- Another strong earnings quarter is taking shape. The third quarter is on track to mark the 8th consecutive quarter of double-digit earnings growth. Roughly half of S&P 500 companies have reported, with 80% beating consensus earnings expectations and EPS growth running at ~28% year-over-year.
- AI and Energy are doing the heavy lifting, but the rest of the market is picking up. Aggregate net income growth is tracking an impressive 35% year-over-year. Excluding AI/tech and Energy, growth is a more modest 13%, but that’s the strongest showing for the rest of the market since the fourth quarter of 2024.
- AI demand remains strong. The harder question is what all that spending will earn. As hyperscaler chip spending surges and new players enter the fray, semiconductor demand shows little sign of cooling, with revenue backlogs pointing to robust demand through early 2028. For hyperscalers and neoclouds alike, monetization and efficiency gains need to keep pace with enormous capital spending requirements.
- Markets are waiting for AI to pay off, but some of the revenue is already here. AI is helping lift existing advertising and search businesses, while cloud capacity and chip sales are generating revenue from the buildout. Consumer subscriptions, devices and AI agents offer new opportunities, though still small relative to capex.
- As rates remain higher for longer, the once impenetrable hyperscaler balance sheets are growing increasingly sensitive to the higher cost of financing more spending, even more so for neoclouds. In theory, higher rates should direct capital to the highest-yielding investments; however, its clear markets will not blindly reward more spend given uncertainty around future profitability.
A stellar earnings song remains on repeat
The tune of strong corporate earnings continues. The S&P 500 3Q26 earnings season is tracking expected earnings growth of 28% y/y, the third consecutive quarter above 25%. Estimates rose 1.4% during the quarter rather than experiencing the typical ~2%+ cut.1
Since the start of the year, earnings have been revised higher for most areas of the market, especially for hyperscalers and semiconductors within the AI value chain. What’s more, 2027 earnings are being revised higher as well generally across the board, apart from power and software. While this provides support in today’s market, upward revisions can cut both ways: if companies are able to beat those elevated expectations, this justifies a bullish view; however, a higher earnings bar becomes more difficult to beat on estimates.
AI’s strongest fundamental story remains upstream
Information Technology earnings are expected to increase approximately 64% year-over-year, supported by revenue growth of roughly 43%. Every reporting Technology company has generated positive earnings growth.2
The more important development is the movement of the AI bottleneck. Constraints initially centered on graphics processing units, but have progressively migrated into memory, advanced packaging, semiconductor equipment, and ultimately manufacturing capacity. Major semiconductor companies continue to report record quarterly sales and indicate that planned supply through 2027 and 2028 may not be sufficient to satisfy stated customer demand. Some customers are already expressing interest in equipment required for 2029 and 2030 deployments.
The implication is straightforward: end demand is not yet the limiting factor. The near-term constraint is the physical ability to manufacture, power and deploy the required compute.
For investors, the AI opportunity is broadening, but so is the need to distinguish between businesses with durable pricing power and businesses benefiting temporarily from scarcity.
Monetization is no longer hypothetical
While investors remain focused on monetization, two things can be true: profitability can be driven by stronger sales in existing businesses, which are already reflected in the data, while new verticals, though more difficult to forecast, could still outperform expectations.
Meta and Amazon have provided an early proof point on monetization lifting existing businesses. Meta’s revenue increased 26% to $64.6 billion, with ad impressions +10% and price/ad +14%. Amazon Web Services (AWS) revenue growth has roughly tripled since 2023, while its profit margin improved to 39% from 35% last year. Amazon is also building its own custom chips and experimenting with agentic AI to help boost ad revenues.
AI monetization is becoming visible through several channels:
- Improving existing businesses. Search, advertising, recommendations and productivity represent the most visible and credible sources of current returns.
- Selling infrastructure. Cloud providers, semiconductor companies and data-center operators are earning revenue by enabling other companies’ AI investment.
- Creating new products. Consumer subscriptions, agents, devices and specialized applications could generate meaningful future revenue, but their economics remain less proven.
New businesses point to promising upside. More than 10 million users in the United States and Canada set up Meta’s Muse within approximately six weeks, with shopping emerging as its most common use case. However, the cost is enormous: Meta spent $34.9bn on 3Q capex and generated -$2.1bn free cash flow.
While demand remains strong, cash flow is increasingly being redirected toward chips and data centers. As the investment debate shifts toward AI capex returns, operating cash flow, depreciation, financing, and compute efficiency, investors should focus on clearer evidence of AI monetization.
Make some Space for X
Similar to chip demand, SpaceX sees an environment of robust demand with supply being the limiting factor. Revenue rose 161% to $14.6bn, with $7.6bn of "Super Intelligence" revenue already exceeding Connectivity and Space, and becoming the company's largest segment. The flip side is that the investment outlook continues to grow: 3Q capex was $19.6bn and management expects $35-40bn in 4Q, with power, cooling and infrastructure identified as deployment constraints. AI demand is spilling well beyond traditional Technology, creating beneficiaries in physical infrastructure, power, connectivity and related capacity. SpaceX’s expanding AI infrastructure business is bringing its investment profile closer to that of the established hyperscalers.
Is capital the next AI bottleneck?
Real interest rates have risen meaningfully this year, which has historically been a headwind to capital spending given the higher cost to fund new projects, and thus weighs on growth performance relative to value. But until now, higher real rates have done little to slow the AI investment cycle. The largest hyperscalers generated enough cash internally to fund much of their capital spending, weakening the traditional relationship between higher rates and investment. That remains broadly true today: Microsoft, Alphabet, Amazon and Meta have ample liquidity and substantial borrowing capacity.
But the financing mix is beginning to change. The broader AI ecosystem is becoming more dependent on bonds, loans, leases and structured financing just as the cost of capital is rising. AI-related companies have raised a record $460 billion globally this year, yet yields on at least 18 data-center financings have climbed into the double digits and demand for some new hyperscaler debt has weakened.3 Higher borrowing costs matter most for neoclouds and data-center developers, whose economics depend more heavily on leverage, refinancing and long-term customer commitments.
For now, this looks more like a rising hurdle than a funding constraint. Hyperscalers still have the balance-sheet capacity to sustain investment, and strong earnings across the AI stack continue to support spending. But with capex expected to reach new highs in 2027 and 2028, capital markets should matter more at the margin. Investors may therefore need to become increasingly selective, particularly among companies whose returns depend on cheap and continuously available financing, but the current environment does not appear to indicate a widespread funding crisis.
From a performance perspective, it is remarkable how rate insensitive growth performance has been in recent years. Prior to 2022, value tended to outperform growth when real interest rates rose and vice versa. Since then, growth has continued to outperform value despite the rise in real yields. Perhaps this dynamic shifts given the once profitable, cash rich companies in growth are now piling on debt. At the very least, investors should look to diversify their growth exposure as fast approaching IPO’s likely only increases the S&P 500’s growth/technology concentration.
Sector spotlight
- Financials: solid earnings, but uneven underlying momentum. Financials earnings continue to benefit from capital-markets activity, with equity trading helping offset more mixed fixed-income trading and investment-banking results. Lending strength is concentrated in corporate and upper-income Americans rather than a broad pickup in household borrowing, in line with our view that corporate activity and higher earning household spend remains robust. Low unemployment suggests middle- to lower-income consumers can make ends meet, but higher-for-longer rates could pressure credit quality if employment weakens. Encouragingly, it’s rare that employment weakens when profit growth is this strong. Financials earnings are expected to grow ~4% in 3Q, or 10% excluding Berkshire Hathaway due to unfavorable comps and lower rates.4
- Health Care: stabilizing insurers and drug innovation, but significant dispersion. After a string of weak quarterly results, healthcare is expected to see modest earnings growth this quarter of ~3%. Improving medical-cost visibility is supporting parts of managed care, although elevated labor costs remain a headwind. Pharmaceutical strength is concentrated in newer oncology and immunology therapies, while patent expirations and competition are weighing on older franchises.
Higher drug prices and expensive specialty treatments add pressure to patient affordability, while negotiated prices for selected drugs and reimbursement scrutiny constrain manufacturers’ pricing power. Biotech is benefitting from innovation and demand for pipeline acquisitions, but access to funding remains uneven, with clinical setbacks, approval uncertainty and commercialization costs separating promising science from durable earnings. Improving insurer cost visibility and strong drug innovation support a cautiously constructive backdrop, but pricing pressure, patent losses, regulatory hurdles and mixed device growth argue against a uniform sector recovery.
- Energy: huge earnings rebound, but different from a typical oil boom. Energy is expected to contribute about $30bn of S&P 500 incremental earnings, though the recovery is being driven disproportionately by refining rather than crude production. Refiners are expected to grow earnings 464%, versus 117% for integrated companies and 63% for producers.
While the U.S./Iran conflict creates headline volatility, oil is getting through the strait, but diesel is not. This is benefiting refiner margins, while drillers would rather rake in today’s profits after years of troubled earnings over ramping up production. The effects extend beyond Energy. Elevated diesel and refined-product costs can raise expenses across freight, agriculture, staples, materials and transportation, all potentially contributing to higher consumer prices. Still, this suggests the current extraordinarily strong earnings momentum, albeit less durable than the AI theme, should slow in 2027.
- Consumer: paying more does not always mean buying more. Chase card spending averaged 4% y/y in 3Q, but company results suggest nominal growth may be masking softer volumes. Costco produced 9% comparable-sales growth with traffic up only 3%, while General Mills had roughly flat organic sales and an 11% decline in adjusted operating profit in constant currency, with management describing substantial stress among middle- and lower-income consumers. Consumers remained strong enough thanks to upper-income households, supported by rising stock prices and real estate, but higher prices and potentially higher-for-longer rates create a clear downside risk to spending and credit quality.
Strong earnings could continue to support a constructive outlook for equities, but the debate is shifting from how much companies will spend to what that spending will earn. Valuations have already corrected by 16% as investors have become more cautious about future earnings and expanding capital budgets.5 AI demand remains robust and monetization is becoming more visible, but the ability to fund investment is not the same as earning an adequate return on it. Consensus expectations of 16% to 17% earnings growth in 2027 and 2028 could provide further support if growth broadens beyond Technology.
Market leadership has rotated across AI “picks and shovels,” hyperscalers, semiconductors, tech-adjacent sectors and defensives. That rotation argues against treating the AI story as the only earnings story. Future profits need to justify today’s spending, but a broader earnings recovery could give equities more than one source of support.