Where we are going

Q1 2026

Foundation

The balancer core, running in one region against testnet.

• JSON-RPC edge with API key auth and method allowlists
• Round-robin, random and weighted routing strategies
• Per-provider circuit breakers and jittered retries
• Robinhood Chain testnet endpoint with four pooled upstreams

Q2 2026

Reliability

Latency-aware routing and the observability layer.

• Latency-based strategy with rolling p50 re-ranking
• Active health probes and block-height drift detection
• Prometheus scrape endpoint and per-provider dashboard
• In-flight deduplication and deterministic-result caching

Q3 2026

Mainnet readiness

Multi-region capacity and the guarantees mainnet traffic needs.

• Three-region edge with regional provider affinity
• WebSocket subscriptions with replay after reconnect
• Consistency guards: reads pinned to a minimum block height
• Transaction submission path with exactly-once semantics
• Public status page and incident history

Q4 2026

Scale

Higher ceilings, tighter controls, more places to run.

• Bring-your-own private nodes into the pool
• Per-key routing policy overrides and provider pinning
• Archive-tier routing for historical eth_getLogs ranges
• Usage analytics with per-method cost attribution
• SOC 2 Type II evidence collection

H1 2027

Adaptive routing

Routing that anticipates degradation instead of reacting to it.

• Predictive breaker trips from latency slope rather than error counts
• Per-method provider scoring - some upstreams are better at logs
• Automatic capacity rebalancing against provider quota burn
• Cross-chain pools beyond Robinhood Chain