Goldman Sachs said third-quarter earnings for the S&P 500 remain powerful, but the phase of sequential acceleration appears to be over.
In its Oct. 30, 2026 report, US Weekly Kickstart: Q3 2026 midseason earnings update, Goldman Sachs’ U.S. portfolio strategy team said 60% of S&P 500 companies had reported by Oct. 30. Its central conclusion was straightforward: earnings growth is still strong, but it is no longer getting stronger quarter after quarter.
Goldman estimated third-quarter S&P 500 EPS growth at roughly 33% year over year, excluding “other income” tied to large-cap technology firms and equity investments. That matched the second quarter’s realized growth rate and stood far above the 10-year median of 10%. Analysts are still raising 2027 earnings forecasts, but the scale and breadth of those upward revisions have weakened. At the same time, expectations for hyperscaler capital spending were barely revised this season, leading Goldman to argue that AI capex growth is peaking.
The internal market picture has shifted as well. Since the start of the third-quarter earnings season, the S&P 500 has fallen 3%, while the S&P 500 excluding AI-related stocks has gained 5%. Goldman’s reading is that an extremely narrow market is broadening through two moves at once: AI stocks are giving back ground, while non-AI shares are catching up.
The bank kept its S&P 500 targets unchanged at 8,000 for year-end 2026 and 8,700 over 12 months. Against a current level of 7,720, that implies upside of 4% and 13%.
Earnings growth held at 33%, but the acceleration has stopped
Goldman’s key earnings figures were as follows:
- Third-quarter S&P 500 EPS growth was about 33% year over year, excluding “other income” linked to large technology companies and equity investments, in line with the second quarter. The 10-year median is 10%.
- The median constituent posted EPS growth of 12%, above the 10-year median of 8%, but slower than the prior quarter’s 14%.
- Excluding non-recurring items, Goldman expects S&P 500 EPS growth to slow from nearly 30% in 2026 to 18% in 2027, still a high level.
- Median company revenue growth eased slightly, while net margins reached a new high.
The share of companies beating expectations has also come down. Among companies that had reported, 60% beat consensus EPS estimates by more than one standard deviation. That is above the historical average of 50%, but below the 64% average of the previous four quarters. At the index level, third-quarter earnings are still beating expectations, though by a smaller margin than in the first and second quarters.
Two sectors drove most of the growth. Information technology and communication services together accounted for more than 60% of third-quarter S&P 500 EPS growth, while energy contributed about 20%.
| Sector | EPS growth | Revenue growth | Margin change |
|---|---|---|---|
| Energy | 138% | 25% | +736bp |
| Information Technology | 67% | 43% | +454bp |
| Communication Services | 46% | 15% | -75bp |
| Materials | 39% | 13% | +232bp |
| Industrials | 23% | 14% | +81bp |
| Health Care | 4% | 3% | +5bp |
| Financials | 4% | — | — |
| Consumer Staples | 2% | 5% | -19bp |
| S&P 500 | 33% | 12% | — |
Energy EPS growth of 138% was tied to a sharp rise in oil prices. In industrials, median company margins have expanded strongly for several quarters and moved higher again in the third quarter. Consumer staples margins remained under pressure.
Earnings beats are being rewarded again, except in tech
Goldman said a pattern from recent quarters has started to reverse. For several quarters, companies that beat earnings expectations had underperformed on average. This season, that changed.
The median stock that beat EPS expectations outperformed the S&P 500 by 152 basis points the day after reporting, above the historical median of 95 basis points. Cyclical sectors saw the biggest reward: industrials at +371bp, consumer discretionary at +326bp, materials at +204bp, and financials at +183bp. Goldman noted that hedge fund positioning in those sectors had already fallen sharply.
TMT was the exception. Communication services posted -419bp, information technology -60bp, and TMT overall -76bp, versus +172bp for non-TMT stocks. Consumer staples, at -181bp, was another sector where earnings beats still led to underperformance.
Goldman’s summary was that good news can still move lightly owned cyclical stocks, while heavily owned technology names had already priced in much of the upside.
The week’s gainers and losers through Oct. 30 reflected the same split. Among the biggest decliners were software and semiconductor names: ServiceNow fell 34%, F5 dropped 26%, Fortinet and Intel each lost 24%, Twilio fell 22%, and Marvell declined 14%. On the upside, Pentair and Vertiv each rose 20%, Carvana gained 18%, and General Motors added 15%.
Market breadth is widening through catch-up and pullback at the same time
Goldman highlighted this as one of the most useful trading signals in the report.
Since the start of the third-quarter earnings season, 70% of S&P 500 constituents have posted positive returns, but those companies represent only 50% of the index by market capitalization. The result is that the S&P 500 has fallen 3%, while the ex-AI S&P 500 has risen 5%.
Goldman said the market’s previously extreme narrowness is easing as AI stocks retreat and the rest of the market moves higher. Earlier data in the report pointed in the same direction: non-TMT stocks that beat expectations outperformed by a median 172 basis points the day after earnings.
Year-to-date total returns, based on pricing through Oct. 29, showed the same divergence:
- Crude oil (S&P GSCI) +128%, energy sector +57%, Russell 1000 Value +26%, information technology +25%;
- Equal-weight S&P 500 +16%, cap-weighted S&P 500 +14%, Russell 1000 Growth +3%;
- Communication services -10%, Bitcoin -11%, 10-year U.S. Treasuries -3%, gold -7%.
Goldman said the next question is whether non-technology sectors can keep rising, which in its view depends heavily on energy prices and interest rates.
AI capex growth is peaking
In prior quarters, the market typically raised hyperscaler capital expenditure expectations after earnings season. Goldman said that did not happen this time.
Consensus capex forecasts for Amazon, Google, Meta, Microsoft, Oracle and peers were listed as follows:
| Year | Consensus capex forecast | YoY growth |
|---|---|---|
| 2026 | About $808 billion | +96% |
| 2027 | About $1.163 trillion | +44% |
| 2028 | About $1.289 trillion | +11% |
After this quarter’s earnings releases, the roughly $1.2 trillion forecast for 2027 was only slightly above the pre-earnings level. Goldman said capex should still rise sharply in 2027, but both capex growth and the scale of upside surprises have already peaked.
That matters directly for the AI infrastructure chain. AI infrastructure stocks such as semiconductors and power names posted third-quarter EPS growth of about 80%, more than five times the 15% pace for the rest of the S&P 500. Consensus estimates suggest that growth rate will slow materially over the next few quarters and fall to around 40% by the end of 2027.
Cash flow pressure has already emerged. Hyperscalers as a group turned free cash flow negative in the third quarter, with capital expenditure exceeding operating cash flow. Analysts expect free cash flow for that group to keep falling in 2027 and turn positive only in 2028.
AI monetization is improving, but investor demands are higher
Goldman said evidence of AI returns is still building and remains one of the few areas where momentum is still improving.
Cloud revenue at Amazon, Google and Microsoft grew 54% year over year in the third quarter, up from 48% and 39% in the prior two quarters. Backlog growth for the three companies also remained strong, though it slowed slightly from the second quarter.
The issue, Goldman said, is expectations. Hedge funds entered earnings season with concentrated positions in hyperscalers. Cloud revenue beat consensus, but still did not clear the bar investors had set.
Third-quarter results also revived concern about AI disruption risk, especially in TMT. Goldman analysts cited disappointing Google search revenue and lower search-driven traffic to Reddit. Goldman economists also said they were seeing early signs of AI’s impact on the labor market.
2027 earnings revisions are still positive, but weaker
Since the start of earnings season, consensus 2027 S&P 500 EPS forecasts have risen 2%.
Upward revisions still outnumber downward revisions, but revision breadth has narrowed sharply from recent months, with only half of sectors still positive.
Over the past month, 2027 EPS revision breadth by sector — defined as the number of upward revisions minus downward revisions, divided by the total number of companies — was strongest in information technology at 64%, energy at 43%, industrials at 24%, and financials at 20%. The S&P 500 overall stood at 10%.
The weakest sectors were real estate at -39%, consumer staples at -36%, consumer discretionary at -28%, and communication services at -18%.
Goldman also pointed to a detail inside the revisions. For companies that had already reported, analysts generally raised 2027 revenue forecasts while cutting margin expectations. Revenue upgrades were largest in energy and information technology. Margin downgrades were tied mainly to higher input costs, including energy, and both management commentary and analyst revisions reflected that pressure.
Goldman’s forecasts and positioning
Using an S&P 500 level of 7,720 as of Oct. 29, Goldman compared its own forecasts with bottom-up consensus and top-down strategist consensus:
| Goldman | Bottom-up consensus | Top-down strategist consensus | |
|---|---|---|---|
| 2026 EPS | $375 (+36%) | $367 (+33%) | $356 (+30%) |
| 2027 EPS | $415 (+11%) | $425 (+16%) | $410 (+15%) |
| 2028 EPS | $460 (+11%) | $495 (+16%) | — |
| 2026 year-end target | 8,000 (+4%) | — | — |
| 12-month target | 8,700 (+13%) | 9,150 (+19%) | — |
Goldman noted that these EPS figures use its adjusted methodology and are not directly comparable with the earlier figures of nearly 30% growth in 2026 and 18% in 2027 after excluding non-recurring items.
On valuation, the S&P 500 trades at 19x forward earnings, in line with bottom-up consensus. Goldman said that even with lower valuation multiples and greater macro volatility, earnings improvement could still support a fourth straight year of double-digit returns for the index.
The bank is more conservative than the market on 2027 earnings. Goldman projects 11% EPS growth for 2027, below the 16% bottom-up consensus. The gap is concentrated in information technology and communication services: Goldman expects +33% for information technology versus +41% consensus, and -18% for communication services versus -9% consensus.
In sector allocation, Goldman is overweight health care and materials. On valuation, communication services trades at 16x forward earnings, in the 4th percentile of the past 10 years, making it the cheapest sector. Health care trades at 20x, in the 96th percentile, while consumer staples trades at 23x, in the 92nd percentile.
The same report’s cross-asset 12-month forecasts were:
- U.S. 10-year Treasury yield: 5.0% now, 4.5% target in 12 months (-53bp);
- Brent crude: $111 per barrel now, $75 target in 12 months (-32%);
- COMEX gold: $4,031 per ounce now, $5,140 target in 12 months (+27%).
On fund flows, Goldman said that over the past 12 months, U.S. mutual funds and ETFs saw cumulative inflows of about $380 billion into equities, about $501 billion into money markets, and about $839 billion into bonds.
Risks and what Goldman is watching next
Goldman flagged several areas to watch.
Energy prices come first. Energy was the second-largest contributor to third-quarter earnings growth and also a major reason margin expectations were cut across many sectors. Goldman expects Brent to fall to $75 over 12 months. Lower oil prices would hurt energy earnings but help margins recover elsewhere.
Rates are another key variable. The U.S. 10-year Treasury yield is at 5.0%, and Goldman expects it to fall to 4.5%. If rates do not decline, the room for expensive non-technology sectors to keep catching up may be limited.
Goldman also pointed to slower AI earnings growth. AI infrastructure earnings growth is expected to slow from 80% to around 40% by the end of 2027, while free cash flow has already turned negative. If capex expectations start to move lower, that chain would be directly exposed.
AI disruption risk is also back in focus. Concerns around disruption in search and software are increasing, and software stocks were among the week’s biggest decliners.
Another issue is the gap between Goldman’s forecasts and consensus. Goldman’s 2027 EPS forecast is about 2% below consensus, mainly in technology and communication services. If consensus estimates move toward Goldman’s view, those two sectors would face the greatest revision pressure.
There is also more earnings data ahead. Next week, another 143 companies are scheduled to report, representing 21% of S&P 500 market capitalization. Within information technology, 29% of sector market cap will report next week.
The targets, earnings forecasts, sector calls and related judgments cited here are Goldman Sachs analyst views from the firm’s Oct. 30, 2026 report. They do not constitute investment advice.
