Goldman Sachs (GS) at 14.5 Times Earnings After a $20.3B Quarter
Goldman Sachs took in $20.3 billion of revenue in its latest quarter, 39% more than a year earlier and 18% more than the quarter before. Multiply that by four and you get an $81.4 billion annual pace, against $58.3 billion for all of 2025. Yet the forward EPS analysts use, $68.23, is only 5% above the $64.76 the firm earned over the last twelve months. Those two facts sit uneasily together, and the space between them is the whole debate about this stock.
My opinion of Goldman fits in five words: excellent franchise, hard to value. The shares trade around $942.00 as I write this, at 14.5 times trailing earnings, and the market value is $274.3 billion. My thesis is short. The quarter proves that Goldman can earn a great deal in a good market, but the price already assumes a good share of it sticks, and the balance sheet multiples, not the earnings multiple, are where that assumption shows.
What a 39% jump tells you
Revenue growth of 39% for a bank of this size is rare. For context, annual revenue went from $47.4 billion in 2022 to $46.3 billion in 2023, then $53.5 billion in 2024 and $58.3 billion in 2025. That is four years in which the firm added roughly 23% in total. One quarter running 39% above its own year-earlier figure compresses a lot of that into three months.
Goldman is not a deposit bank. Its earnings come from advising on mergers, underwriting stock and bond sales, and above all trading with and for institutional clients. Trading revenue rises when markets move and clients reposition, and it thins out when everything is calm. I read a quarter like this one as a description of the environment first and of the firm second. When the desks are busy, Goldman collects. When they are not, the fixed costs of the people and technology stay.
The sequential number matters as much as the yearly one. Revenue was up 18% from the prior quarter, so the pace was still accelerating, not fading. That is the tricky part. A stock that is priced after the best quarter in a series is being priced on a peak-adjacent number, and the honest question is how much of the peak repeats.
I looked at what the results imply for the yearly totals in the Second-Quarter 2026 side-by-side of four big US banks, where Goldman’s trading-heavy mix is the obvious outlier. Its peers earn more from lending spreads, which move slowly. Goldman earns more from activity, which does not.
Trailing versus forward earnings
Here is the arithmetic that bothers me. Trailing EPS is $64.76. Forward EPS is $68.23. The difference is 5%. If the latest quarter really were a new run rate, the forward figure would sit far above the trailing one, because the trailing figure still includes three older, weaker quarters.
There are two readings. One is that analysts are simply slow: they update forecasts in steps, and the next few weeks of revisions will push $68.23 higher. The other reading is that analysts treat the quarter as partly one-off and model the following quarters at a lower level. I lean toward the second, mainly because of what I said about trading. But I hold that at about sixty percent, not ninety.
| Metric | Now | Five-year average | Gap |
|---|---|---|---|
| P/E (trailing) | 14.5 | 13.3 | about 9% above |
| P/E (forward) | 13.8 | n/a | n/a |
| Price to book | 2.6 | 1.5 | about 73% above |
| Price to sales | 4.3 | 3.0 | about 43% above |
Why book value is the number to check
The earnings multiple of 14.5 looks calm. Slightly above the 13.3 five-year average, nowhere near a bubble. But earnings for a trading-heavy bank are the noisiest number on the page, and a low P/E on a peak year can be the trap. Price to book is steadier, and it says something different: 2.6 times book today against a 1.5 average.
A rough conversion helps. At 2.6 times book and $942, book value is about $362 per share. Trailing EPS of $64.76 over $362 gives a return on equity near 18%. That is a good return, and part of why the multiple is high. But a bank that trades at 2.6 times book is pricing in a return well above its cost of capital for years. If the return slips back to the low teens, which is more typical for a large investment bank across a full cycle, the multiple gets pulled back toward the old average.
I do not want to overdo this. A higher price to book is not a sell signal by itself; the quality of the franchise in advisory and prime financing has improved, and part of the premium is deserved. What I would refuse is the idea that the premium is free.
Investment banking is the steadier part
Trading is volatile, but advisory and underwriting are the reason I do not think this is a pure cyclical peak. Deals that were put off while rates were uncertain do not disappear. They wait. Once financing conditions stabilize, those companies come back to sign mergers and list shares, and Goldman sits at the front of most of those queues.
Fees from that work are lumpy too, but they follow the deal calendar more than daily price swings. If banking fees keep climbing while trading returns to normal, the earnings base is sturdier than the headline quarter suggests. If banking fees stall the moment markets get choppy, then the record was a market gift and nothing else. Comparing the two revenue lines each quarter is the single most useful thing an owner can do.
For the same reason, I would read the JPMorgan record-high piece alongside this one. JPMorgan earns from a wider base, so it shows what the same environment looks like for a bank that is less exposed to trading.
Payout, targets and the last earnings move
The dividend is $17.00 per share over the last twelve months, a yield of 1.80%. That is about 26% of trailing EPS, a low payout. It leaves the firm room to keep raising the dividend and to buy back shares, and it means income is not the reason to own this. Total return here is about the price.
On the price, thirteen analysts have an average target of $1,203, which is 28% above $942. The lowest target is $1,075, 14% above, and the highest is $1,325, 41% above. So even the most cautious analyst sees the stock higher. Only 46% of them rate it a Buy, though, which is a low share for a stock whose lowest target is above the price. That gap says the ratings are more careful than the targets. I take it as a sign that many analysts think the stock is fair but the forecast could be raised, not that the stock is cheap.
The last report came on July 14, and the shares moved +9.0% that day, more than twice the 4.4% average move around earnings. A move that size shows how much surprise was in the numbers and how little of it the market expected. It also raises the bar: the next report has to be judged against a higher starting price than the last one had.
The stock sits between a 52-week low of $726 and a high of $1,148, so the range is wide. Short interest is 2.4% of shares, unremarkable.
A sensitivity table I would actually use
Take forward EPS of $68.23. At the five-year average P/E of 13.3, the shares would be worth about $907, 4% below the current price. Now cut earnings by 20% to about $54.6, a bad-but-not-terrible year for trading, and keep the same average multiple: about $726, which is exactly the 52-week low. Push in the other direction: if EPS reached $80 and the multiple held at 14.5, the value is about $1,160.
None of these are forecasts. They are a way to see that the stock has little cushion between “earnings hold” and “earnings slip.” A holder is paid for a strong environment continuing, and punished quickly if it does not.
Where I would be wrong
My view fails if two things happen. First, if forward EPS climbs quickly toward $80 over the next two quarters while banking fees keep growing, the 14.5 times earnings would look like about 12 times, and the book multiple would start to look fair. I would have been too cautious. Second, if the trading share of revenue falls back and the pieces that replace it are deals and asset management fees, the business really is more stable than I give it credit for.
The uncertainty I cannot remove is timing. I do not know when the market environment changes, and neither does anyone who publishes a price target. I only know that the firm earns the most when I would least want to be paying for it.
The two figures I would watch next
I would follow the $68.23 forward EPS figure: a rise to $75 or more before the next report would tell me the analysts expected the quarter to repeat, and I would revisit my thinking. And I would follow price to book: with the shares around $942, I would want to see the multiple settle below 2.2 times, or the earnings to catch up with it, before calling the stock inexpensive for a bank. For a first reading of how much risk sits in a position like this, my note on sizing positions when volatility is high is a practical companion.
Analysis and opinion only, not investment advice. Figures come from Goldman Sachs’s filings on SEC EDGAR and the CNBC report on big-bank second-quarter earnings; market data and valuation multiples are approximate and were checked on September 18, 2026.