Key takeaways: A recent economic analysis found that the current scale of capital expenditure flowing into AI far exceeds any previous capital boom cycle in history, including the dawn of the railway era and the internet wave.

In 2026, mega-cap technology stocks are seeing heightened volatility and negative free cash flow; bond market yields are rising, and some AI-related issuers have fallen into junk bond territory.

This situation is pushing some investors toward a quality asset safe haven, with a focus on balance sheet metrics: ample cash, return on capital, and net debt levels. The quality stock pool also includes the U.S. Magnificent Seven (Mag 7).

On February 13, 2025, the Wall Street Charging Bull in New York's financial district. A new estimate shows that AI infrastructure investment will exceed most major economic transformations in U.S. history.

Columbia Business School professor Stijn Van Nieuwerburgh wrote in an article published at a Brookings Institution event last week: Relative to the size of the economy, this AI infrastructure investment will exceed the investment booms in U.S. canals, railways, electrification, highways, and telecommunications.

Even though these leading tech companies are already among the world's most profitable firms, massive AI spending is still squeezing their cash flow.

At the same time, the bond market is struggling to absorb companies' huge debt financing needs, with more and more bond issuance falling into junk status; the U.S. Treasury selloff and rising yields are partly driven by AI capital expenditure pressure.

TMX VettaFi head of research and editorial Todd Rosenbluth said that in a volatile environment, investors are focusing more on core balance sheet metrics, with free cash flow being the primary indicator (free cash flow refers to cash remaining after expenses, interest, taxes, and long-term capital investment).

Return on equity, net debt leverage, and earnings stability have also become key metrics for investors.

Behind this quality asset safe haven approach lie two key questions: how much AI capital expenditure companies are willing to commit for long-term growth, and which companies can withstand such capital spending pressure.

Rosenbluth said: The market is beginning to reassess AI investment expectations, and investors are turning to companies they are confident can sustain growth even if the environment changes.

During periods of low rates and high growth, investors were sometimes willing to pay a premium and take on higher risk for future growth. But as rates rise and monetary policy tightens, market sentiment has shifted somewhat, and some investors have become more cautious.

Some AI infrastructure-related IPO projects have been shelved, and at least for now, the market's scrutiny of such offerings has become stricter.

The broad S&P 500 market showed a surprising phenomenon in 2026: the index is near record highs, but earnings expectations and stock price movements have diverged. Wall Street analysts keep raising earnings forecasts, yet investors are unwilling to speculate on valuation growth.

Sean Snyder, economic strategist at Potomac Fund Management in Bethesda, Maryland, said investors are increasingly reluctant to allocate to highly indebted companies, or those that borrowed excessively and lost control of risk in a high-rate environment.

Most corporate debt is tied to the U.S. 10-year Treasury yield. Snyder said that when investors see the 10-year Treasury yield hovering around 5%, they look for higher quality names in their portfolios.

Focus on companies with stronger free cash flow and lower reliance on high-interest debt. Cash today is worth more than future cash promises.

U.S. 10-year Treasury real-time quote: 5.293%, +0.038. In Iowa City, the August report from Raymond James' James Investment Group reminded investors that the market is beginning to focus on whether AI sector growth can continue.

Morningstar fund manager research analyst Zachary Evans said investors increasingly want to see solid support for AI growth. When evaluating stock prices, they need to confirm that companies have viable business models and user bases to support AI infrastructure investment.

Tech giants' free cash flow situation. Not long ago, cloud hyperscalers' cash reserves were more than enough to easily cover investment. The Raymond James investment team wrote: But by the second quarter of 2026, these companies' combined capital expenditure had exceeded operating cash flow, and free cash flow turned negative.

This does not necessarily mean a bearish view on these stocks. The report added: The key is that these companies remain extremely profitable; the problem is not profitability itself. Rather, the scale of AI investment is so large that even their abundant cash flow cannot fully cover it.

Its analysis shows that the S&P 500 technology sector's free cash flow yield (free cash flow/market cap) is roughly in line with the S&P 500 overall average, both at 4%.

But this clearly reflects the market contradiction and is the core question more and more investors are considering. The report concluded: Free cash flow is not constant, and capital expenditure is cyclical. AI is still in the early adoption stage, and data center investment is surging. Over the long term, as more data centers come online, hyperscalers are expected to reduce capital expenditure and reap new revenue from these facilities.

According to data from the U.S. Census Bureau and the St. Louis Fed, since December 2023, AI data center construction investment has increased by $51 billion; all other private construction project spending combined has fallen by $120 billion.

For many investors, free cash flow is a metric worth tracking. Morgan Stanley's July research report wrote: Companies that can generate sustainable free cash flow, maintain strong competitive moats, and allocate capital efficiently are generally better able to withstand economic uncertainty, changes in the competitive landscape, and market volatility.

There are not many funds focused specifically on free cash flow, but investor interest is clearly rising. The VictoryShares Free Cash Flow ETF (ticker VFLO) has seen significant inflows over the past month, reflecting investor demand. The ETF has $11 billion in assets, and the latest ETFAction data shows monthly net inflows of more than $900 million, with an annual expense ratio of 0.39%. There are other thematic ETFs in the market, but none have recently been as hot as this product.

The $18 billion Pacer U.S. Cash Cows 100 ETF (COWZ) had net inflows of only $24 million over the past month, with a 0.49% expense ratio. The top 10 holdings of such ETFs define quality stocks far more broadly than simply concentrating on top tech giants.

Although both funds' largest holdings are tech stocks (Micron for VFLO, Qualcomm for COWZ), healthcare, energy, professional services, insurance, and defense stocks also rank in the top 10.

Another product in the same space, the Amplify COWS Covered Call ETF (HCOW), overlays covered calls on a cash flow strategy to generate income for investors. As of September 25, monthly net inflows were about $804,000. The product focuses on income and is gaining popularity among baby boomer investors; with the options overlay, its expense ratio is higher at 0.65%. The fund was launched in September 2023 and has now been around for three years, but its assets are only about $17.2 million, making it relatively small.

ETF strategy feature: read more. More funds focused on free cash flow continue to be brought to market. Pacer expanded its strategy ETF lineup this year, launching the Pacer S&P 500 Quality Free Cash Flow R&D Leaders ETF (QFRD) and the Pacer S&P 500 Quality High Dividend Free Cash Flow ETF (QFHD).

The R&D Leaders ETF tracks the S&P 500 Quality Free Cash Flow R&D Leaders Index, screening 50 companies within the S&P 500 with high R&D intensity and excellent free cash flow margins. The High Dividend ETF tracks the S&P 500 Quality High Dividend Free Cash Flow Index, and constituents must meet at least five consecutive years of stable dividend payments while also meeting high standards for free cash flow quality.

Other quality factor ETFs, and the comeback of the Lagging Seven. In addition, quality factor ETFs that also consider free cash flow, return on equity, net debt leverage, and earnings stability have also recently attracted fund attention.

Evans said: Quality ETFs have generally performed well - even without an explicit high free cash flow focus, their stock selection core is still quality, and free cash flow is often part of the quality metric.

One important point: such funds still regard the Magnificent Seven (Mag 7) as core quality holdings.

The $48 billion iShares MSCI USA Quality Factor ETF (QUAL) saw strong inflows over the past month, with ETFAction data showing monthly net inflows of about $308 million and an expense ratio of only 0.15%. Microsoft, Apple, Nvidia, and Meta are among the ETF's largest holdings. Morningstar data shows that as of September 28, the ETF was up about 13% year to date, roughly in line with the S&P 500 ETF (VOO); but other quality factor ETFs performed better.

The $9 billion JPMorgan U.S. Quality Factor ETF (JQUA) rose about 18% year to date, attracted about $300 million over the past month, and has an expense ratio of 0.12%, lower than QUAL. Its holdings are entirely tech stocks, including Meta, Microsoft, Nvidia, and Alphabet.

The $3.2 billion WisdomTree U.S. Quality Growth Fund (QGRW) rose nearly 17% year to date, saw about $153 million in monthly inflows, and has an expense ratio of 0.28%, also holding Magnificent Seven names.

Even though some quality factor ETFs have outperformed the S&P 500, they still lag the momentum trading leader in the ETF space. The iShares MSCI USA Momentum Factor ETF (MTUM) is heavily invested in popular chip and memory stocks and is up more than 26% year to date. But enthusiasm for the strategy has recently faded, with the fund down 7.7% quarter to date, potentially its worst quarterly performance since the second quarter of 2022. ETFAction data shows that this $20 billion fund saw about $5 billion in outflows over the past month.

The Magnificent Seven, which underperformed the S&P 500 in 2026, is showing signs of a rebound. The Roundhill Magnificent Seven ETF (MAGS) is up less than 9% year to date, weaker than the S&P 500; but it has risen nearly 3.5% over the past month, while the S&P 500 has struggled to stay in positive territory over the same period.

Richard Rell, chief investment officer at Questar Capital Partners in Paramus, New Jersey, wrote in a recent market commentary: Throughout the year, the Magnificent Seven became the Lagging Seven, with hot money flowing into AI infrastructure favorites such as memory stocks; but now the market is full of concerns about AI spending, and the Magnificent Seven is once again in favor.

The market widely believes that, apart from Nvidia, Microsoft, Amazon, and Alphabet have large traditional cash flow businesses that do not directly depend on AI, enabling them to withstand a cooling in AI spending or negative sentiment. If AI investment contracts, their diversified business models can rely on traditional operations for support.

Charles Kantor, managing director and senior portfolio manager at Neuberger Berman, said in a recent interview that although the market questions these companies' negative free cash flow, the long-term return on this investment will eventually materialize.

When companies buy back stock, everyone cheers, but when they invest in business expansion, they face heavy skepticism. He said investors find it hard to understand the sheer scale of this investment. Companies have also not disclosed enough information to convince the market (of course, disclosure itself is difficult while maintaining competitive advantage).

But he added: We calculate the returns on such investments and believe that as long as one accepts the reality of no short-term payoff, the long-term returns will be highly attractive... At some point, money will return to quality factors and to companies with moats, and that will be a very attractive entry window.