This week’s high-stakes tech earnings are a test of AI’s “lift vs. load”
The stock market Super Bowl… Microsoft, Amazon, Alphabet, and Meta are set to report on Wednesday, followed by Apple on Thursday. While traders are pricing in a 99.5% chance that the Fed will keep its target rate steady on Wednesday, there’s a lot more uncertainty around tech results. AI enthusiasm largely carried the market’s rally last year, but this year’s been marked by worries over AI splurges. The earnings stakes are high, especially given how much weight these tech names hold:
- Mag 7 minus 493: If you exclude the Magnificent Seven from the S&P 500, the index’s market value is down since the start of the Iran conflict in late February.
- AI spending is bolstering GDP, too. Makes sense when Alphabet, Microsoft, Meta, and Amazon alone are expected to collectively shell out nearly $700 billion (about the GDP of Belgium) on AI this year.
Business AI adoption crossed 50% for the first time in March, according to the Ramp AI index. Demand is clearly there, but the results of this week’s earnings hinge on one question: what’s AI’s lift vs. load on the bottom line?
Lift: How much is AI driving growth?Areas of focus for this:
- Ads: Meta has cited AI-driven improvements (think: better targeting, recommendations) for strong growth in advertising, which still makes up 97% of its revenue. The social giant is projected to unseat Google as the largest digital-ad company this year.
- Cloud: AI workloads are driving massive demand for computing titans like Amazon, Microsoft, and Google. That’s translated to extremely strong growth for cloud units.
- AI infusions into core products have led to bundle price increases. Think: Google’s Gemini-fication of Google Workspace and Microsoft’s Copilot integrations.
Load: How much is AI weighing on profit? The stronger the profit cushion from core businesses, the more lenient investors will be with AI spend forecasts. This week, there will likely be heightened sensitivity to any slowdowns in earnings growth. Markets are asking whether lift + load will be enough to justify hundreds of billions in CapEx.
- Load could lessen if AI costs come down: Gartner predicts that by 2030, inference on an LLM with 1T parameters will cost GenAI providers over 90% less than in 2025. Sounds great, except the firm suggested enterprises won’t see total cost reductions because demand for more intensive applications (like agentic AI) will lead token consumption to rise faster than costs fall.
The bottom line:
The “lift vs. load” equation applies to all businesses… Higher lift + lower load = larger AI investments justified (and likely).Ramp data shows businesses' monthly AI spend quadrupled from February 2025 to February 2026. We’re still in the early innings, but continued long-term investment will depend on 1) measurable ROI, like AI-driven revenue lifts or savings, and 2) a growing profit cushion (whether AI-related or not).
For more stories like this, subscribe to the Leading Indicators newsletter.
