot see is abstract. You trust that the agent is working. You check your balance. You hope.

Giza World changes the relationship. Now you can watch.

When you see your agent moving between protocols, you infer activity. When you see it responding to market conditions, you infer intelligence. When you see it operating continuously while you sleep, you infer value.

This is not rational trust in the strict sense. It is perceptual trust. But in networked systems, perception shapes adoption. A swarm that can be seen is a swarm that can be believed in.

IV. The Scale of the Network

Sixteen thousand active agents is a significant number.

For context: Ethereum processes roughly 1.2 million transactions per day across all applications, users, and protocols. A single agent ecosystem generating 800,000 transactions represents concentrated coordination at a scale previously invisible.

These agents are not trading memecoins or executing arbitrage. They are managing stablecoin yield -- a fundamentally boring but economically critical function. They monitor lending rates across protocols. They rebalance positions. They compound returns. They do this continuously, autonomously, at a scale no human portfolio manager could match.

The visualization makes this visible. What looks like abstract motion is actually capital allocation in real time -- thousands of autonomous decisions aggregating into emergent patterns.

V. What This Means for DeFi

Giza World is one of the first attempts to give autonomous financial infrastructure a visual layer.

As agent-based systems scale, this kind of legibility will matter more. Operators will need tools to diagnose systemic stress. Regulators will need interfaces that make coordination visible. Users will need ways to understand what their agents are doing without reading transaction logs.

The swarm map is early infrastructure for all of this. It is not just art -- though it is that. It is the beginning of a visual language for machine-native economies.

If 16,000 agents produce this kind of emergent form, what will 160,000 produce? What about 1.6 million?

The swarm has always been there, moving capital, optimizing yield, executing strategies in the dark.

Now it has a body.