Zero audits. Zero user data. Zero team identities. In a market where trust is the only currency, WhatPay launches with all three missing. The project claims to be an AI-native multi-chain wallet that lets users trade, query, and analyze on-chain data through natural language — a 'conversation-as-trading' paradigm. But as a data detective who has spent years tracing the scars left by every transaction, I find the wound before the hype. This article is not a review; it's a forensic breakdown of what the official launch announcement actually reveals — and what it hides.
Context: The Setup
WhatPay is positioned as an application-layer wallet infrastructure. Its core innovation: replacing traditional menu-driven interfaces with a large language model (LLM) that interprets user intent, retrieves off-chain and on-chain data, and executes transactions — all within a chat window. The technical backbone combines MPC (multi-party computation) self-custody with support for 65 blockchains and Layer-2s. The official narrative is that this reduces friction for non-technical users and unifies the fragmented DeFi experience.
But here’s the problem: the entire announcement is a closed box. No public code repository. No third-party security audit. No team names. No user metrics. No tokenomics. No governance structure. As I wrote in my 2017 ICO audit pipeline, I rejected 80% of projects due to missing technical specifications. The 2017 code was honest; the humans were not. Today, the pattern repeats — but with a shiny AI wrapper.
Core: The On-Chain Evidence Chain
Let’s dissect the technical claims against the data we can verify — or more precisely, cannot.
1. The AI Backend: A Centralized Black Box
WhatPay states that its AI automatically performs intent recognition, data retrieval, and result generation. But it does not disclose which LLM it uses, how it processes on-chain structured data (e.g., through indexed RPCs or GraphQL), or — critically — how it prevents hallucinations that could generate incorrect transaction parameters like wrong token addresses. Every transaction leaves a scar; I find the wound. Here, the wound is the lack of transparency around the AI reasoning pipeline. In my experience tracking DeFi liquidity during Summer 2020, I built custom SQL dashboards to verify every data point. WhatPay offers no such verification layer. The user must trust the AI’s output without inspection.
2. The 65-Chain Claim: Depth vs. Breadth
Supporting 65 chains is a headline grabber, but the critical question is: what does “support” mean? In practice, multi-chain wallets typically fall into two tiers: (a) read-only balance display, and (b) native interaction (swap, bridge, DApp access). The official announcement lists chain names but provides zero granularity. From my work on the 2022 Terra collapse forensics, I know that multi-chain support often masks thin liquidity. In May 2022, the algorithm ate its own tail — Terra’s ‘support’ on many wallets meant nothing when the peg broke. Without disclosing which chains allow native swapping or which DEX aggregators are integrated, the 65-chain claim is noise, not signal.
3. The MPC Self-Custody: Missing Thresholds
MPC is a mature technology used by Fireblocks and ZenGo. WhatPay says it fragments the private key across multiple parties, and the platform cannot access user funds. But it does not reveal the threshold scheme (e.g., 2-of-3, 3-of-5), who holds the shards, or whether offline backup recovery is supported. In my 2024 ETF inflow model, I correlated institutional wallet creation with price movements — but only after verifying the custody architecture. Without such details, the MPC claim is an empty promise.
4. No User Data, No Market Validation
The announcement contains zero metrics: no DAU, no TVL, no transaction volume. In a sideways market where chop is for positioning, the absence of data is a red flag. Projects that have real traction typically lead with numbers. WhatPay’s silence suggests it is in the pre-product stage, targeting early adopters and investors, not actual users. Structure reveals the chaos hidden in the noise — and here, the structure is entirely missing.
Contrarian: The Counter-Intuitive Angle
One might argue that the AI wallet narrative is so hot that being early gives WhatPay a first-mover advantage. But correlation ≠ causation. The hype around AI+Web3 does not automatically translate to product success. In fact, the very feature that sets WhatPay apart — conversational trading — introduces a new attack surface: if the AI backend is compromised, a malicious prompt could return a fake contract address, and the user, trusting the AI, might sign a transaction that drains their wallet. The platform claims “all transactions require user confirmation,” but confirmation is meaningless if the user cannot verify the AI’s output. This is a classic case of responsibility transfer: the project shifts security risk to the user while maintaining central control over the AI.
Furthermore, mainstream wallets like MetaMask and OKX are already integrating AI features. Their existing user base and brand trust give them a structural advantage. WhatPay’s differentiation is a thin layer of UX — easily replicable, not defensible. The 2017 code was honest; the humans were not. In 2025, the code (AI) is opaque, and the humans (team) are anonymous. That’s a double trust deficit.
Takeaway: The Next-Week Signal
Over the next 90 days, watch for three signals: (1) publication of a security audit by a reputable firm (SlowMist, Trail of Bits), (2) disclosure of team identities and backgrounds, and (3) release of any user growth data. If none of these materialize, WhatPay will join the 80% of projects I rejected in 2017 — ambitious but unverifiable. The market is not forgiving in a chop; liquidity is a mirror, and it shows who is fleeing. Until the evidence arrives, treat this as a research sample, not a portfolio allocation. The question is not whether AI wallets are the future — but whether this particular wallet has the scars to prove it.