Pyth DAO commits all product revenue to PYTH buybacks after monthly purchases fell 75%

A rule that spends one-third of a balance each month, with nothing coming in, leaves about 13% of that balance after five months ^[26]. OP-PIP-87 was built that way ^[3]. Counted in tokens, August's purchase was about 24% of March's ^[27]. That is a slower fall than pure drawdown, though token counts also move with the PYTH price, and neither report gives the price. The September 9 Ideas Bank post behind the change named two causes: the non-PYTH treasury was shrinking, and Douro Labs was paying the DAO more in PYTH and less in USDC, so there was less stablecoin to convert ^[14]. The post's author said the old method was "increasingly disconnected" from the network's revenue growth ^[18].

OP-PIP-136 fixes the first cause ^[1]. The budget now comes from a flow: the DAO's share of revenue from Pyth Pro subscriptions, Listing as a Service, the Data Marketplace and Pyth Indices ^[11]. The second cause is up to Douro Labs, which gained the option to pay in PYTH after an April forum proposal ^[15]. The DAO's cut is 60% of Pyth Pro revenue, 90% of Listing as a Service and 60% of the Marketplace ^[4]. Any part of that cut Douro pays in tokens is already PYTH when it arrives, so it is never bought on the market ^[14].

The size of the bid can only be estimated. Crypto Briefing puts the DAO's share at roughly 60%, with Douro keeping 40% ^[5]. Applied to $11.5 million of annual recurring revenue ^[6], that comes to about $6.9 million a year, or $575,000 a month ^[25]. The 90% cut on listings pushes the real figure somewhat higher ^[4]. Each buyback transaction is capped at $25,000, with up to 5% slippage allowed ^[16]. So $575,000 a month means at least 23 trades ^[28]. According to Allium Labs data, crypto projects spent a record $638 million on buybacks in 2026 ^[21]. Pyth's run rate is about 1.1% of that ^[29].

The first weeks will make the rule look better than it is. The legacy treasury is being spent through the same process ^[17]. The reserve went from about 41 million to 42 million tokens after the first purchases on September 30 ^[8], and that increase can include old cash as well as new revenue.

Douro Labs chief executive Mike Cahill said committing all of this revenue to the reserve is "the strongest alignment this network has had" ^[19]. Allium's head of research, Elton Shehdula, warned that buybacks reduce supply and add demand but do not guarantee price gains, Cryptopolitan reported ^[20]. Crypto Briefing wrote that the effect on supply will depend on how revenue compares with PYTH's trading volume and with unlocks entering circulation ^[24].

There are three ways this can go. Revenue keeps growing and arrives as USDC, so buying grows with it. Douro keeps paying in tokens, so the reserve fills while market buying stays small. Or growth slows, and buying slows with it ^[12]. I think the rule removes the old formula's main flaw, which was spending a fixed fraction of a shrinking pot. Demand for the data points to the first outcome. Pyth says its data underpinned more than 94% of tracked real-world-asset perpetual futures volume over the past three months ^[9], and Kalshi named it the only price source for its CFTC-cleared gold and silver perpetuals ^[22]. The case against is how Douro pays. If monthly dollar purchases come in well below one-twelfth of the DAO's annual revenue share while revenue is still rising, the 100% figure overstates the demand the rule creates.