When a DEX protocol upgrades its pool incentive mechanics, it changes the equilibrium of where liquidity providers want to deploy capital. That shift in LP behavior shows up as a change in pool depth, which in turn affects routing efficiency for anyone trading on or through those pools. The Osmosis v23 upgrade, which modified how liquidity mining rewards are distributed across concentrated liquidity positions, produced exactly this kind of structural depth shift in several major pairs, and the ATOM-USDC pair is the clearest case study.
We started tracking ATOM-USDC depth on Osmosis DEX immediately after v23 went live and continued for 30 days. The pattern we saw was not what most participants expected going into the upgrade, and it has implications for anyone routing cross-chain trades that touch Cosmos.
What v23 changed in pool incentive mechanics
Prior to v23, Osmosis concentrated liquidity pools distributed incentive rewards proportionally across all active positions within the tick range, with equal weight per unit of liquidity provided. The v23 upgrade introduced a modified distribution formula that weights rewards toward positions that are closer to the current price, specifically those within a configurable spread of the current tick. The intent was to encourage tighter liquidity concentration and reduce the practice of LPs placing wide positions to collect rewards passively without meaningfully contributing to depth at executable prices.
In theory, this should increase depth at the 1% price impact threshold because more liquidity is being incentivized to concentrate near the current price. In practice, the transition period created an unusual depth pattern as existing LPs repositioned their capital to capture the new reward distribution. During that repositioning window, total active liquidity dropped temporarily while LPs exited wide positions before re-entering at tighter ranges.
The 30-day depth pattern
In the first 72 hours after v23 activated, depth at the 1% price impact level for ATOM-USDC on the primary Osmosis concentrated liquidity pool declined by roughly 30-40% relative to the 7-day pre-upgrade baseline. This is the repositioning dip: LPs withdrawing wide positions and the new tight positions not yet filling the gap. If you were routing a moderate-sized ATOM-USDC trade in that window, the actual execution quality was meaningfully worse than the pre-upgrade average.
By day 7 post-upgrade, depth had recovered to approximately pre-upgrade levels. By day 14, depth at 1% exceeded the pre-upgrade baseline in most measurement windows, consistent with the theoretical expectation that tighter concentration improves depth at standard execution notionals.
The interesting part is what happened between day 14 and day 30. Depth at the 1% threshold continued to improve on average, but intraday variance also increased. Tight positions require more active management: LPs who set positions near the current price need to rebalance as the price moves, or their positions go out of range. This creates a pattern where depth is very high when the price is within a well-managed concentration band, but drops more sharply when price moves push it near the edges of active positions. For execution purposes, this means depth is higher on average but more volatile than before, which changes how you should estimate execution cost for trades placed at non-peak-depth moments.
Routing path implications
Before v23, ATOM-USDC routing on Osmosis was relatively predictable: depth was moderate and stable, and the execution cost estimate from a 1-hour-old snapshot was usually within an acceptable margin of the realized cost. After v23 settled into its new equilibrium, the pool is often deeper than before at favorable moments but occasionally shallower at others.
For a cross-chain routing system, this changes the right data cadence. Pre-upgrade, hourly depth snapshots were a reasonable input for routing decisions on moderate trade sizes. Post-upgrade, hourly snapshots miss the intraday depth variance pattern. The routing system that performs best now is one that can read current pool depth at routing time rather than relying on a cached snapshot, and that has a view on how tight the current concentration band is relative to the current price.
A practical signal: if the ATOM price has moved more than 1.5-2% from the center of the largest concentrated position in the past hour, the available depth at 1% price impact is likely lower than the recent average suggests. This is not a hard rule but a useful heuristic until the pool composition data is available explicitly.
Comparative routing: direct vs. multi-hop
The v23 depth dynamics also changed the relative attractiveness of the ATOM-USDC direct path versus multi-hop alternatives. Before the upgrade, the direct ATOM-USDC pool was consistently the best-depth option for trades up to a certain notional. After v23, there are windows when the multi-hop route through OSMO or another intermediate pair has competitive depth because those pools have not yet fully migrated to concentrated liquidity mechanics with the same volatility profile.
This is not a permanent situation. As all major pairs on Osmosis DEX continue to mature under v23 incentives, the depth hierarchy will stabilize. But it illustrates a general principle: protocol upgrades that change LP economics are depth events, not just governance events. Research desks that follow protocol governance closely but do not update their depth data accordingly will be working with a model that no longer reflects current pool behavior.
What to watch going forward
The most useful forward-looking indicator for ATOM-USDC depth on Osmosis post-v23 is the distribution of active LP positions around the current price. When many positions are concentrated within 1-2% of the current price, depth is high and stable. When positions are spread wider following a price move that pushed some LPs out of range, depth is lower and recovery depends on when those LPs reposition.
For routing systems and research desks, the takeaway is straightforward: treat depth as a dynamic pool-state property that changes with protocol upgrades, not just with market conditions. A routing model calibrated on pre-v23 depth data will systematically misprice execution costs under v23 dynamics, particularly for larger trade sizes where the depth curve shape matters more than for small trades.
We are continuing to index ATOM-USDC depth post-v23 along with several other concentrated liquidity pools on Osmosis DEX. The 30-day dataset is the beginning, not the conclusion. Upgrade-induced depth dynamics typically take 60-90 days to reach a new stable equilibrium as LP behavior adapts to the changed incentive structure.