Does Heavy Capex Predict Future Stock Returns? Capital Expenditure Analysis in Python

May 16, 2026

What’s the question?

Capital expenditure (capex) represents a company’s investment in future productive capacity — factories, data centers, equipment, technology. When a company dramatically increases capex, it is making a bet on future growth. The question is whether the stock market rewards or punishes this bet. Does heavy capex spending precede strong stock performance, or does it signal overinvestment and diminishing returns? Capex intensity — defined as capital expenditures divided by revenue — normalizes spending across companies of different sizes, allowing direct comparison between a $717B revenue giant like AMZN and a $68B industrial firm like CAT.

The approach

Select 8 large-cap stocks across sectors. Pull 5 years of annual fundamentals including revenue, capital_expenditures, and free_cash_flow. Compute capex intensity (capex divided by revenue) and year-over-year capex growth for each period. Then examine the multi-year trajectories for AMZN and META — the two most aggressive capex spenders in recent years — to understand how their investment strategies have diverged.

import xfinlink as xfl
import pandas as pd

xfl.api_key = "YOUR_API_KEY"  # free at https://xfinlink.com/signup

# -- Configuration ----------------------------------------------------------
tickers = ["AAPL", "MSFT", "AMZN", "META", "XOM", "CAT", "UNH", "HD"]

# -- Fetch 5Y annual fundamentals ------------------------------------------
df = xfl.fundamentals(
    tickers,
    period_type="annual",
    fields=["revenue", "capital_expenditures", "free_cash_flow"],
    period="5y",
)

# -- Compute capex intensity and YoY growth --------------------------------
df = df.sort_values(["ticker", "period_end"])
df["capex_abs"] = df["capital_expenditures"].abs()
df["intensity"] = df["capex_abs"] / df["revenue"]
df["capex_yoy"] = df.groupby("ticker")["capex_abs"].pct_change()

# -- Latest annual period per ticker ---------------------------------------
latest = df.sort_values("period_end").groupby("ticker").tail(1).copy()
latest = latest.sort_values("intensity", ascending=False)

print("=== Capex Intensity: Who Is Investing the Most? (Latest Annual) ===")
header = (
    f"{'Ticker':6s}  {'Period':>12s}  {'Revenue':>12s}  {'Capex':>12s}  "
    f"{'Intensity':>10s}  {'Capex YoY':>10s}"
)
print(header)
print("-" * 65)

for _, r in latest.iterrows():
    rev_str = f"${r['revenue'] / 1e3:>8.0f}B"
    capex_str = f"${r['capex_abs'] / 1e3:>8.0f}B"
    intensity_str = f"{r['intensity']:>9.1%}"
    yoy_str = f"{r['capex_yoy']:>+9.1%}" if pd.notna(r["capex_yoy"]) else "      N/A"
    print(
        f"{r['ticker']:6s}  {str(r['period_end'])[:10]:>12s}  {rev_str:>12s}  "
        f"{capex_str:>12s}  {intensity_str:>10s}  {yoy_str:>10s}"
    )

# -- AMZN trajectory -------------------------------------------------------
print("\n=== AMZN Capex Trajectory ===")
amzn = df[df["ticker"] == "AMZN"].dropna(subset=["capex_abs"])
for _, r in amzn.iterrows():
    yoy_str = f"YoY={r['capex_yoy']:+.1%}" if pd.notna(r["capex_yoy"]) else "YoY=N/A"
    print(
        f"  {str(r['period_end'])[:10]}  "
        f"capex=${r['capex_abs'] / 1e3:.0f}B  "
        f"intensity={r['intensity']:.1%}  "
        f"{yoy_str}"
    )

# -- META trajectory -------------------------------------------------------
print("\n=== META Capex Trajectory ===")
meta = df[df["ticker"] == "META"].dropna(subset=["capex_abs"])
for _, r in meta.iterrows():
    yoy_str = f"YoY={r['capex_yoy']:+.1%}" if pd.notna(r["capex_yoy"]) else "YoY=N/A"
    print(
        f"  {str(r['period_end'])[:10]}  "
        f"capex=${r['capex_abs'] / 1e3:.0f}B  "
        f"intensity={r['intensity']:.1%}  "
        f"{yoy_str}"
    )

Output:

=== Capex Intensity: Who Is Investing the Most? (Latest Annual) ===
Ticker        Period     Revenue       Capex   Intensity   Capex YoY
-----------------------------------------------------------------
META      2025-12-31  $     201B  $      70B       34.7%     +87.1%
MSFT      2025-06-30  $     282B  $      65B       22.9%     +45.1%
AMZN      2025-12-31  $     717B  $     132B       18.4%     +58.8%
XOM       2025-12-31  $     332B  $      28B        8.5%     +16.7%
CAT       2025-12-31  $      68B  $       3B        4.2%     +41.9%
AAPL      2025-09-27  $     416B  $      13B        3.1%     +34.6%
HD        2026-02-01  $     165B  $       4B        2.2%      +5.6%
UNH       2025-12-31  $     448B  $       4B        0.8%      +3.5%

=== AMZN Capex Trajectory ===
  2022-12-31  capex=$64B  intensity=12.4%  YoY=+4.2%
  2023-12-31  capex=$53B  intensity=9.2%  YoY=-17.2%
  2024-12-31  capex=$83B  intensity=13.0%  YoY=+57.4%
  2025-12-31  capex=$132B  intensity=18.4%  YoY=+58.8%

=== META Capex Trajectory ===
  2022-12-31  capex=$31B  intensity=27.0%  YoY=+68.2%
  2023-12-31  capex=$27B  intensity=20.0%  YoY=-14.0%
  2024-12-31  capex=$37B  intensity=22.6%  YoY=+37.8%
  2025-12-31  capex=$70B  intensity=34.7%  YoY=+87.1%

What this tells us

META is the heaviest investor relative to revenue at 34.7% capex intensity — nearly doubling capex from $37B to $70B in a single year (+87% YoY), driven by AI infrastructure buildout. MSFT and AMZN follow at 22.9% and 18.4% respectively. AMZN’s trajectory is equally striking: capex nearly tripled from $53B to $132B in two years, pushing intensity from 9.2% to 18.4%. META’s trajectory shows a clear inflection: after pulling back from peak metaverse spending (27% intensity in 2022) to $27B in 2023, capex re-accelerated sharply — $37B in 2024, then $70B in 2025 — as the company pivoted its capital strategy toward AI data centers. Both AMZN and META are now in aggressive capex expansion mode, with META’s acceleration even steeper. UNH at 0.8% intensity confirms that health insurance is essentially a capital-free business model.

So what?

Capex intensity is not inherently good or bad for shareholders — the question is whether the investment earns above its cost of capital. META and AMZN together spent $202B on capex in 2025 — an unprecedented combined bet on AI infrastructure. META’s 87% YoY capex surge and AMZN’s 59% growth are wagers that AI compute capacity will generate returns exceeding their roughly 12% cost of equity. If it does, these stocks are undervalued at current levels because the market is discounting future earnings that have not yet materialized. If it does not, the excess capex becomes a drag on free cash flow and the stocks reprice lower. Monitor the ratio of capex intensity to revenue growth: when capex grows faster than revenue for multiple consecutive years, the company is either building future capacity (bullish) or overinvesting (bearish). The distinction depends on whether new capacity translates to incremental revenue.