Coinbase Ventures' 2026 Bet: RWA Perps, Specialized Trading, and AI-Native Dev Tools
Coinbase Ventures outlined three areas it wants to fund in 2026: real-world asset (RWA) perpetuals, specialized trading infrastructure, and AI-powered tools for onchain development. If you build in crypto and care about product-market fit, this is a clear signal: ship where new liquidity, developer leverage, and defensibility converge.
RWA Perpetuals and Exotic Markets
Perpetual futures tied to offchain assets offer a faster path than full tokenization. You get synthetic exposure to assets without custody or transfer of the underlying. That opens markets around private companies, macro instruments (oil, credit spreads), and even time-series data like CPI prints.
Perp DEX volume is hitting about $1.4T monthly, up ~300% year over year. The rails are here; the surface area just expanded.
- What to build: Exotic perps (private equity baskets, event-driven spreads), oracle frameworks for non-price data, risk engines for thin markets, and settlement logic that survives volatility.
- Tech notes: Design funding-rate mechanics per venue, use latency-aware oracles, implement circuit breakers and position caps, and plan liquidation paths that work under stress.
Trading Infrastructure: Aggregation, Prediction, and LP Defense
Prediction market liquidity now tops ~$600M across fragmented venues. Expect demand for pro-grade terminals that consolidate quotes, size, fees, and settlement risk into one screen-with routing that actually improves expected value, not just fills orders.
On Solana, there's interest in AMMs that blunt predatory order flow. Think batch auctions, private or delayed mempools, JIT liquidity filters, inventory-aware curves, and intents-based matching to protect LPs.
- What to build: Unified prediction market order books, best-execution routers, cross-venue settlement hedging, and LP-first AMMs that price adverse selection.
- Tech notes: For Solana, optimize for QUIC, local fee markets, cranker networks, and GPU-friendly math. Add MEV-aware routing and quote decay to reduce toxic flow.
Next-Gen DeFi: Unsecured Credit and Transaction Privacy
Unsecured lending is on the table. The U.S. revolving credit market is ~$1.3T; onchain equivalents could blend reputation, cashflow data, and risk limits to lend without heavy collateral.
- What to build: Onchain reputation graphs, verifiable income/cashflow attestations, probabilistic risk scoring, dynamic credit lines, and programmatic collections.
- Data strategy: Combine offchain data (banking, payroll, commerce) with onchain behavior and maintain audit trails for regulators.
Privacy is a requirement for institutions and pro traders who can't leak strategies. Viable stacks include zero-knowledge proofs, fully homomorphic encryption, and trusted execution environments. Pick the model based on latency, cost, and disclosure rules.
- What to build: Selective-disclosure trading, private vaults, private intents, and auditor modes with bounded reveals.
- Primer: A solid starting point on ZK concepts is here: Zero-knowledge proofs (ethereum.org).
AI and Robotics: Dev Acceleration and DePIN for Embodied AI
AI coding assistants for smart contracts are nearing a "Copilot moment." The useful ones won't just autocomplete; they'll reason about specs, run property tests, point out reentrancy or overflow risks, and propose formal-verification hints. That levels the field for non-solidity engineers and speeds senior teams.
- What to build: Spec-to-code generators, audit copilots, fuzz + invariant harnesses with suggested fixes, and change-impact analyzers across contracts, scripts, and infra.
- Skill up: If you're adding AI to your dev workflow, this track helps: AI certification for coding.
For robotics, Coinbase Ventures is eyeing incentivized data collection. Decentralized physical infrastructure networks can reward high-quality sensor data and physical interaction logs to train embodied AI. Quality oracles, reputation, and anti-spam are the hard parts.
Expect "proof of humanity" to matter more as synthetic media floods feeds. Approaches blend biometrics with cryptography-Worldcoin is one example-raising the bar for Sybil resistance while keeping privacy in play.
Build Checklist: From Thesis to Traction
- Pick a narrow wedge: one market (e.g., credit spreads perps, private-company baskets) and one user (prop desk, market maker, or quant fund).
- Model risk first: funding rate math, oracle failures, liquidation paths, and venue downtime scenarios.
- Ship a testnet with guardrails: small caps, circuit breakers, kill switches, and real-time telemetry.
- Make execution your moat: best-ex routing, intent resolution, and LP protection that actually improves fill quality.
- Wire privacy where it matters: selective disclosure for regulators and counterparties; keep PII out of public state.
- Integrate AI where it saves hours: spec linting, audit suggestions, and diff-aware testing in CI.
- Prove demand with one metric: spread captured, PnL uplift for LPs, or default rate vs. APR-then scale.
- Partner early: oracle providers, custodians, credit data sources, and market makers willing to post day-one liquidity.
If you're building in these lanes, the bar is clear: ship faster than tokenization, route smarter than incumbents, protect LPs, and use AI to cut dev cycles without cutting safety. The teams that turn these theses into reliable systems will own the flow.
Your membership also unlocks:
