A protocol can spend millions buying its own token and still leave holders worse off. Because buying tokens is only half the equation. The other half is how much supply enters circulation, whether purchases are burned, and what investors paid for the token. A tracker using DefiLlama data estimates $1.51B in buybacks across 23 programs over the year through October 15, 2026. The latest 30 days accounted for $155M. Three protocols made up roughly 72%: > @HyperliquidX: $71.4M (46.1%) > @Pumpfun: $30.4M (19.6%) > @Uniswap: $9.76M (6.3%) > Other programs: $43.44M (28%) But who actually benefits? … α/ Spending Isn’t Value Creation A $100M buyback against a $1B market cap represents 10% of valuation. A $150M buyback against $10B is only 1.5%. The bigger spender isn’t necessarily offering better economics. And token treatment matters. Purchased tokens might be burned permanently, held in a treasury, or redistributed later. Those aren’t equivalent outcomes. … β/ Three Programs, Three Different Risks Hyperliquid: Trading fees finance $HYPE purchases through its Assistance Fund. A September SEC-filed document described 99% of protocol fees going to the fund and 46.7M $HYPE acquired and permanently removed by August 23. But future emissions and contributor allocations complicate the supply picture. An unlock isn’t automatically a sale, yet token destruction doesn’t guarantee shrinking circulating supply either. Pumpfun: Its dashboard targets 50% of revenue for $PUMP buybacks and burns. The tracker estimates a 17.8% annualized buyback yield. That’s not a 17.8% investor return. If speculative trading cools, fee revenue and buybacks can cool with it. Its dashboard also flags reporting limitations after custom-pairs changes. Uniswap: $UNI’s UNIfication framework connects eligible fees to token burns through contracts. The proposal also included a 100M $UNI treasury burn and a proposed 20M $UNI annual growth budget. Burning treasury tokens isn’t necessarily the same as buying circulating tokens. And trading volume isn’t automatically revenue captured by $UNI. … γ/ What Net Absorption Reveals Imagine three hypothetical tokens: > A: $100M bought; $150M of newly liquid supply. Net: -$50M. > B: $80M bought; $20M newly liquid. Net: +$60M. > C: $50M bought; no new liquid supply. Net: +$50M. Relative to market caps of $1B, $2B and $500M respectively, those simplified net absorption rates are -5%, +3% and +10%. C has the smallest buyback, but the strongest relative absorption. These are illustrations, not actual protocol measurements. They assume purchases permanently remove liquid tokens and all new supply becomes liquid. Real unlocks aren’t automatic sales, and treasury-held tokens can return to circulation. … δ/ Buyback Yield Isn’t Earnings Yield A governance token doesn’t automatically confer shareholder rights or a claim on protocol profits. And fee-funded purchases can be cyclical. In bull markets, trading rises, revenue grows and buybacks accelerate. In downturns, buying power can contract just as holders want to sell. Annualizing a strong month doesn’t make its revenue durable. … ε/ My Take Buybacks can connect protocol success to token value. But dollars spent are a poor standalone leaderboard. The better comparison combines sustainable fee revenue, permanent burns, actual circulating-supply changes and valuation. The winner isn’t necessarily the protocol buying the most tokens. It’s the one creating the most durable value per dollar investors pay.