Q2 2026 - Pyth Express Relay: A Path Forward

Co-authored by KemarTiti (Pythian Council, Pyth Data Association) and @zenyas (Pythian Council, Douro Labs)

Summary

Q1 fees are live and working as designed. With two months of data in hand, the Pythian Council and Douro Labs have reviewed Express Relay’s performance and want to share our findings with the DAO. This post covers what’s working, where the challenges are, and a proposed path forward.

What’s Working: X Stocks

Before diving into the numbers, worth highlighting what’s actually performing well. X Stocks account for ~45% of current volume and ~60% of fee revenue. They’ve held up better than other asset classes (-63% vs -90%+ elsewhere) because xStocks have limited onchain liquidity outside Express Relay, giving searchers better pricing power.

This is a real use case — tokenized assets with thin liquidity benefit from the Express Relay model.

1 Fees: Implementation Successful

In February, OP-PIP-92 activated protocol fees across 35 tokens. The implementation went smoothly — fees are collecting as designed.

Month Revenue Context
Jan 2026 $2,250 Pre-fees (tips only)
Feb 2026 $1,360 Fees activated Feb 9
Mar 2026 $2,200 Full month with fees

Revenue has been steady but hasn’t grown with the fee implementation, largely due to volume shifts.

Volume Changes

Metric 2025 (Annual) Post-Fee (Annualized) Change
Total Volume ~$600M ~$59M -90%
WSOL $123M $2.5M -98%
USDC $238M $27M -89%
Memecoins $77M $4M -95%
X Stocks ~$30M ~$11M -63%

The volume decline isn’t fee-related — at 1-10 bps, protocol fees are negligible (a $10k trade costs $1-10). The primary factor: Jupiter launched their own RFQ system and, from a competitive standpoint, removed Express Relay from their aggregator. Jupiter had been the main distribution channel for spot token flow.

Economics

Running Express Relay costs between $5,000-$15,000 monthly (infrastructure, tech support). Current revenue is around $2,000/month. For every $1 Douro spends operating Express Relay, the DAO sees roughly $0.20 in revenue.

The DAO isn’t directly bearing this cost, but it raises the question of whether Douro’s resources could have more impact elsewhere.

Market Context

A few observations from looking at the broader landscape:

  1. Distribution matters significantly in OEV/MEV. Products that own their flow (Jupiter) or are integrated with major existing flow (Chainlink SVR via Aave) have an advantage that’s hard to replicate through protocol improvements alone.

  2. Solana aggregator dynamics changed. Jupiter’s RFQ launch shifted the competitive landscape. This wasn’t unexpected from their perspective, but it affected Express Relay’s distribution.

  3. Tokenized assets remain promising. X Stocks performance shows the model works for assets with limited onchain liquidity — there may be future opportunities here even if the broader product winds down.

Proposal

Based on this review, we’re proposing an orderly wind-down of Express Relay.

What this means:

  • Douro Labs will begin a gradual sunset of Express Relay infrastructure

  • Communicate timeline to integrated protocols

  • Help current integrators transition

  • Maintain service during the transition period

  • Explore opportunities to license or sell the codebase to interested parties (AMMs, RFQ protocols, tokenized asset issuers) — any proceeds to the DAO treasury

Why: Current revenue (~$2K/mo) doesn’t justify operating costs ($5-15K/mo). Douro’s resources would have more impact on higher-traction areas like Pyth Pro and Pyth Terminal.

Timeline

Phase Timing Action
Discussion April 2026 Gather DAO feedback
Decision Last week of April OP-PIP vote to confirm direction
Deprecation End of Q2 2026 Full sunset, support integrator transitions
Code Exploration Ongoing Explore licensing/sale opportunities

Next Steps

  1. Community feedback — Does this direction make sense? Anything we should consider?

  2. Integrator feedback — Reaching out to protocols using Express Relay to understand their timeline needs.

  3. Searcher feedback — Talking to active searchers about transition and potential interest in the codebase.

5 Likes

From a community perspective:
Makes complete sense. The numbers don’t add up to continue this, especially given the situation with Jupiter. Pyth was there flying the flag for MEV protection from the beginning in July 2024. This helped set up the standard for swap users from their aggregators.

Pyth led the way on end-user priority, and does not need to become a martyr now that MEV protection is mainstream.

6 Likes

Agree with the above from @Derrp .

This move is inline with the sunsetting PythCore, allowing the team to further consolidate resources into more valuable ventures.

Seems like a logical choice to me.

3 Likes

Agree with @lowkeigh and @Derrp above. The business model should work with revenue not matter what - without it the expenditures will always be negative. Though X-stocks is a very very promising tool, considering the overall liquidity in the market for it and how people are trading it - it needs some time for adaptation!
Sunsetting means the tools are always in the box there and ready to deploy any time until better times and conditions will present it!

4 Likes

I think it makes total sense to sunset PER. The revenue generated from PER is not enough to cover its costs.

We should place more emphasis on Pyth Pro, which is currently the main revenue stream.

From SCP

3 Likes

Community: Agree with all the above - focusing efforts/resources into products with higher upside and winding down those with excessive operating costs compared to revenue generation makes sense

1 Like