GS — Q2 2026 Financial Report Analysis

Overview

Goldman Sachs earned $6.63 billion in the second quarter of 2026, up 78% from $3.72 billion a year earlier, on net revenues of $20.34 billion (+39%). Diluted earnings per share of $20.98 nearly doubled from $10.91. Annualized return on equity — ROE, the profit a bank generates each year for every dollar of shareholder money it holds — was 23.5%, against 12.8% in the second quarter of 2025.

Almost all of the increase came from one segment. Global Banking & Markets, the institutional trading and dealmaking arm, produced $15.52 billion of net revenues, 53% more than a year ago, and $7.50 billion of pre-tax profit, up 84%. Within it, Equities revenue rose 72% and investment banking fees rose 55%. The two other segments moved in opposite directions: Asset & Wealth Management grew 20%, while Platform Solutions — the consumer-lending remnant GS is winding down — fell 64% on writedowns of the Apple Card loan book.

The prior-year comparison is unusually soft. In the second quarter of 2025 Goldman still carried the Apple Card and GM card portfolios on its balance sheet at amortized cost, which produced a $384 million charge for expected credit losses; this year the same line was $102 million. That $282 million swing is worth roughly 6 percentage points of the 78% profit growth on its own, before any of the trading improvement.

Headline figures

"Net revenues" for a bank means revenue after subtracting interest paid out — Goldman took in $22.05 billion of interest income in the quarter and paid $18.09 billion of interest expense, leaving $3.95 billion of net interest income, which is then added to fee and trading revenue. "Pre-tax margin" is profit before tax as a share of net revenues; it is the closest equivalent to the operating margin of an industrial company. The "efficiency ratio" is simply total costs divided by net revenues, so a lower number is better.

What actually drove the quarter

Equities was the single largest contributor. Equities net revenues were $7.42 billion, up 72%, split between $4.16 billion of intermediation (making markets in stocks and equity derivatives for clients, +60%) and $3.26 billion of financing (+91%), which the filing attributes primarily to "significantly higher net revenues in prime financing" — lending cash and securities to hedge funds against collateral. Prime financing revenue scales with how much money clients have deployed and borrowed, and the market backdrop was strongly supportive: the 10-Q records that the S&P 500 rose 15% and the MSCI World Index 14% during the quarter. That is a real result, but it is a market-conditions result, and a flat or falling quarter for equities would not repeat it.

Investment banking fees rose 55% to $3.40 billion, and the mix matters. Equity underwriting more than doubled, from $428 million to $985 million, "primarily reflecting significantly higher net revenues from secondary and initial public offerings." Debt underwriting rose 75% to $1.03 billion on "leveraged finance and asset-backed activity." Advisory — fees for advising on mergers, historically Goldman's signature business — grew the least, up 17% to $1.38 billion, "reflecting an increase in industry-wide completed mergers and acquisitions volumes." So the fee surge is led by capital-markets issuance, which is the most cyclical part of the franchise, not by the steadier advisory book.

FICC was solid rather than spectacular. Fixed income, currency and commodities revenues were $4.59 billion, up 32%, with intermediation up 39% to $3.38 billion. Management's explanation is candid about where that came from: revenues in interest rate products and currencies "reflected the impact of improved market-making conditions on our inventory," and the same phrase appears for commodities alongside higher client activity. That is mark-to-market gains on positions Goldman already held as conditions improved, not a durable increase in customer flow. Credit products went the other way, on lower client activity.

Asset & Wealth Management grew steadily, with one soft spot. Segment net revenues of $4.60 billion (+20%) came from management and other fees of $3.36 billion (+19.7%) and a jump in Investments revenue from $137 million to $441 million on "significantly higher net gains from investments in private equities" — a mark-to-market gain on Goldman's own balance-sheet holdings, and therefore not a recurring fee stream. Private banking and lending revenue fell 12.7% to $689 million because of "lower net interest margin related to Marcus deposits": Goldman is paying up for retail deposits, and the spread it earns on them has compressed even as balances grew.

Assets under supervision — client money Goldman manages or advises on, and the base its management fees are charged against — reached $4.04 trillion, up from $3.29 trillion a year ago. The quarter added $391 billion, of which $230 billion was net new client money and $161 billion was market appreciation. Notably, $139 billion of the inflow went into liquidity products (money-market funds and private bank deposits), the lowest-fee category, so AUS growth will translate into fee growth at less than a proportional rate.

Platform Solutions is a shrinking drag. Net revenues fell to $221 million from $619 million, "primarily reflecting net markdowns related to the Apple Card loan portfolio, which was transferred to held for sale in the fourth quarter of 2025." That portfolio stood at $19.5 billion at 30 June. Goldman agreed in December 2025 to transition the Apple Card program to another issuer over roughly 24 months, so further marks are likely as the book is carried at the lower of cost or fair value until it moves. The segment lost $48 million pre-tax in the quarter.

Costs, capital and tax

Operating expenses of $11.67 billion rose 26%, driven by compensation and benefits (+30% to $6.10 billion, which the filing ties to "improved operating performance" — that is, bonus accruals tracking the revenue surge) and transaction-based expenses (+56% to $3.05 billion, largely brokerage and clearing costs that rise mechanically with trading volume). Because revenue rose faster than costs, the efficiency ratio improved to 57.4% from 63.4%. Headcount was 46,200, essentially flat year over year and down 2% from March 2026 — the cost increase is almost entirely pay-per-head and volume-linked, not hiring.

One comparison needs care. Goldman's effective tax rate for the first half of 2026 was 18.5%, flattered by roughly $965 million of tax benefits on the settlement of employee share-based awards, which the filing says cut the half-year rate by 6.5 percentage points and added about $3.15 to half-year diluted EPS. Most of that landed in the first quarter (13.2% rate). The second quarter's own rate was 22.6%, against 24.9% a year earlier, so the Q2 EPS figure is not materially inflated by that benefit — a distinction that gets lost if you read only the half-year numbers.

EPS grew faster than net income (92% versus 78%) because the share count keeps falling. Goldman repurchased $4.00 billion of stock in the quarter and paid $1.36 billion in dividends, $11.74 billion of total capital returned in the first half; average diluted shares fell 4.2% to 304.9 million. Book value per share ended at $367.67, up 2.8% since December. The Common Equity Tier 1 ratio — the regulatory measure of loss-absorbing capital against risk-weighted assets — was 12.9% on the Standardized approach, leaving room to keep buying back stock.

Takeaway: This was a genuinely excellent quarter, but the quality of the earnings is more cyclical than the 23.5% ROE suggests. The three largest drivers — prime financing into a 15% S&P 500 rally, equity underwriting that more than doubled, and FICC gains the filing explicitly attributes to "improved market-making conditions on our inventory" — are all market-dependent. The recurring fee engine, advisory (+17%) and management fees (+19.7%), grew respectably but nowhere near 39%. Sustaining mid-20s ROE requires the market backdrop to hold.

Outlook

Goldman gives no revenue or earnings guidance, but the filing contains three forward indicators.

The investment banking backlog — management's estimate of revenue from deals it considers more likely than not to close — rose versus March 2026, with "higher estimated net revenues from potential advisory transactions, partially offset by significantly lower estimated net revenues from potential debt underwriting transactions (primarily from leveraged finance activity)." That mix shift is mildly encouraging for revenue quality, since advisory fees are less volatile than leveraged-finance underwriting, but it points to the second-quarter debt underwriting strength not repeating at the same level.

On tax, management states it expects the 2026 annual effective rate to be approximately 20%, above the 18.5% booked in the first half — so the second half carries a modestly higher tax drag.

On capital, the board raised the quarterly dividend to $5.00 per share from $4.50 on 13 July 2026, payable 29 September. Combined with $9.00 billion of first-half buybacks, that signals management sees no near-term capital constraint. In Asset & Wealth Management, the stated medium-term targets are a high-teens segment ROE (roughly 17–19%) and about 30% pre-tax margin within three to five years from year-end 2025; the segment delivered 13.8% ROE this quarter and a 24% half-year pre-tax margin, so both remain some distance away.

Our read: The structural improvements are real — the efficiency ratio is 6 points better, headcount is flat, the consumer-lending experiment is being dismantled, and the fee base in wealth and asset management keeps compounding at high-teens rates. But roughly half of the incremental revenue this quarter came from businesses whose results move with asset prices and market volatility, and the year-ago quarter was depressed by credit-card provisions that no longer exist. A reasonable expectation is that ROE settles below 23.5% as those comparisons normalize, with the durable question being whether Asset & Wealth Management's fee margin can improve fast enough to close the gap to its own targets. Management's own framing of risk is explicit: if "market-making activity levels decline or investment banking activity levels decline," segment revenues "would likely be negatively impacted."