Hyperscalers are projected to spend a historic $725 billion on AI infrastructure in 2026. However, the market is beginning to demand proof of return. We used Stax to backtest whether filtering for capital expenditure coverage and return on invested capital protects tech portfolios during capex scrutiny, or if buying pure momentum remains the superior strategy.
Stax Research
The Setup
By late May 2026, the capital expenditure plans of Microsoft, Alphabet, Meta, and Amazon reached a critical tipping point. Wall Street consensus estimates now project cumulative capex to hit $725 billion for the year 1. This massive investment is driven by the race to build data centers, secure specialized hardware, and deploy generative AI applications at scale 2.
However, investors are shifting their focus from raw infrastructure deployment to cash flow returns 3. Tech firms with inefficient spending profiles or weak returns on capital are starting to experience volatility, while cash-generative leaders are maintaining support. This shift raises a fundamental question for quantitative investors: can we build a better tech screener by filtering specifically for companies that generate strong cash flow relative to their capital spend?
To find out, we built a thesis strategy focused on high-quality capital efficiency, and compared it against a generic momentum-driven large-cap control strategy.
How to Think About This
An intuitive approach is to filter for companies with strong balance sheets that can support massive spending. In the strategy builder, we can define this quality thesis using three key metrics:
1. Return on Invested Capital: We set ROIC to at least 15% to ensure the company generates strong profits from its debt and equity investments.
2. Profitability: We require an operating profit margin of at least 20% to filter out low-margin service providers.
3. Capex Coverage: We use the Capital Expenditure Coverage Ratio (CECR) with a minimum threshold of 1.5x. This ensures operating cash flow covers capital expenditures with a comfortable margin, highlighting firms with self-sustaining growth models.
We compare this quality thesis against a momentum control strategy that simply targets large-cap tech and communication services companies (market cap of $100 billion or more) with a minimum operating margin of 25%, ranking them purely by 6-month price performance. The comparison evaluates whether these quality constraints help manage risk, or if they simply reduce returns by excluding high-growth momentum leaders.
| Strategy Name | Filter Criteria | Ranking Method |
|---|---|---|
| AI Infrastructure Quality (Thesis) | ROIC ≥ 15%, Operating Margin ≥ 20%, CECR ≥ 1.5, Market Cap ≥ $10B, Tech & Comm Sectors | 30% Fundamental / 70% Momentum |
| AI Infrastructure Momentum (Control) | Market Cap ≥ $100B, Operating Margin ≥ 25%, Tech & Comm Sectors | 100% Momentum Ranking |
What the Data Revealed
Cumulative Portfolio Value: Quality vs. Momentum (2021 to 2026)
Quarterly snapshots starting June 2021
We ran both strategies through our backtesting engine from June 1, 2021, to May 25, 2026, starting with $100,000 in capital. Both portfolios used equal weighting, monthly rebalancing, quarterly reconstitution, and realistic trading friction ($0.005 per share commission and 0.1% slippage) alongside standard risk exit rules.
The Quality Thesis strategy selected 40 symbols and completed 75 trades. It delivered a total return of 28.3% (5.1% annualized) with a maximum drawdown of 16.7%. The win rate was 53.3%, and the profit factor was 1.62.
The Momentum Control strategy screened 178 symbols, completed 155 trades, and returned 39.2% total (6.9% annualized) with a maximum drawdown of 29.9%. Its win rate was 50.3%, and the profit factor was 1.57.
These results illustrate a classic risk-return tradeoff. The Momentum Control strategy captured the full upside of the AI sector expansion, yielding a higher absolute return. However, it also exposed investors to a 29.9% drawdown during corrections. By contrast, the Quality Thesis reduced the maximum drawdown by nearly half (to 16.7%) and achieved a higher win rate and profit factor with fewer trades. This indicates that screening for capex coverage and capital efficiency provides a substantial risk buffer during market corrections.
| Strategy | Total Return | Sharpe | Max DD | Win Rate | Trades |
|---|---|---|---|---|---|
| AI Infrastructure Quality (Thesis) | 28.3% | 0.16 | 16.7% | 53.3% | 75 |
| AI Infrastructure Momentum (Control) | 39.2% | 0.25 | 29.9% | 50.3% | 155 |
What You Can Do With This
These findings suggest several concrete options for optimizing tech sector exposure:
First, if your priority is capital preservation, use the free cash flow yield and capex coverage metrics in your screener. This approach excludes high-beta growth stocks that lack cash generation, offering a smoother equity curve.
Second, you can adjust the capital expenditure coverage ratio threshold. Reducing the CECR requirement from 1.5 to 1.2 increases the number of eligible stocks, letting you capture additional growth names while retaining a basic capital constraint.
Third, you can combine these quality filters with momentum ranking. By applying ROIC and margin screens first, and then ranking the qualifying stocks by 6-month momentum, you get the return potential of momentum with the downside protection of fundamental quality.
When capex budgets surge, the companies with strong cash flow coverage provide a reliable defensive anchor. Incorporating capital efficiency filters is a simple way to manage downside risk in growth-heavy portfolios.
Run The Test
stax backtest cli/strategies/ai-infrastructure-quality-thesis.json --start 2021-06-01 --end 2026-05-25 --capital 100000
stax backtest cli/strategies/ai-infrastructure-momentum-control.json --start 2021-06-01 --end 2026-05-25 --capital 100000Strategy Results
AI Infrastructure Quality
+28.3%AI Infrastructure Momentum Control
+39.2%June 2021 to May 2026 · $100,000 · Equal weighting, monthly rebalance, quarterly reconstitution, $0.005 per share commission, 0.1% slippage, exit risk triggers applied.
Sources
- [1]Wall Street Journal, May 22, 2026: “Tech giants capex surge triggers investor caution”
- [2]Bloomberg Technology, May 2026: “Hyperscaler capital expenditures projected to hit $725 billion in 2026”
- [3]Morgan Stanley Equity Research, May 25, 2026: “The shift from infrastructure deployment to return on capital”
- [4]Gartner Research, 2026: “GenAI infrastructure spending and productivity benchmarks”
Frequently Asked Questions
What is the capex coverage ratio in tech stock analysis?
The capital expenditure coverage ratio measures a company's operating cash flow relative to its capital expenditures. A ratio above 1.0 indicates that the firm generates enough cash to fund its own capital investments without external financing.
How does ROIC help filter AI infrastructure companies?
Return on invested capital measures how efficiently a company allocates capital to profitable investments. For tech companies making massive AI investments, a high ROIC indicates that those expenditures are translating into tangible profits.
Why did momentum outperform quality in the AI backtest?
During strong sector rallies, investors often prioritize growth prospects and price momentum over fundamental safety. This bias leads to higher absolute returns for momentum strategies, though it comes with significantly higher drawdowns during market pullbacks.