Goldman Sachs Analysts Increase Their Forecasts After Strong Q2 Results

Goldman Sachs reported Q2 EPS of $20.98, far above the $14.40 analyst consensus, and net revenue up 39% YoY to $20.34B versus $16.13B expected, led by strength in Global Banking & Markets. CEO David Solomon attributed rising capital demand to the early AI investment cycle expanding beyond core tech into infrastructure, energy, and data centers. Shares were up 0.3% pre-market to $1,143 as analysts adjusted price targets after the earnings beat.

Analysis

GS is signaling that the higher-beta parts of financials are still the cleanest way to monetize an AI-led capex cycle. The immediate implication is not just a one-quarter EPS reset; it is a broader read-through that underwriting, ECM, and strategic advisory should stay bid as AI spend migrates from software into power, cooling, networking, and real assets. That favors infrastructure suppliers and financing conduits over pure software names, because the near-term cash outlay comes before monetization.

Second-order losers are the balance-sheet owners of the spend: hyperscalers and AI-platform leaders may see near-term margin pressure if the market continues to reward growth but punish capex intensity. That makes GOOGL a useful hedge against the “AI spend is all good news” consensus. If the cycle is real, the revenue capture for banks and picks-and-shovels should show up over 1-3 quarters; if it is just a burst of one-off financing, the trade will fade quickly.

Contrarian takeaway: the market may be underestimating how much of GS’s beat is cyclical beta rather than a permanently larger fee pool. The falsifier is simple: if capital markets activity stalls next quarter or if AI-related capex does not translate into visible lending/fee wins, the multiple expansion on GS should reverse. On the other hand, a sustained pickup in data-center financing would make OZK an interesting secondary beneficiary, but only after loan growth confirms the thesis.

Key Decisions for Investors

• Add GS on any 1-2% post-earnings pullback; target 5-8% upside over 1-3 months as estimates for trading and IB fees reset higher. Cut if next-quarter capital markets revenue fails to hold current run-rate.
• Pair trade: long GS / short GOOGL for 1-3 months to express AI capex monetization lag. GS should capture financing/advisory flow while GOOGL bears the capex drag; thesis breaks if GOOGL capex begins to convert into outsized revenue acceleration.
• Set a watchlist entry for OZK as an indirect data-center financing proxy. Only initiate a small long if loan growth tied to AI infrastructure becomes visible and CRE credit metrics stay stable; otherwise keep it as an alert, not a trade.