Key points:

• Expectations are already high. S&P 500 earnings are expected to rise around 29.5% year-on-year, and analysts have actually raised estimates going into the reporting season. A routine earnings beat may therefore carry less weight than usual.
• Valuations face a tougher test when bond yields are high. The US 10-year Treasury yield has recently traded above 5.3%, raising the return hurdle for equities and putting greater pressure on stocks where a lot of future growth is already priced in.
• Earnings quality and balance sheets matter more. Investors may increasingly distinguish between companies that can turn growth into cash and finance the next phase of investment internally, and those that need ever more capital to keep the growth story alive.

What is different about this earnings season?

1. Expectations are unusually high

The setup for Q3 earnings is strong, but that is precisely what makes this season more demanding.

FactSet expects S&P 500 earnings to rise 29.5% year-on-year, alongside revenue growth of 12.3%. More unusually, analysts raised Q3 earnings estimates by 1.4% during the quarter. Over the previous five years, estimates have typically been cut by around 2.2% over the course of a quarter. Of the 116 S&P 500 companies that have issued Q3 EPS guidance, 72 have issued positive guidance, well above historical averages.

Companies are therefore heading into earnings with much less of the usual expectations cushion. Instead of asking simply whether a company beat consensus, investors may need to ask whether the result was better than an already optimistic market expected.

That distinction matters because an earnings beat is only a surprise if investors were not already positioned for it.

• Positive signal: Revenue and earnings beat expectations, forward guidance rises and analysts continue upgrading future earnings estimates.
• Negative signal: The company beats current-quarter EPS, but guidance or future earnings estimates fail to move higher.

2. Valuation: how much good news is already priced in?

Expectations become even more important when combined with valuation.

A company trading at a modest valuation may only need to deliver solid results. A stock priced for exceptional growth may need to beat earnings, raise guidance and convince investors that elevated growth can continue for years.

This is particularly relevant after strong gains across parts of technology and AI. A company can deliver excellent operational results and still produce a disappointing share-price reaction if investors had already priced in something better.

The bond market raises that hurdle further. The US 10-year Treasury yield recently reached around 5.35%, its highest level in roughly 24 years. Investors can therefore earn a substantially higher return from relatively low-risk government bonds, while the present value of companies' distant future cash flows falls as discount rates rise.

This is why a good company and a good stock are not always the same thing. Investors need to consider not just how quickly earnings are growing, but how much they are being asked to pay for that growth.

• Positive signal: Earnings and guidance rise fast enough to support the valuation, while forward estimates keep moving higher.
• Negative signal: The business continues growing, but earnings revisions stall while the valuation still assumes exceptional future growth.

3. Earnings quality and balance sheets: can the growth keep paying for itself?

The third test is the quality of the growth itself.

Headline EPS can improve for many reasons, including cost cuts, lower taxes, share buybacks and one-off gains. A stronger earnings result would normally combine revenue growth, resilient margins, healthy operating cash flow and good conversion of accounting profits into free cash flow.

Higher bond yields make this balance-sheet test more important. Companies that generate plenty of cash can continue investing even when financing becomes more expensive. Businesses dependent on debt markets or repeated equity issuance face a much tougher equation.

This is also where cost of capital enters the discussion. It is not enough for a company to spend heavily and generate growth; the returns earned on that investment need to justify the increasingly expensive capital being committed.

Investors may therefore want to pay more attention to free cash flow, leverage, interest expense, refinancing needs, capex and returns on incremental investment alongside the headline EPS number.

• Positive signal: Revenue, earnings and cash flow rise together, while growth can largely be funded internally and returns on investment remain healthy.
• Negative signal: EPS beats, but cash flow deteriorates, debt rises or capex needs to keep increasing much faster than the revenues it generates.

These three tests apply across the market, but they show up very differently by sector. For AI, the debate is increasingly about monetisation and returns on investment. For banks, it is about whether higher yields remain a benefit or start becoming a problem. For consumer companies, the question is whether real demand can continue supporting both earnings and valuations.

What would be a positive or negative signal by sector?

AI hyperscalers: monetisation has to catch up with capex

For Microsoft, Alphabet, Amazon and Meta, the ability to finance AI investment is not the main concern. Their substantial cash generation gives them much greater flexibility than most companies.

The more interesting question is return on capex.

The market already expects enormous spending on data centres, chips and AI infrastructure, so another increase in capex is not necessarily bullish in itself. Investors increasingly need evidence that the spending is translating into faster cloud growth, incremental AI revenues, stronger customer commitments and eventually greater cash generation.

This represents an important shift in the AI narrative. The first phase rewarded companies for announcing investment. The next phase may demand proof that those investments are producing sufficiently attractive economic returns.

• Positive signal: AI and cloud revenues accelerate alongside capex, utilisation rises and companies provide more tangible evidence of customer monetisation.
• Negative signal: Capex forecasts rise again, but revenue benefits remain difficult to quantify and free cash flow comes under greater pressure.

Valuation raises the bar further. The more future AI success already embedded in the share price, the less patience investors may have for a long gap between spending and monetisation.

Software: AI needs to become revenue, not just a feature

Software deserves a separate test from the companies building AI infrastructure.

For software businesses, AI has the potential to create new products, improve productivity and support higher pricing. But investors increasingly need proof that those benefits are translating into financial results.

The central question is whether AI becomes genuinely incremental revenue, or simply another capability customers expect to receive within an existing subscription.

Metrics such as AI-related upselling, net revenue retention, bookings, remaining performance obligations, customer expansion, operating margins and free cash flow could therefore become particularly important.

There is also a longer-term competitive question. AI can make existing software companies more productive, but it can also make it easier and cheaper to build competing products. Strong platforms with distribution, proprietary data and deep customer relationships may benefit, while less differentiated software could face greater pressure.

Higher yields add another challenge because many software stocks derive a large part of their valuation from profits expected well into the future.

• Positive signal: AI drives new paid usage or upselling, bookings and retention remain healthy, and productivity gains improve margins and cash flow.
• Negative signal: AI adoption looks impressive but customers are reluctant to pay more for it, while bookings, retention or forward growth weaken.

For high-multiple software stocks, merely maintaining current growth may no longer be enough. Investors may need either faster growth, improving margins or stronger cash generation to offset the valuation pressure from higher yields.

Memory: watch what companies do with the boom

Memory offers a different challenge. Strong AI demand, tight HBM supply and improving pricing can support powerful earnings growth, but memory remains an inherently cyclical industry.

The key risk is therefore the industry response to today's strong demand.

High prices and attractive margins encourage manufacturers to add capacity. If supply eventually grows faster than demand, today's shortage can become tomorrow's oversupply.

More capacity is not automatically bearish, particularly if it is supported by contracted customer demand. The more important issue is whether capacity additions remain disciplined relative to the visibility manufacturers have on future orders.

• Positive signal: Pricing remains firm, inventories stay controlled and additional capacity is supported by long-term commitments from customers.
• Negative signal: Producers accelerate capacity expansion well ahead of visible demand, inventories start rebuilding or pricing momentum weakens.

Valuation also needs to be treated differently for cyclical companies. A low P/E ratio does not necessarily mean a memory stock is cheap; sometimes it simply tells investors that current earnings are unusually high. Looking at normalised earnings across the cycle can therefore be more useful than relying only on trailing multiples.

AI infrastructure, networking and components: demand must convert into cash

For networking, optical, semiconductor-component and other AI supply-chain companies, strong demand is only the first part of the story.

Investors also need to see orders converting into revenue and revenue converting into cash. Customer concentration, working capital, inventories and receivables become useful ways to test whether reported demand is translating into sustainable economics.

This group can also be more exposed to financing conditions than the hyperscalers themselves. Companies expanding production rapidly may need substantial new capital just as borrowing costs are rising.

Physical constraints add another layer of risk. Power availability, data-centre construction and other infrastructure bottlenecks can delay deployment even when underlying appetite for AI computing remains extremely strong.

• Positive signal: Orders convert into revenue and cash, margins remain healthy and expansion is funded largely from internal cash generation.
• Negative signal: Inventories or receivables rise much faster than sales, customer concentration increases or expansion becomes increasingly debt-funded.

Here, a strong AI story does not automatically equal a strong investment case. Balance-sheet quality and execution become much more important as the industry moves from ordering equipment to deploying it at scale.

Banks: when higher yields stop being good news

Banks provide one of the clearest tests of the recent bond sell-off.

Higher rates can initially be positive because banks can earn more on their assets and potentially widen the spread between lending rates and deposit costs. But the relationship changes when long-term yields rise too far or too quickly.

Funding costs can rise, loan demand may weaken and borrowers can find refinancing increasingly difficult. Banks can also suffer mark-to-market pressure on securities portfolios, while prolonged high rates may eventually lead to higher defaults and credit provisions.

The question this earnings season is therefore not simply whether higher yields are helping net interest margins. It is whether the bond sell-off is starting to affect the broader credit cycle.

Investors should pay particular attention to net interest income, deposit costs, loan growth, delinquencies, provisions and commercial real estate exposure. Trading and investment-banking revenues may provide an offset if market activity remains strong.

• Positive signal: Net interest income remains resilient, deposit costs stabilise, loan growth holds and credit quality remains healthy.
• Negative signal: Funding costs continue rising, loan demand weakens or banks start increasing provisions in anticipation of greater credit stress.

Valuation again matters. Banks that have already de-rated may only need evidence that conditions are stabilising, while stocks still priced for strong profitability may require a much cleaner earnings outcome.

Consumer: is growth coming from price or genuine demand?

For consumer-facing companies, headline revenue growth may hide an important distinction.

Over the past several years, companies have often maintained sales growth by raising prices even when customers purchased fewer units. That strategy becomes harder as borrowing costs remain high and households become more price-sensitive.

Investors should therefore pay close attention to price versus volume. A business capable of maintaining both volumes and pricing power has a much healthier earnings profile than one where price increases are masking weakening underlying demand.

Margin performance will also matter. If companies need heavier promotions or discounts to maintain volumes, headline sales may hold up while profitability deteriorates.

• Positive signal: Volumes stabilise or improve, companies retain pricing power and margins hold without aggressive discounting.
• Negative signal: Revenue growth is increasingly price-driven, volumes deteriorate or management highlights more pronounced consumer stress.

Valuation can amplify these reactions. When investors pay a premium for a company's brand strength or pricing power, even modest evidence of customers trading down can challenge the investment case.

Key earnings dates to watch

*Dates marked with an asterisk remain subject to company confirmation.

Bottom line: a beat is only the starting point

The headline earnings numbers may still be very strong. But with expectations elevated, valuations demanding and Treasury yields near multi-decade highs, the investment hurdle has become higher as well.

This season should therefore be less about simply identifying who beat consensus by the largest amount. The stronger businesses may be those that can deliver growth above expectations, convert that growth into cash, fund future investment from a strong balance sheet and still offer enough earnings upside to justify their valuation.

In other words, investors may increasingly reward profitable, self-funded growth at a defensible valuation rather than growth at any price. That could also mean greater dispersion between companies and index constituents.