Question about Tokenomics Phase 1 outcomes — what happens to staking APR under a fee-based model?

I’m a blockchain student researching Cosmos Hub, and I’ve been reading the ATOM Tokenomics Phase 1 outcomes thread. The Gauntlet finding that only ~42% of withdrawn rewards get sold the same week surprised me — it suggests stakers aren’t just farming and dumping.
My question: if the Hub eventually moves to a fee/revenue-based model as discussed since the November kickoff, what realistically happens to the ~15% staking APR in the transition period? Is the expectation that enterprise/Eureka fees can actually replace inflation as the security budget, or is a much lower APR just accepted as the price of sound tokenomics? Curious how long-time community members see this trade-off.

Less inflation, more staking revenue from fees. I like it and want it.

Spot-on question by `@guyangel18`. This touches on the core risk-reward dilemma that proponents of fee-driven models often overlook. If ATOM slashes its staking APR down to a 4%–6% range to mimic top-tier assets like ETH or SOL, it creates a fundamental capital allocation problem: Why should stakers stay in ATOM instead of migrating to blue-chips? When you evaluate the proposition, it becomes a clear case of “offering less for less”: • Asymmetric Risk/Reward: Why take on the volatility and lower market-cap risk of ATOM for a 4%–6% yield when you could hold BTC, ETH, or SOL for comparable returns/security and far greater institutional backing? • The 21-Day Lockup Friction: Top assets don’t force you into a strict 21-day unbonding period for a modest yield. Locking up capital for three full weeks during market downturns for a tiny 4% APR on a medium-cap token makes zero financial sense. • Competitive Pressure: If ATOM lowers its returns without first establishing ETH-level fee generation, it destroys its main incentive for capital retention without offering superior security or liquidity in return. Unless the Cosmos Hub introduces unique native value-capture…

Excerpt (1194 of 1349 characters). Read the whole post on the forum ↗