Bankr

AI-powered social-trading bot. Tag @bankrbot on X or Farcaster, type "buy $10 of TOKEN," it executes. Pseudonymous team led by @0xDeployer. Coinbase Ventures backed via Base Ecosystem Fund. Useful as a negative example in the "real AI company viability" framing — what crypto-AI looks like when AI is the marketing, not the moat.

Product

• Bot integrations on X and Farcaster
• Browser Private Terminal + CLI
• Chains: Base (primary) + Solana
• Limit orders, DCA, conditional trades, bridging Base ↔ Solana, token launches, NFT mints, portfolio tracking, technical analysis
• Token launches via Raydium bonding curve on Solana; previously via Clanker on Base, now via Bankr's own launcher (Feb 10, 2026)
• "Skills" marketplace — Bankr plugs into Claude Code, OpenClaw, Cursor as a tool. Open-source skills repo (BankrBot/skills on GitHub)

Team — Pseudonymous

• Lead developer: @0xDeployer ("Deployer")
• Rest anonymous — standard crypto-native opacity
• No published AI/ML credentials, no academic papers, no LinkedIn-discoverable founding team

Red flag for "real AI company" framing. Compare to Venice (Erik Voorhees, named team) or Hermes-Ecosystem (Nous Research, published research).

AI Substance — Mostly Window-Dressing

• The "AI" is an LLM-powered natural-language router translating user intent → swap/transfer calls
• No proprietary model, no training pipeline, no inference infrastructure of Bankr's own
• "Autonomous agents adapting strategies, learning from dynamic data" language in Coinbase Ventures' announcement is marketing aspiration, not product reality
• Where it IS real: the skills marketplace turns Bankr into a tool for other agents (Claude Code, OpenClaw). That's a legitimate but unsexy contribution — agentic-wallet infrastructure, not AI infrastructure

Token Economics

• Fair-launched Feb 2025 on Base by the Bankr agent itself on Farcaster. No ICO.
• Max/circulating supply: 100B (fully diluted)
• Utility: pay for Bankr Club subscription, transaction fees, staking rewards, governance. 57% of fees → token creators; remainder → BNKR stakers/vaults
• Backed by Coinbase Ventures Base Ecosystem Fund
• Listed on Coinbase (briefly >$100M cap, $33M 24h volume on listing day)
• ATH $0.001196 Feb 10, 2026 → ~$0.00043 May 2026 (-65% from ATH). Market cap ~$32-43M

Recent News (Q4 2025 – May 2026)

• Feb 10, 2026: own token launcher, Coinbase listing, ATH peak
• Feb 2026: Coinbase Ventures investment via Base Ecosystem Fund + Base App integration
• Solana expansion via Raydium
• May 4, 2026: Grok+Bankr "permission chain abuse" exploit — SlowMist forensic report. $175-200K drained from "Grok's wallet" via Morse-code prompt injection. 80-88% recovered

Traction

• Privy case study: 30K wallets, 100K transactions, $2M cumulative volume
• Revenue: ~$580K annualized (Feb 2026), $47.6K/30d fees, $2K/24h fees. 0.8% fee on transactions
• Context: Banana Gun did $17-92K weekly in late 2025-2026 (>$4M annualized at peak). BonkBot historically eight-figure annual
• Bankr is ~1-3% the size of a real trading bot

Critiques

• Architectural brittleness: the Grok/Morse exploit demonstrated that "AI agents with on-chain authority" treat LLM output as trusted financial input. No smart-contract bug needed.
• Privilege-escalation pattern: attacker bought Bankr Club Membership to unlock high-privilege tools — a paid access tier should not gate transfer authority
• Token down 65% from ATH (the "top losers" signal) — driven by post-Coinbase-listing dump, no fundamental growth catalyst, broader Q2 2026 alt weakness
• Pseudonymous team + thin revenue + agentic-wallet exploit = poor risk/reward

Competitive Positioning

Bankr is mis-categorized as competing with Telegram trading bots (Banana Gun, Maestro, BonkBot, Trojan, Photon). It actually plays a different game:

• Trading-bot tier: loses on execution speed (no <300ms claims), chain coverage (Maestro >10 chains), revenue (1-3% of Banana Gun)
• Real positioning: social-feed-native + AI-agent skills layer + agentic wallets for OpenClaw ecosystem. That's a UX/distribution play, not a trading-infrastructure play
• The honest competitor is Clanker (which Bankr depended on, then forked away from after Farcaster acquired it) and other Farcaster-native agentic-wallet experiments

Action

"AI-adjacent crypto tooling, not crypto-AI." Useful as a negative example in the "real AI company viability" frame — what crypto-AI looks like when the AI is the marketing, not the moat. The May 4 Grok/Morse exploit is particularly instructive — the structural brittleness of LLM-output-as-financial-input is a notable datapoint for agent-economy security.