GS · Q4 2025 Earnings

Narrative arc: Q2 (M&A reignites, capital becomes the story) → Q3 (dominance quantified, GS 3.0 launched) → Q4 (multi-year structural targets, excess capital framing, AI reorg formalized into six work streams).

The forward framework moved from quarterly cycle calls to multi-year structural targets. Q4 set a $750B fee-paying alts AUS target for 2030, a 5% long-term net inflow target for AWM, a 30% AWM pre-tax margin target, and reaffirmed mid-teens through-cycle ROE. The anchoring quote: "We believe we can raise between $75 and $100 billion annually on a sustainable basis...fee-paying alternative assets under supervision to reach $750 billion by 2030." Goldman doesn't typically issue 2030 targets — doing so signals management wants investors to underwrite the trajectory, not the next print.

Capital posture moved from "deployable" to "excess." Q4 explicitly used the word "excess": "given our strong earnings generation capability and excess capital positions, we also have capacity to return more capital to shareholders." The 50% dividend raise (to $4.50/quarter) and CET1 at 14.4% Standardized are the action and the headroom that match the language. Worth noting the discipline: when Poonawala asked whether the firm would reset the mid-teens ROE target higher, Solomon declined — capital flexibility is being used for dividends and growth investment, not for resetting the through-cycle bar.

GS 3.0 moved from announcement to operating-model specifics. Q3 launched "Goldman Sachs 3.0" as a "multi-year effort" with detail promised for January. Q4 delivered the next layer: "we announced the launch of One Goldman Sachs 3.0, our new operating model propelled by AI. We are excited to embark on this effort, starting with six work streams we identified as ripe for disruption." Six identified work streams is incremental specificity, but headcount targets, charge magnitude, and an efficiency-ratio glide path are still absent — the answer to the Q3 watch item is partial.

Wealth strategy explicitly clarified as manufacturer + distributor, not direct competitor. Schorr's Q&A produced the cleanest articulation yet: ultra-high net worth remains direct, while mass-affluent and HNW is being scaled via third-party RIA partnerships, with the United Capital experiment now formally retired as a strategic precedent. The new 5% long-term fee-based net inflow target is the public anchor, but note management also disclosed that over the last five years, long-term fee-based inflows grew at a 6% annual pace — meaning the new external target sits below the trailing run-rate. Q4 long-term fee-based inflows were $66B, comfortably above a 5% pace against the AWM base; investors should size the 5% bar against the fee-based denominator each quarter.

Risk language stayed thin and macro. The risk register cited — economic growth, policy, geopolitics, market volatility — is the same boilerplate as prior quarters. There is no incremental caution embedded anywhere in the prepared commentary; the bullishness is genuine, not hedged.