Goldman Sachs Q2: What Actually Powered the 23.5% ROE?

Goldman Sachs turned every $100 of year-over-year revenue growth into roughly $94 from Global Banking & Markets—and still posted a credible counterweight in financing and fee revenue. That tension, not the headline alone, is the Q2 story. The 23.5% return on equity matters only if the machine can keep producing after the loudest trading quarter fades.

A 23.5% return, with one big dependency

Goldman's Q2 supports a balanced conclusion. The quarter was not merely a trading spike: financing revenue, management fees and private banking revenue all contributed to a more-durable proxy that grew by one-third. But the firmwide growth bridge was overwhelmingly driven by Global Banking & Markets, where Equities alone produced 54.1% of the firm's entire year-over-year revenue increase.

That distinction matters because a high-return quarter and a repeatable earnings system are different claims. Client activity, underwriting and asset prices can lift Global Banking & Markets quickly. Compensation and infrastructure costs do not necessarily fall at the same speed when activity normalizes. Goldman's 57.4% efficiency ratio shows excellent Q2 conversion; it does not prove the same conversion under a weaker mix.

The constructive case strengthens if Q3 and Q4 each hold ROE near or above 20%, the more-durable proxy remains near or above 42% of firm revenue, efficiency stays at or below 60%, and tangible book value per share continues growing. Those are editorial monitoring tests, not Goldman guidance or an investment rating.

Follow every $100 of added revenue

The official financial overview establishes the quarter's scale. Net revenues were $20.338B, net earnings were $6.628B and diluted EPS was $20.98 per share. Annualized ROE reached 23.5%, annualized ROTE reached 25.5%, and the efficiency ratio was 57.4%.

Net revenues increased from $14.583B to $20.338B, a $5.755B gain. Global Banking & Markets rose from $10.133B to $15.520B, a $5.387B gain. Dividing those amounts gives 93.6%. Asset & Wealth Management added $766M, while Platform Solutions declined $398M. The result is a clean arithmetic bridge: almost all firmwide growth came from GBM.

GBM's official detail shows why the concentration was so large. Investment banking fees rose to $3.395B from $2.191B. FICC reached $4.592B from $3.487B, and Equities rose to $7.416B from $4.301B. Equities alone added $3.115B, or 54.1% of the firm's total revenue increase. Equities financing nearly doubled to $3.259B.

The official 2025 annual report describes FICC financing, Equities financing, management and other fees, and private banking and lending as “more durable” revenues. Applying that classification to Q2 creates a transparent proxy: $1.216B + $3.259B + $3.355B + $0.689B = $8.519B. The same lines totaled $6.361B a year earlier, so the proxy grew 33.9%.

There is an important limit. The proxy represented 41.9% of firm revenue, down from 43.6% a year earlier because transactional revenue grew even faster. We also exclude AWM Investments revenue, which rose to $441M from $137M; Goldman attributed the increase mainly to higher private-equity gains. That is real revenue, but it does not belong in this recurrence test.

This source hierarchy follows the five-gate bank earnings guide: preserve the reported denominator, rebuild the earnings mechanism, then separate durable evidence from cyclical evidence. It also makes Goldman's question different from the JPMorgan Q2 adjusted-ROTCE test, where significant gains had to be removed before evaluating the operating quarter.

The flywheel behind the headline

No live quote or private consensus is needed to define the expectation test. The reported quarter creates its own hurdle: after an unusually strong GBM contribution, a greater share of the next earnings step must come from financing, management fees and private banking while expenses remain controlled.

Source-derived expectation map: Goldman Sachs Q2 2026 earnings presentation, pages 2 · 4 · 5 · 11 and article calculations. The sequence is an editorial monitoring framework, not company guidance or a valuation model.

The official capital table supplies the final checkpoint. Standardized CET1 was 12.9%, the supplementary leverage ratio was 4.3%, and tangible book value was $336.61 per share. Goldman also repurchased $4.0B of stock during Q2. Capital return is useful only if it coexists with per-share tangible-book growth; the balance sheet cannot be treated as an unlimited fuel tank.

Where the return engine can slip

The first risk is revenue concentration. GBM represented 76.3% of firm revenue and 88.0% of common net earnings. A decline in underwriting, client activity or asset prices can therefore affect both the top line and compensation-driven operating leverage.

The second risk is confusing financing with fixed revenue. FICC and Equities financing are more durable than directional trading under Goldman's classification, but balances, spreads and client demand can still change. The label improves the analytical grouping; it does not turn market-sensitive revenue into a subscription.

The third risk is private-asset marks. AWM Investments revenue increased because of higher private-equity gains. Those gains supported Q2 earnings but were excluded from the durable proxy. A reversal would create a headwind even if management fees remain healthy.

The fourth risk is capital dilution by ambition. Repurchases, dividends, financing growth and regulatory buffers all compete for the same equity base. If capital requirements rise or tangible-book growth slows, a strong income statement can still produce weaker per-share compounding.

The two-quarter checkpoint

Goldman schedules its Q3 2026 results for 2026-10-13. One quarter can test the direction, but two are needed to judge whether Q2 established a stronger return regime.

The conclusion becomes more constructive if both Q3 and Q4 keep ROE near 20% or better while the more-durable revenue mix holds around 42%, efficiency remains below 60% and TBV per share rises. That combination would show that Q2's high return was supported by a broader earnings system.

It weakens if GBM activity cools and all three cushions fail together: financing and fees lose share, efficiency moves above 60%, and tangible book value stops growing. A single slower trading quarter would not flip the call. A weaker revenue mix plus weaker conversion would.

How the durable proxy was rebuilt

The firmwide growth bridge subtracts Q2 2025 net revenue from Q2 2026 and repeats that calculation for each segment. The more-durable proxy applies Goldman's annual-report classification to the exact Q2 financing, management-fee and private-banking lines; it excludes AWM Investments revenue. The expectation tests compare reported operating evidence across future quarters; none is a fair-value estimate.

• Goldman Sachs Q2 2026 earnings release, published 2026-07-14
• Goldman Sachs Q2 2026 results PDF, published 2026-07-14
• Goldman Sachs Q2 2026 earnings presentation, published 2026-07-14
• Goldman Sachs 2025 annual report, checked 2026-07-15
• Goldman Sachs 2026 earnings calendar, checked 2026-07-16

Facts, calculations, image captures and links were rechecked as of 2026-07-16. AI assisted with structure, chart production and EN/KO consistency checks; final editorial responsibility remains with Hynexly's owner-operated desk. No sponsorship or affiliate relationship with Goldman Sachs is disclosed. This is general information, not individualized investment advice; it does not issue an investment rating or share-price objective.