Market & Threats

The domain's pain over the past ~12 months — and why it adds up to a new security category.

Pull the year's expensive failures together and they rhyme: the contracts did exactly what they were told, and what they were told was false, stale, or fed in by an attacker through the data/RPC layer. Health checks never catch it — the endpoint is "up," the answer is just wrong.

A year of "the system trusted bad data"

The tell that this is a real category and not a run of bad luck: Chainalysis bought Hexagate and now sells monitoring for exactly this. The market is putting money behind "is the on-chain data correct?" — which is the question Smart Router is built to answer.

Competitors

The full field — what's commoditized, and the one lane that's still open.

The plumbing is a commodity

Routing across providers, health-based failover, retries, caching — all of it is now free and open-source (eRPCOpen-source fault-tolerant RPC proxy. Its case studies claim 45–67% cost/request cuts from caching alone., dRPC's NodeCore) and matched by every managed provider (Alchemy, QuickNode, Chainstack, Infura, Ankr, Tatum, Blockdaemon, Pocket/Grove). They compete on uptime %, price, and chain count. So does RouteMeshA routing service marketing 20+ providers, 1000+ chains, "40–70% cheaper.", which markets being "40–70% cheaper." The certifications (SOC 2, ISO 27001) we once led on are now everywhere too. None of it is a moat.

The one lane nobody holds: proof against the chain

The interesting players all sound like they verify — and each verifies something else. The clearest two:

• RouteMesh "Sentinel"Replays a sample of requests to 3 random providers and takes the majority — providers checked against each other, not the chain. checks providers against each other (replay to 3 random nodes, take the majority). Their own words: "there is no absolute ground truth available to an individual RPC consumer." That's the model we reject — and poison the providers (Kelp) and "majority" agrees on the wrong answer.
• dRPCAlready ships signed responses a client can re-verify — so "we sign" isn't unique. already ships signed responses you can re-check. So "we sign / cross-check" is table stakes, not a moat.

The honest read: this is a head start, not a wall. The chain read itself is easy to copy. What's hard to copy — and where we should aim — is being the filed record a regulator accepts (see Opportunities).

Customers

What our buyers did over the past ~6 months — and the one pattern across all of them.

The strongest signal we're onto something: our customers are already racing toward the destination on their own — audit-ready proof that their on-chain data was correct.

• FireblocksPaid $130M for TRES (Jan 2026); added Canton (Feb 2026). — Jan 2026: paid $130M for TRES (reconciliation/accounting across 280+ chains); Feb 2026: added Canton support. They're stitching on-chain truth to audit-grade books.
• dfnsPowers IBM's Digital Asset Haven (Oct 2025); 2026 plan names "evidence and audit primitives." — Oct 2025: became the engine behind IBM's Digital Asset Haven; public 2026 plan names "evidence and audit primitives."
• Kraken — Mar 2026: got a Fed master account; May 2026: applied for a national bank charter. Now federally supervised, on top of Proof-of-Reserves — they must prove their on-chain data was right at decision time.
• GK8 (by Galaxy) — air-gapped custody, "no single-vendor risk." Still needs trustworthy reads and proof a transaction landed, without trusting one provider.

The pattern, and the pains we're probably missing

Every one is racing up-stack — from "is the data correct" toward "can I prove, account for, and report this to a regulator." So the receipt's real destination isn't "best-execution proof," it's the compliance record they file. Ranked by how much they'd grow or defend an account:

• Compliance-grade evidence (the filed MiCA/DORA record) — highest leverage, most defensible.
• Correctness-under-degradation (finality/reorg-aware reads) — why Hypernative bought us; what Kraken needs to credit deposits.
• Reconciliation signal — "your providers disagree on this balance" — a sticky layer above routing.

Regulation

The clock that turns the receipt from a nice feature into something firms must buy.

This is the surest, best-dated part of the picture — the reason the receipt-as-record is a deal this quarter, not someday.

• MiCA — hard deadline July 1, 2026. After it, a firm without authorization can't serve EU clients. New licensees are standing up exactly this record-keeping right now.
• DORA (live since Jan 2025) — report major incidents within hours; keep an auditable register of critical infrastructure providers. A signed, timestamped router log proving when something diverged is a native DORA artifact.
• ESMA's machine-readable record format (ISO 20022 / JSON, published late 2025) — regulators will request records in this format, so a receipt born in it doubles as the regulator-ready record.
• SOC 2 "processing integrity" (complete, valid, accurate) maps almost word-for-word onto multi-source reconciliation — a control a customer can inherit from us.

Opportunities & the call

Where to expand, how durable the edge really is, and the one test that settles the bet.

The bets, ranked

• 1. Sell the receipt to the customers we already have — first. The receipt is our proof: a record anyone can re-check against the chain itself that the data we served was the real, canonical answer — issued in the exact shape EU regulators (MiCA, DORA) now demand. Our customers owe auditors that evidence regardless. Hand them ours and we become load-bearing — ripping us out means re-papering their compliance. Fastest revenue, deepest lock-in.
• 2. Then go after the stablecoin and payments crowd — BVNK, Bridge, Conduit, the outfits moving real money across chains. Their whole job is making the numbers reconcile from one chain to the next, and the EU now obliges them to keep audit-grade records of exactly that. It's our product, aimed at fresh logos.
• 3. Tokenized real-world assets — enter, but with eyes open. RWA means wrapping real assets — treasuries, funds, property — as on-chain tokens. Real money, real demand. But the prize they chase, proving a token is genuinely backed, is already ChainlinkThe incumbent that already provides the "is this token really backed?" proof. We won't take it from them.'s turf — and we won't take it. So don't try. Win the layer beneath instead: the trustworthy chain data their books depend on. A slower, supporting sale — behind the first two.
• 4. Oracle networks — leave them alone. Oracles pipe outside data, like prices, onto the chain. Chainlink owns the category. Feed it or partner with it; don't pick that fight.

How durable is the edge, honestly?

A head start, not a wall. Routing and signed responses are table stakes anyone matches in a quarter or two. The one thing that becomes durable is becoming the record a customer files — once they do, swapping us out means re-papering their compliance. So the job of the next two quarters is to manufacture that switching cost before the rest commoditizes.

The call — one bet, one test

Ship the receipt as the compliance record our customers file. It does double duty — revenue from the customers we have, and the only thing that turns the edge durable. But the core claim is still unproven, so it gets a two-week gate first:

✦ The Maverick — sideways bets

The wildcard lens: bold, non-obvious plays — each with the cheapest test that would prove or kill it in weeks. Bets the Critic hasn't fought yet, not conclusions.

One pattern jumps out before you even weigh the ideas: three of the five sell to a buyer we don't talk to — risk, insurance, legal, compliance — not the infra team. That reframe may matter more than any single bet.

Source: the Maverick agent, Cycle 01 — before the Critic. Trend grounding: x402 (Coinbase / Linux Foundation), 2026 RWA + stablecoin reports.