Inside Goldman Sachs (GS) Shares: Advisory, Trading and Fee Income

Few names on Wall Street are as closely watched, or as widely misunderstood, as Goldman Sachs Group Inc (GS). Any serious Goldman Sachs stock analysis has to start with a simple observation: this is no longer the firm that spent the early 2020s trying to become a consumer bank. Over the past three years management has steadily narrowed the business back toward what it has always done best, and the financial statements now read very differently as a result. This article walks through how the company actually earns its revenue, what the most recent reported numbers show, how the balance sheet and capital returns are structured, and which risks matter most to anyone following the shares.

How Goldman Sachs Makes Its Money

Goldman reports through three segments, and the concentration is striking. Global Banking and Markets is the engine. It houses investment banking advisory work on mergers and acquisitions, equity and debt underwriting, and the firm's two large trading franchises: FICC (fixed income, currencies and commodities) and Equities. Asset and Wealth Management is the second pillar, running money for institutions, ultra high net worth families and retail investors, and increasingly for private credit and alternatives clients. Platform Solutions, the remnant of the consumer experiment, is now a rounding error by design.

The distinction between those first two segments matters more than the segment labels suggest. Global Banking and Markets revenue is inherently cyclical. It rises when deal volumes recover, when corporate treasurers issue debt, when volatility gives trading desks something to price. Asset and Wealth Management revenue is closer to an annuity: management fees are charged on assets under supervision, and those fees reprice slowly. Investors who follow Goldman tend to watch the mix between the two, because a higher share of fee income is what would narrow the gap between how the market values Goldman and how it values a pure asset manager such as BlackRock (BLK).

There is a third revenue layer worth understanding. A large and growing part of both trading segments is financing rather than risk-taking: prime brokerage balances, securities lending, margin lending against client portfolios, and structured lending in FICC. These are balance-sheet-intensive but relatively steady, and they are one reason trading revenue has become less erratic than the pre-2008 caricature of an investment bank would imply.

What the Most Recent Results Show

Goldman reported second-quarter 2026 results on 14 July 2026, and the quarter was an unusually strong one. Net revenues were $20.34 billion, 39% higher than the second quarter of 2025. Net earnings were $6.63 billion, and diluted earnings per common share came in at $20.98. Annualized return on common equity was 23.5%, against 12.8% in the same quarter a year earlier.

The segment detail explains where that came from. Global Banking and Markets produced net revenues of $15.52 billion, up 53% year over year and 22% on the first quarter of 2026. Within that, investment banking fees were $3.40 billion, 55% higher than a year earlier, driven by equity and debt underwriting. FICC net revenues were about $4.6 billion, up 32%. Equities set a quarterly record at roughly $7.4 billion.

Asset and Wealth Management net revenues were $4.60 billion, 20% higher than the second quarter of 2025 and 13% higher than the first quarter. Management and other fees reached a record $3.4 billion, up 20%, reflecting higher average assets under supervision. Total assets under supervision finished the quarter at a record $4.04 trillion, having risen $391 billion during the three months, split between $230 billion of net inflows and $161 billion of market appreciation.

Context is useful here, because a 23.5% return on equity is not the firm's normal run rate. For the full year 2025, Goldman reported net revenues of $58.3 billion, up 9%, diluted earnings per share of $51.32, up 27%, and a return on equity of 15.0%. The first quarter of 2026 delivered net revenues of $17.23 billion, EPS of $17.55 and an annualized return on equity of 19.8%. The trajectory has been upward, but the second quarter sits well above both the prior quarter and the prior full year, and capital markets activity is the reason.

Costs, Credit and the Consumer Retreat

Operating expenses were $11.67 billion in the second quarter of 2026, 26% higher than a year earlier and 12% higher than the first quarter. In a business where a large share of the cost base is variable compensation, expenses rising alongside revenue is expected rather than alarming; the question is whether they rise more slowly. On that measure the firm made progress: the efficiency ratio for the first half of 2026 was 58.8%, against 62.0% for the first half of 2025, helped by a compensation ratio net of provisions of around 31%.

Credit costs were modest. The provision for credit losses was $102 million in the second quarter of 2026, compared with $384 million in the second quarter of 2025 and $315 million in the first quarter of 2026, with the quarter's provisions primarily reflecting impairments on wholesale loans. That is a very different credit profile from a deposit-funded consumer lender such as Bank of America (BAC) or Citigroup (C), and it is a direct consequence of the strategic retreat described below. Wells Fargo (WFC) belongs in the same comparison, a big-four lender whose profits lean on deposit-funded lending rather than on underwriting and trading.

That retreat is now essentially complete. In January 2026 Goldman announced an agreement to transition the Apple Card program to JPMorgan Chase (JPM), ending its partnership with Apple (AAPL) and closing the most visible chapter of its consumer banking push. The card portfolio had already been moved to held for sale, and the markdowns associated with it are what crushed Platform Solutions revenue: the segment produced just $221 million of net revenues in the second quarter of 2026, 64% below the prior year, after $411 million in the first quarter. Management has described the transaction as substantially completing the narrowing of the consumer business.

For a Goldman Sachs stock analysis, the significance is structural rather than arithmetic. The consumer venture consumed capital, generated elevated credit losses and diluted returns for several years. Removing it simplifies the earnings model considerably, but it also removes the diversification argument that originally justified it. What remains is a firm whose results are once again tightly geared to capital markets conditions.

Capital, Dividends and Shareholder Returns

Goldman's capital position strengthened through the second quarter of 2026. The standardized Common Equity Tier 1 ratio rose to 12.9% from 12.5% in the prior quarter, and the advanced ratio to 13.7% from 13.3%. Following the Federal Reserve's 2026 Comprehensive Capital Analysis and Review, published on 24 June 2026, the firm's stress capital buffer remains 3.4% through 30 September 2027, leaving a standardized CET1 requirement of 11.4%. The reported ratio therefore sits comfortably above the regulatory floor, which is what creates room for distributions.

Those distributions have been growing. Goldman stated its intention to raise the quarterly common dividend from $4.50 to $5.00 per share from 1 July 2026, an increase of 11% on the prior level and 25% on the year-earlier rate. During the second quarter the firm returned $5.36 billion to common shareholders, comprising roughly $4.0 billion of share repurchases and $1.36 billion of common dividends.

Book value per common share ended the second quarter of 2026 at $367.67, 1.8% higher over the three months. For a capital markets firm, book value per share is arguably the most useful single anchor: it grows through retained earnings and shrinks through buybacks executed above book, and the ratio of price to book value is the multiple most commonly used to compare Goldman with peers such as Morgan Stanley (MS) and JPMorgan Chase (JPM). What that ratio measures is how much investors are willing to pay for each dollar of accounting equity, and it tends to expand when returns on that equity are high and persistent, and compress when they are not. For a contrast drawn from outside the capital markets cycle, Standard Chartered (STAN) is a bank whose returns rest on cross-border trade corridors and Asian wealth rather than on underwriting and trading, and setting the two side by side shows how differently a banking multiple can be earned.

The Risks That Matter

The first and largest is cyclicality. Investment banking fees of $3.40 billion in a single quarter reflect a buoyant underwriting calendar. Deal pipelines can close quickly when rates move, credit spreads widen or equity markets turn, and advisory and underwriting revenue is the line item that falls fastest. A 23.5% annualized return on equity in a strong quarter tells you what the franchise can earn at the top of a cycle, not what it earns through one.

Second, trading concentration. Equities generated a record quarter, and Equities plus FICC accounted for the majority of Global Banking and Markets revenue. Trading results depend on client activity and on volatility, neither of which the firm controls, and record quarters are by definition unusual.

Third, regulatory capital. The stress capital buffer is recalculated annually and is sensitive to the Federal Reserve's scenario design. A higher buffer in a future cycle would raise the CET1 requirement and mechanically reduce the capital available for buybacks, which is currently a meaningful component of shareholder returns.

Fourth, credit and private markets exposure. Provisions are low today, but Goldman has been expanding in private credit and alternatives, where losses tend to appear late and in concentrated form. The wholesale loan impairments recognized in the second quarter are a reminder that this book is not risk-free.

Fifth, compensation. A compensation ratio near 31% is competitive by historical standards, but talent is mobile and pay is the principal cost lever. If revenue growth slows while compensation stays sticky, operating leverage reverses quickly.

What to Watch From Here

Several measurable things would change the picture. The clearest is the durability of investment banking fees: whether the underwriting and advisory strength seen in the first half of 2026 persists into subsequent quarters, or proves to have been a pull-forward. The second is the mix of assets under supervision: net inflows of $230 billion in a single quarter are exceptional, and the composition of those flows, particularly the share going into higher-fee alternatives rather than low-fee liquidity products, determines how much management fee revenue they actually produce.

The third is the efficiency ratio. Moving from 62.0% to 58.8% between the first halves of 2025 and 2026 was real progress; holding below 60% in a weaker revenue quarter would be more informative than achieving it in a record one. The fourth is book value per share growth alongside buyback pace, since repurchases funded out of a strong capital position compound book value per share only if executed at sensible prices. The fifth is credit provisioning, where any sustained move back toward the $300 to $400 million quarterly range seen in 2025 would signal that the wholesale and private credit books are deteriorating.

Finally, it is worth watching whether the Platform Solutions line settles at a genuinely immaterial level now that the Apple Card transition is agreed, or whether residual markdowns continue to appear. A clean exit removes noise from reported earnings and makes year-on-year comparisons meaningful again, which matters for anyone trying to assess the underlying trend.

Taken together, the evidence is of a firm that has simplified its business, restored high returns on equity in favorable conditions, strengthened regulatory capital and raised its dividend, while remaining tightly exposed to the capital markets cycle it deliberately chose to concentrate on. Those are the variables to track, and the reported figures above provide the baseline against which the next set of results can be measured.