For three years, the crypto world has categorized Worldcoin as a glorified airdrop machine—scan your iris, get your tokens, and watch the price dance. The narrative was simple: distribute WLD to millions of verified humans, build a base, and figure out the business later. But with the announcement of Phase 3, the narrative has fractured. The project is no longer paying you to be human; it's asking businesses to pay for the privilege of knowing you are. This is not a pivot. It is a fundamental restructuring of the economic relationship between user, network, and token.
Worldcoin launched in 2021 with a promise that felt both utopian and dystopian: a global identity network based on iris scans, secured by zero-knowledge proofs, and incentivized by a native token. The deployment of Orbs—hardware devices that capture iris patterns—accelerated through 2022 and 2023, amassing millions of registrations across Asia, Latin America, and Africa. The token, WLD, became a top-100 asset by market cap, buoyed by the cachet of co-founder Sam Altman and the AI boom. Yet beneath the surface, a structural tension simmered: the entire user growth engine was fueled by inflation. Every new user meant new token emissions, diluting holders and funding operational costs. It was, in the language of venture capital, a classic “burn for growth” model—sustainable only if you eventually unearth a revenue stream.
Phase 3 is the attempt to unearth that stream. The official line: “We are moving from incentivizing registration to selling proof-of-human verification services to enterprises, applications, and AI agents.” The shift is from a consumer reward model to a B2B SaaS model. The product is no longer a token; it is an API that answers one question: Is this agent a human? For AI platforms drowning in bot traffic and sybil attacks, that question is increasingly urgent. But urgency does not automatically translate into willingness to pay—especially when the buyer is a decentralized AI agent with no credit card.
The key metric has shifted from number of verified humans to number of verification API calls per day. That is the core insight. World’s new valuation thesis rests on its ability to become the default trust layer for the AI economy. If every ChatGPT interaction, every decentralized science (DeSci) proposal, and every social network login requires a quick check against World’s registry, the network effects could be exponential. But if demand is tepid—if enterprises prefer cheaper, centralized KYC or homegrown solutions—then the token loses its narrative anchor and becomes a governance relic with no functional role.
This transition reminds me of a pattern I dissected during DeFi Summer in 2020. Compound’s COMP token was distributed to liquidity providers, creating a frenzy of “yield farming.” I calculated then that 40% of the liquidity was speculative arbitrage, not long-term commitment. When the farming rewards were reduced, the TVL collapsed. I called it “The Hollow Yield Trap” in my newsletter. Worldcoin faces a similar danger: its user base is incentivized, not intrinsically motivated. Phase 3 ends the incentives, but does not automatically replace them with organic demand. The human registry is large, but its active commercial value is untested.
Let’s examine the mechanism. World operates an L2 on Optimism’s OP Stack. When an Orb scans an iris, it generates a cryptographic commitment stored on-chain. The user’s biometric data never leaves the device in raw form—zero-knowledge proofs verify the uniqueness without exposing the template. This is technically elegant. The real infrastructure challenge is the API layer: a scalable, low-latency oracle that can answer thousands of queries per second from AI agents. During the 2017 ICO mania, I modeled the economic incentives of early Chainlink nodes, and I see a parallel here. Both projects are building oracles—one for price data, one for identity data. But Chainlink had a clear paying customer base (DeFi protocols). World’s customer base for identity queries is still hypothetical. The network effect is real only if the buyer side materializes.
Mechanisms, not memes. World’s tokenomics have an additional complexity. WLD has a capped supply, but a significant portion is held by the foundation and distributed over time. Phase 2 saw high inflation as new users claimed rewards. With Phase 3, those claims will be “gradually wound down,” implying lower future supply growth. If the service revenue is collected in WLD—or used to buy back WLD—the token could become deflationary. But if revenue is in fiat or stablecoins, the token’s value capture is zero. The official announcement is silent on the payment mechanism. That silence is a red flag. A token without a revenue flywheel is a governance shell. In my experience auditing narrative decay—from the FTX collapse to the Luna crash—the projects that survive are those where the token is an integral part of the value chain, not an afterthought.
Regulatory headwinds compound the uncertainty. World’s data collection practices have drawn investigations from the UK Information Commissioner’s Office, Germany’s data protection authority, and multiple Latin American regulators. Phase 3 increases regulatory surface area: selling identity verification services falls squarely into KYC/AML territory. In the United States, the SEC could argue that WLD is a security under the Howey test—especially if the foundation’s efforts are seen as driving the token’s value through service sales. The transition from airdrop to business model may actually heighten securities risk because now there is an enterprise revenue story to evaluate. I discussed this with a compliance lawyer in Toronto last month; the consensus was that any revenue model that relies on token appreciation to compensate participants looks like a common enterprise.
But let’s play the contrarian. The market may be overestimating the demand for “proof of human.” AI agents are advancing rapidly, but most commercial applications today use centralized identity via OAuth or email verification. Decentralized identity standards like ENS and EIP-712 are lighter and cheaper. Moreover, World’s Orb is expensive to manufacture and deploy; the hardware moat is also a cost burden. If a software-only solution emerges—say, a zero-knowledge proof of a SSN or passport—World’s value proposition weakens. The contrarian thesis: World is building a luxury infrastructure in a world that is already satisfied with utility solutions.
During the 2022 bear market, I wrote a 10-part series titled “The Death of Faith-Based Finance,” dissecting how marketing outperformed audits in the FTX narrative. The same dynamic could apply here: World’s brand association with Sam Altman and AI has generated a trust premium that may not survive a failed commercial launch. If Phase 3 yields no major partnerships in the next 6 months, the crypto market will move on, and WLD will trade as a meme with diminishing attention.
Yet there is a bullish counter-contrarian scenario. Artificial intelligence is entering a phase where agentic systems—autonomous programs that act on behalf of users—will require a global identity standard. If World becomes the default human-proof layer for protocols like Bittensor, Virtuals, or even a future decentralized social network, its API could become a critical piece of internet infrastructure. In that case, the network effects are explosive: more agents mean more queries, which justify more Orbs, which bring more users, which attract more agents. Trust is a feedback loop.
In the end, every token faces its own contradiction. World’s is that it must stop bribing users to prove they are human and start convincing businesses to pay for that proof. Phase 3 is the hinge. If it works, World is no longer a crypto project—it is the identity layer for the AI century. If it fails, it becomes a cautionary tale about confusing distribution with demand. The market will start pricing that binary in the coming quarters. The question is whether we are witnessing the birth of a new trust protocol or the end of the free lunch.
—