MW This earnings number is the only one that matters as a gut check on Big Tech's AI spending

By Jurica Dujmovic

S&P 500 index-fund investors are locked into a big bet that AI capital expenditures will pay off

The one place hype cannot hide? The cash-flow statement.

With four of the "Magnificent Seven" tech companies reporting earnings this week, artificial intelligence is about to show up in the one place hype cannot hide: the cash-flow statement.

The question for investors is no longer whether AI is transformative. The question is whether the largest capital-spending cycle in modern tech history is starting to produce measurable returns before the market loses patience.

Microsoft $(MSFT)$, Meta Platforms (META) and Tesla $(TSLA)$ report on Wednesday; Apple $(AAPL)$ follows on Thursday. Alphabet $(GOOG)$ $(GOOGL)$ and Amazon.com (AMZN) close out the wave in early February. Together, these companies are running what amounts to a live stress test of the AI investment thesis, and the results will ripple through investment portfolios.

The S&P 500's AI risk

Investors in plain-vanilla S&P 500 SPX index funds are actually making a concentrated bet on AI infrastructure. Four Big Tech companies - Amazon.com, Microsoft, Alphabet and Meta Platforms - account for 25.6% of total index capital expenditures. AI-related tech spending across the S&P 500 has exceeded $1.25 trillion over the past 12 months, with the Magnificent Seven responsible for roughly 28% of that total.

This concentration is risky - and not in a conventional way. Those four Magnificent Seven companies command almost 20% of the S&P 500's market capitalization; adding the other three - Nvidia (NVDA), Apple and Tesla - brings the Magnificent Seven's collective weighting to almost 35% of the benchmark index.

With so much of the S&P 500 riding on a single theme, the performance of the index depends on that theme delivering.

If AI spending converts into durable revenue and margin expansion, the S&P 500 will benefit. If it does not, the weight of disappointment will be difficult to diversify away.

The spending itself is no longer debatable. Data-center equipment and infrastructure investment reached roughly $290 billion in 2024, with projections suggesting the market could approach $1 trillion by 2030. What remains uncertain is the timeline for payback, and whether capacity constraints like power availability and cooling infrastructure will delay returns even as dollars continue flowing out the door.

Capex reality check

The table below shows where six of the Magnificent Seven companies stand on the most basic measure of AI investment pressure: How much of their operating cash flow is being consumed by capital spending.

   Microsoft                             $45.1B   $19.4B   $25.7B  43.0% 
   Meta                                  $30.0B   $19.4B   $10.6B  64.6% 
   Apple                                 $111.5B  $12.7B   -       11.4% 
   Alphabet                              $151.4B  $77.9B   $73.6B  51.4% 
   Amazon                                $130.7B  $115.9B  $14.8B  88.7% 
   Tesla                                 $6.2B    $2.2B    $4.0B   36.0% 
   Data: Trailing 12-months, annualized 

Amazon stands out immediately: Close to 90% of its operating cash flow is being reinvested into property and equipment - the majority of it tied to AWS and the broader data-center buildout. That leaves just $14.8 billion in trailing free cash flow, a razor-thin margin for a company of its scale. If AWS growth or AI-service adoption does not accelerate visibly, that ratio becomes difficult to defend.

Meta sits at the other end of intensity, with capital spending running above 64% of operating cash. The company has already guided $70 billion to $72 billion in total 2025 capex, explicitly tying much of the increase to AI and infrastructure. Advertising revenue remains the engine funding this buildout, which makes any weakening in ad pricing or engagement a structural problem, not just a cyclical one.

Apple provides the counterpoint: Its capex intensity remains below 12%, a reflection of its device-centric AI strategy. Apple does not need to build hyperscale data centers to monetize AI. Instead, it relies on on-device processing and potential partnerships to deliver AI features without matching the infrastructure arms race. That lower intensity is an advantage if AI proves to be a demand catalyst for hardware upgrades and services growth. It is a vulnerability if cloud-based AI models dominate the competitive landscape.

Microsoft and Alphabet occupy the middle ground. Microsoft's capex has been rising, but the company highlighted in its most recent quarter that free-cash-flow growth has held up in part because of a higher mix of finance leases, which spread the cash impact over time. Alphabet has maintained trailing 12-month free cash flow above $73 billion even as it guided toward $91 billion to $93 billion in 2025 capex, suggesting the business can still generate substantial cash while investing aggressively.

Tesla is the outlier in this group: AI spending matters for autonomy and manufacturing automation, but investors still anchor primarily on automotive margins and production efficiency. The company posted positive free cash flow in its most recent quarter despite ongoing capex, but the AI narrative here is secondary to near-term profitability in the core business.

Read: I refused to invest in Tesla for years - but now's the time to bet on Elon Musk

What to listen for on earnings calls

The financial statements provide the baseline. The earnings calls will reveal whether management still believes the payback clock is on schedule.

Microsoft (reports Jan. 28): The critical question is whether Azure demand is genuinely outrunning capacity, or whether growth is decelerating as utilization normalizes. With Copilot, investors should listen for specifics on commercial seat adoption and whether enterprises are renewing and expanding deployments.

Meta (Jan. 28): Meta's higher capex intensity makes advertising performance the central issue. If AI-driven improvements in ad targeting and engagement are translating into pricing power, the buildout becomes easier to justify.

Tesla (Jan. 28): AI and robotics investments will be discussed, but the market will focus on automotive cash generation and whether the company is tightening or loosening capital-allocation discipline.

Apple (Jan. 29): The iPhone cycle and services growth are the real tests. AI should show up as a feature that drives upgrade intent or increases services attachment, not as a standalone revenue line.

Alphabet (Feb. 4): Search monetization is the key. AI integration into search has been widely discussed, but the impact on ad load, click-through rates and cost per click remains opaque.

Amazon (Feb. 5): With capex consuming almost 90% of operating cash flow, AWS performance is nonnegotiable. Investors need to hear that incremental capacity is converting into billable services quickly, and that utilization rates are rising rather than stalling.

Suppliers flash an early warning

Magnificent Seven earnings are only a part of the story; supplier results tell it first.

Broadcom $(AVGO)$ reported in its most recent quarter that AI contributed more than half of its semiconductor revenues, with the company securing over $10 billion in AI rack orders. That kind of visibility suggests hyperscaler demand is real and sustained, not speculative.

If networking vendors such as Arista Networks (ANET) begin reporting elongated lead times and supply bottlenecks, or Marvell Technology $(MRVL)$ signals softening data-center demand, that would suggest either capacity saturation or weakening demand - both of which would undercut the AI investment narrative before it shows up in any Magnificent Seven guidance.

Power and thermal infrastructure companies like Vertiv Holdings $(VRT)$, which recently reported a $9.5 billion order backlog and strong organic bookings, and Eaton $(ETN)$, with backlog and book-to-bill expansion in its electrical segments, are worth monitoring. Bookings growth and backlog conversion at these firms tend to lead data-center revenue by one or two quarters.

Semiconductor-equipment providers - Applied Materials (AMAT), Lam Research $(LRCX)$ and KLA $(KLAC)$ - offer another signal. If advanced packaging demand or AI-specific wafer-fab equipment orders begin to decelerate, that could signal chip makers are pulling back on future production capacity and indicate slower fab capex - undercutting the broader AI investment narrative before it shows up in Magnificent Seven guidance.

Enterprise software companies such as ServiceNow (NOW), whose platform automates mission-critical workflows and increasingly integrates AI and automation layers, provide a different angle. If net retention and renewal rates - such as the roughly 98% renewal rate ServiceNow reports - remain high or expand as customers adopt AI features, that suggests enterprises are expanding usage rather than treating these tools as discretionary spend.

The AI clock is ticking

The optimistic case is straightforward: If AI drives productivity improvements that show up as revenue-per-employee gains, operating leverage and margin expansion, and if hyperscaler capex converts into high-margin cloud- and AI-services revenue, the spending will have been justified.

The pessimistic case is equally clear: If capex intensity remains elevated for multiple quarters without corresponding revenue acceleration or free-cash-flow improvement, and if depreciation from prior AI buildouts begins compressing margins without offsetting productivity gains, the market will reprice the investment thesis.

The Magnificent Seven's earnings will not settle the question definitively. But they will show whether the AI payback clock is still running on time - or starting to slip.

More: With Tesla earnings on deck, investors reveal their most pressing questions for Elon Musk

Also read: Is Meta's stock a steal? Why a new bull sees big gains ahead.

-Jurica Dujmovic

(MORE TO FOLLOW) Dow Jones Newswires

January 27, 2026 07:50 ET (12:50 GMT)

Copyright (c) 2026 Dow Jones & Company, Inc.