# Pyth Token Phase 3

**URL:** <https://forum.pyth.network/t/pyth-token-phase-3/2350>\
**Category:** Ideas Bank\
**Created:** [February 13, 2026, 4:54pm UTC](https://forum.pyth.network/t/pyth-token-phase-3/2350 "2026-02-13T16:54:36Z")\
**Posts on this page:** 1\
**Showing post:** 18

<div class="post-metadata">

**Author:** ![Derrp](https://sea1.discourse-cdn.com/flex001/user_avatar/forum.pyth.network/derrp/32/670_2.png) [@Derrp](https://forum.pyth.network/u/Derrp)\
**Post date:** [March 10, 2026, 1:30pm UTC](https://forum.pyth.network/t/pyth-token-phase-3/2350/18 "2026-03-10T13:30:05Z")

</div>

In general, I thought the intent of OIS was a great use of the token to actually stake against quality data.  
HOWEVER: It depended largely on the publisher staking $PYTH against their own data to be effective as a “punishment” metric.

Whilst some publishers confidently backed up their data with tokens, the overwhelming majority of publishers staked zero tokens.

Furthermore, not one data misprint (for a slashing proposal) was submitted to the Pythian Council, as far as I can tell. So there was only carrot, and no stick?

In practice, OIS has been largely an inflationary exercise in token distribution, and not necessarily a successful experiment in data quality adherence.

If OIS has not been used for its initial intent (to reward good data + punish bad data), then we need to reassess if inflation is really necessary.

I will always welcome PYTH token utility, provided that it isn’t inflation wrapped up in a word-veil. Unfortunately I don’t think OIS really succeeded and would welcome a sunset.

For the record, however, I am glad that we ran the experiment to test the theory.

---

_[View the full topic](https://forum.pyth.network/t/pyth-token-phase-3/2350)._