The S&P 500 is staring at a big number for 2026. Earnings are expected to pop more than 26% year over year. On paper, that sounds great. In practice, it ties the whole market to a handful of companies spending staggering amounts on AI infrastructure.

This piece breaks down what that 26% hurdle actually means, how the AI capex wave could help or hurt margins, and what to watch in the next few quarters. No hype here. Just the moving parts that decide whether the math adds up or not.

If you only have a few minutes, start with the quick answer. Then dive into the questions around margins, concentration, rates, and what could go sideways first.

Yes, the 26% profit hurdle can survive, but it needs a near-perfect setup. Consensus has 2026 EPS in the $340–$350 range, underpinned by AI demand and hyperscaler spending. The catch is the bill. Hyperscaler capex near record levels has to translate into durable, monetizable revenue, not just headlines. If margins slip or AI monetization lags, that EPS path gets tight fast.

• Consensus calls for 26%+ EPS growth in 2026 per LSEG IBES, a very high bar Reuters (reported on Investing.com).
• JPMorgan now sees S&P 500 EPS at $350 and a 7,800 year-end target for 2026 Investing.com.
• Wells Fargo lifted its 2026 EPS view to $340 with a 7,950 target, citing hyperscaler tailwinds Reuters (reported on Investing.com).
• Five tech giants plan roughly $730 billion of 2026 capex combined, mostly AI buildout, per JPMorgan estimates Reuters (reported on Investing.com).

What exactly is the 26% earnings hurdle, and why is it tricky?

It is simply the gap between where S&P 500 earnings finished the prior year and where analysts think they land in 2026. LSEG IBES shows consensus calling for more than 26% growth. You can think of it like a high jump bar. Markets already ran the approach. Now they need clean execution to clear it Reuters (reported on Investing.com).

The tricky part is concentration. A large slice of index-level EPS growth is expected to come from the usual mega-cap suspects tied to AI, cloud, and digital ads. If two or three of them miss on margins or push monetization out by a couple of quarters, the whole index math shifts. The rest of the index can help, but cyclicals and defensives usually do not sprint like software or semis.

Another wrinkle is timing. Street models tend to assume a pretty clean quarterly ramp. Real businesses are lumpy. A delayed product cycle or an enterprise budget freeze can put a dent in a single quarter that ripples across a full-year number.

Will AI capex lift or crush margins in 2026?

Both forces are in play. On the lift side, AI spend can create new workloads for cloud providers, pull in fresh software revenue, and boost high-margin services layered on top. On the crush side, the infrastructure bill is enormous. JPMorgan pegs combined 2026 capex for five hyperscalers around $730 billion, which will flow into depreciation, energy, and maintenance lines for years Reuters (reported on Investing.com).

Short term, heavy capex usually depresses free cash flow. Income statements will also feel it once assets go live and start depreciating. The key offset is monetization. If AI services command premium pricing and sticky adoption, gross margin can hold up even as D&A rises. If pricing gets competitive and inference costs do not fall fast enough, you get margin squeeze.

One more practical point. Energy is not a footnote. Power availability and pricing influence both capex timing and operating expenses. A tight grid or delayed data center hookups can push revenue recognition out while fixed costs keep rolling in.

How much of the S&P 500’s EPS depends on the Big Seven?

There is no single official number, and weights shift with prices. Still, it is fair to say a heavy chunk of the growth delta sits with the mega-cap platform companies. That is why consensus targets lean on them. JPMorgan’s raised EPS forecast to $350 and 7,800 price target, along with Wells Fargo’s $340 and 7,950, both highlight hyperscaler capex as a positive driver if it converts to revenue efficiently Investing.com Reuters (reported on Investing.com).

Think of the index like a pyramid. The top layer, a small group of huge companies, moves the headline EPS the most. The middle layer provides stability. The base is wide but slower. That mix can work beautifully when the top layer is hitting on all cylinders. It also makes the system sensitive to disappointments up top.

Group Revenue visibility Margin trajectory Capex intensity Sensitivity to rates Valuation risk AI leaders and hyperscalers High, multi-year contracts Mixed, depends on AI pricing vs costs Very high Moderate, via discount rates and demand High, expectations loaded Broader tech and suppliers Medium, tied to product cycles Improving if volume ramps High but more variable Moderate Medium, selective Non-tech cyclicals Medium to low, macro driven Tight, depends on input costs Low to medium High, via financing and demand Low to medium Defensives High, steady demand Stable but capped Low Lower, but rates hit valuation Low

If the AI leaders meet or beat, the whole index gets pulled higher. If they blink, even strong showings from other sectors may not fully backfill the gap.

What needs to go right for $340–$350 EPS to stick?

Framing it as a checklist helps. JPMorgan’s $350 EPS call and Wells Fargo’s $340 assume AI demand converts into paid workloads, cloud margins hold, and the macro backdrop stays roughly constructive through 2026 Investing.com Reuters (reported on Investing.com).

• Paid AI adoption: Enterprise pilots must become paid seats and usage-based revenue, not endless proofs of concept.
• Cloud margins: Infrastructure gross margin needs to offset growing depreciation and power costs.
• Semiconductor supply: Enough advanced chips, memory, and networking gear to support deployment timelines.
• Power and permitting: Data centers get energization on schedule to avoid revenue slippage.
• Macro stability: Rates drift lower or at least do not spike, labor markets cool without breaking demand.
• Credit spreads: No sudden tightening that crimps buybacks or capex plans beyond AI.

If most of that happens, $340–$350 is achievable. If two or three fail at once, the EPS ladder gets a few rungs shorter pretty quickly.

Where could the consensus stumble first?

The softest spot is unit economics. If inference costs per user fall slower than expected, or if vendors subsidize usage to win share, gross margin takes the hit. Cloud providers can hide that for a bit in blended margins, but not forever.

Another early warning sign would be slower than expected conversion of pilots to production. Plenty of companies love announcing AI initiatives. Fewer have the budget and change management muscle to deploy at scale. Delays are common, even in good markets.

Pro tip: Separate reported “AI revenue” from revenue that is actually profitable at scale. Some lines grow fast because they are underpriced relative to compute cost. That is growth, but not necessarily earnings.

Finally, energy. If power is late or more expensive than planned, you might see revised deployment timelines, which pushes revenue right while opex and D&A stay put. That is classic negative operating leverage.

How do rates, inflation, and buybacks change the math?

Rates affect both sides. On valuation, lower discount rates support higher multiples. On earnings, lower rates cut interest expense and free up cash for buybacks. If the Fed stays cautious and cuts are shallow, the direct EPS help is smaller, though still positive over time as legacy debt is refinanced.

Inflation now matters more in services and wages than in goods. For many S&P 500 names, that shows up as pressure on operating expenses rather than cost of goods. If wage growth cools a touch while demand holds, operating margin has room. If wage growth stays hot and pricing power fades, margins compress.

Buybacks are the quiet lever. Even modest net buybacks can add a point or two to EPS growth. But if credit spreads widen or regulators get more vocal, that lever weakens at the wrong time.

What should investors watch in Q3 and Q4 2026 results?

A few simple tells go a long way. You do not need a 200-line model to see if the 26% track is intact.

• Cloud gross margin and capex cadence. Are vendors guiding to steady or improving margins while capex stays elevated?
• AI monetization granularity. Management should split out usage that is monetized at or above cost, not just total AI engagement.
• Data center energization dates. Look for specific timelines, not hand-waving. Slips tend to cluster by region.
• Enterprise conversion metrics. Pilots to production ratios, contract lengths, and minimum commitments.
• Working capital trends. Inventory for AI hardware and related networking should normalize, not balloon.
• Share count. Quiet buyback acceleration can absorb small earnings misses at the index level.

Put differently, if capex is still roaring and the revenue line is not keeping up, that is your hangover signal. If both scale together and margins hold, consensus looks reasonable.

Common Mistakes

1. Assuming all AI revenue is high margin. Some workloads are subsidized or priced for share. Ask about contribution margin, not just top line.
2. Ignoring depreciation and power. The bill shows up after the ribbon cutting. Rising D&A and energy can nick margins for several years.
3. Overweighting one vendor narrative. Spread bets across infrastructure, platforms, and applications. Single-vendor risk is real in a fast-moving stack.
4. Forgetting the middle of the index. Industrials, financials, and healthcare can surprise on efficiency and pricing. Do not assume they are dead weight.
5. Reading guidance too literally. Management sandbags or overreaches. Track trailing execution against guidance quality, not just the headline raise or cut.

If you want more ongoing reads like this, Crypto Daily covers macro, digital assets, and the crossover where AI spending meets market structure. Visit Crypto Daily for weekly breakdowns.

Frequently Asked Questions

What happens if hyperscalers cut capex midyear?

Short term, free cash flow improves. But if cuts reflect weaker demand or power constraints, growth expectations reset too. Markets would likely cheer the cash in the moment and then rerate growth down if guidance confirms softer workloads.

Can energy constraints alone derail the EPS path?

They can delay it. If data centers cannot energize on time, revenue recognition slips to later quarters while fixed costs keep accruing. A few months here and there is manageable. A year of delays across multiple regions would force consensus down.

Which non-tech sectors could offset a tech miss?

Financials can benefit from stable credit and fee growth, healthcare from product cycles and services, and industrials from automation and reshoring. None have the same torque as mega-cap tech, but together they can cushion the blow if tech only modestly underdelivers.

Do semiconductor and equipment vendors offset hyperscaler margin risk?

They help the ecosystem, yes. Strong orders at chipmakers and gear suppliers support earnings breadth. But the index relies more on the platform companies’ margins. Healthy suppliers cannot fully compensate if platform monetization disappoints.

How do accounting rules treat AI training costs?

Most model training runs through operating expenses or cost of revenues rather than being capitalized. The big capitalized pieces are hardware, data centers, and certain software development. That is why D&A and power matter so much to the margin story.

What if the economy weakens but inflation stays sticky?

That is the awkward mix. Slower demand cuts revenue growth while sticky wages and energy keep costs high. In that scenario, the 26% EPS hurdle becomes much tougher and buybacks do less to offset it.

Is the $340–$350 EPS range too bullish?

It is ambitious, not impossible. It assumes AI ramps monetize fairly cleanly, macro holds, and execution is steady. If those pieces line up, it works. If not, expect revisions, especially in the first half as guidance tightens.

Disclaimer: This article is provided for informational purposes only. It is not offered or intended to be used as legal, tax, investment, financial, or other advice.