2025 Third Quarter Investment Market Report

Markets at a Crossroads?

Is the U.S. economy prospering from the “A.I. boom” or is it struggling as the jobs market slows markedly? Last quarter we predicted that economic data would begin to take shape in Q3 to inform us if last spring’s tariff turmoil would result in a slowing economy or if the underlying strength from the boom in A.I.-related capital expenditures would overcome any drags on the economy.

…and the verdict is… both?? Markets and policymakers seem to be at a crossroads of conflicting outlooks, with an unusually divergent array of substantial data supporting both economic optimism and pessimism, although markets seemed focused on the optimism in Q3:

U.S. Equities

Source: Morningstar

• On the one hand, the U.S. job market appears to be deteriorating to such an extent that the Federal Reserve felt compelled to begin cutting interest rates again. The need to cut interest rates is generally considered a negative for the stock market.
• But on the other hand, massive capital investment on artificial intelligence infrastructure, continued growth in corporate profits, and unabated consumer spending indicate that the U.S. economy, which has repeatedly surprised with its upside strength, is still cruising comfortably.

Among other looming uncertainties,

• The federal government has been collecting substantial tariff revenues from importers, but only a small portion of that cost has been passed through to consumers so far, meaning the jury is still out on whether the result will be lower corporate profits or higher costs to the consumers.
• The fourth quarter will start with yet another government shutdown, but the markets have become almost completely desensitized to these exercises in brinksmanship as none of the previous shutdowns have led to sustained economic damage.

…the U.S. economy, which has repeatedly surprised with its upside strength, is still cruising comfortably.

Despite a widening range of economic possibilities, the stock market seems to be pricing in a “best case” scenario, with most major indexes reaching record highs in the quarter.

An A.I. Crossroads for Equities?

Will A.I. be the fuel for a transformative productivity boom or become a source of such concentrated risk that if one domino falls, it leads to a cascade of collapses? If it resembles previous investment cycles in transformative technologies, it might be both.

Markets are cyclical. Investment booms are cyclical. Dating back to at least the construction of the railroads, potentially transformational innovations have gone through five distinct phases:

1. Early-stage excitement and cases of immense wealth creation

2. Market hysteria where investors feel compelled to get involved, even though they have no idea where eventual profits may come from

3. Overinvestment and overcapacity

4. Disappointing returns, sometimes crashes, and investor disenchantment

5. Longer-term, widespread adoption, often made more economical by overcapacity, leading to increased productivity

What does this mean in today’s A.I. market boom??

As Mark Twain is credited with saying,

“history doesn’t repeat itself, but it often rhymes,”

which explains in the market context how every boom/bust cycle looks different enough from those that preceded it that investors can be convinced that “this time is different.”

Is it possible that “this time is different”? Yes.

Is it likely? No.

…every boom/ bust cycle looks different enough that investors can be convinced that “this time is different.”

A quarter century ago, WorldCom, Cisco Systems, Intel, and Microsoft were among the very largest market capitalizations in the S&P 500. WorldCom was building out fiber optic cable nationwide as fast as it could, not knowing where the profits would come from, but cheered on by investors that the internet would be so transformative, all that mattered was building out fiber capacity. WorldCom went bankrupt within a few years. In fact, only one of the four (Microsoft) can claim today to be among the S&P 500’s largest, and it still took 15 years to recover its losses from the last “transformative technology” boom/bust. Another (Cisco) took nearly 20 years to recover its losses, and the last (Intel) trades at a lower valuation today than 25 years ago and is a shell of its former self in terms of technology leaders.

The chart below shows how the technology-laden NASDAQ composite was “dead money” for 15 years after our last transformational boom went bust.

NASDAQ Composite % Change 1997-2015

In addition to being the biggest stock in the S&P 500, Nvidia holds a more concentrated share of the index than any other stock in history. You can also see in the chart below that the only times the index’s largest stock traded a higher P/E multiple was Microsoft in 1998 and 1999.

Largest company share in S&P 500

Today, the biggest contributor to U.S. economic growth is the massive buildout of A.I.-related capacity, which has reached the point, as in past cycles, that capacity is being built without a knowledgeable basis in future profits. But investors are cheering on the largest stocks in the S&P 500 who are making hundreds of billions of dollars of such capital investments because they believe that A.I. will be so transformative, all that matters is building capacity. Sound familiar?

Today, the biggest contributor to U.S. economic growth is the massive buildout of A.I.-related capacity…

Could it be different this time?

Even though this cycle has already brought the S&P 500 to a price/earnings valuation level never seen before except during the internet bubble, the A.I. bulls argue that it is indeed different this time. Rather than a market populated by unprofitable companies financed by debt and venture capital that found themselves unable to meet their debt obligations and that went bankrupt when the music stopped, this market is led by established companies who are using their own extremely strong levels of free cash flow to make these A.I.-related investments, so the downside from overinvestment is not the same.

So, just to be clear – these non-cyclical technology companies with “asset light” balance sheets generating hundreds of billions in free cash flow from software and services are using most of that free cash flow to become “asset heavy” by investing in property, plant, and equipment as fast as they can, trending toward balance sheets that look more like cyclical industrial companies (who trade at much lower valuation multiples due to cyclical vulnerabilities).

Capex share of operating cash flow for hyperscalers: 60%

In addition, capital is not without a cost, whether it is financed by debt or by cash from operations. Poor allocation of capital, regardless of its source, can hamstring a company for years. Perhaps a Mag 7 company doesn’t completely collapse like so many of those internet-fueled stocks from the last cycle, but could it drop 50% because investors no longer think it is worth a sky-high price/earnings multiple when it is saddled with supporting a trillion dollars’ worth of data centers that it discovered to be beyond its needs? In a 2000s redux, could its stock price recovery from such a depth be stalled for a decade because its cash flow is already spoken for and can’t be used for technological development?

Keep in mind that technology continually gets cheaper and smaller with new competitors emerging. Is there a startup out there that will build a better A.I. model with a fraction of the computing need? History would argue that is a very real threat. We carry computers in our pockets more powerful than those more than 100 times their size not that long ago. Will advances in technology make data centers the “fiber optic buildout of 2025”?

Keep in mind that technology continually gets cheaper and smaller with new competitors emerging.

Market Crossroads: What are the caution flags?

While the markets have clearly focused on the positives this past quarter, there are plenty of caution flags waving beyond the historically high valuation multiples and unprecedented concentration of value in the U.S. stock market in a small group of A.I.-fueled stocks (the top 10 stocks in the S&P 500 now account for over 40% of the value of the entire index!):

Looking beyond the cash-rich Mag 7, many of the high-flying “A.I.” companies in the headlines, as well as an entire wave of startups, are, in fact, not profitable and don’t expect to be for years to come, meaning they are very reliant on debt and/or venture capital! Let’s look at perhaps the most famous example and its potential ripple effects on other companies:

…many of the high-flying “A.I.” companies are, in fact, not profitable and don’t expect to be for years to come…

• OpenAI’s most recent valuation is approximately $500 billion on revenue of just $13 billion
• While the most famous, ChatGPT finds itself in an increasingly crowded market of generative A.I. tools
• Reportedly, attempts to convert free users into paying subscribers have been disappointing due to competition and issues with quality and reliability
• OpenAI expects to lose $44 billion over the next four years
• Yet, OpenAI signed a deal to pay Oracle $300 billion over the next 5 years for computing power
• And, OpenAI unveiled plans last week for a $1 trillion build-out of data centers
• At a valuation of almost 40x revenues with tens of billions of dollars in losses, the market seems to be pricing in near limitless potential for OpenAI, despite an increasingly crowded and capable field of competitors

Cisco Systems was a very profitable company in 2000 – its problem was that it was reliant on commercial relationships with unprofitable customers. Like a dot.com that ordered thousands of servers that it never paid for, what is the downside for OpenAI to sign all these deals? Especially if the vendor will finance them!

“Circularity” has become a popular term to describe how some profitable A.I. firms are providing financial support to their customers to enable the customers to buy more of the seller’s product, essentially leveraging up their exposure to those clients. Let’s look at potential sources of ripple effects on two very profitable companies if OpenAI’s plans stumble:

• Oracle is currently the 12th largest company by value in the S&P 500 after jumping approximately 40% on announcement of its $300 billion deal with OpenAI
• Presumably, Oracle has now committed a substantial share of it’s A.I. computing capacity to an unprofitable company whose annual revenue is less than ¼ of its annual commitment to Oracle
• Nvidia has made investments in several high-profile clients, including OpenAI
• Nvidia’s market value jumped by $160 billion upon announcement of a $100 billion investment in OpenAI, which illustrates some concerning market alchemy.
• Reports estimate that for every $1.00 Nvidia invests in OpenAI, it gives OpenAI the financial capacity to spend $3.50 on Nvidia chips

However, despite the signs of froth, timing a market peak is virtually impossible. “Irrational” exuberance can continue for years, and, as the graph on page 3 shows, a “wise” investor who sold technology stocks two years early in 1997 sacrificed enormous return potential to avoid losses on gains that they never realized. What if there is another stage to this boom before the bust?

Market Crossroads: Green Flags

As we mentioned earlier, the U.S. economic data is unusually divergent, and for all the caution flags, there are plenty of green flags as well.

• With 3.8% GDP growth in the second quarter, the U.S. economy continues to surprise to the upside with its strength.
• Interest rate cuts by the Federal reserve when the economy is not in a recession usually imply very strong performance for U.S. stocks.
• The massive capital investments being made, largely by the “hyperscalers,” such as Microsoft, Amazon, and Google, are providing a substantial boost to the U.S. economy, not just to buy A.I. semiconductor chips, but for building the data centers and the massive energy capacity to power them.
• Consumer spending continues to be surprisingly resilient, although growth is largely focused among higher income brackets who are benefitting most from strong financial markets.
• For all the reports of a weakening jobs market, the unemployment rate remains historically low at 4.3% with layoffs and jobless claims remaining subdued as fewer job openings are offset by fewer available workers in a “no hiring, no firing” paradigm.
• Market anxiety about trade policy and tariffs has greatly improved

For all the reports of a weakening jobs market, the unemployment rate remains historically low at 4.3%…

Not surprisingly given the new market highs, “FOMO” or “fear of missing out” on short-term market rallies seems to be supplanting the fear and caution brought on by the tariff-related April market turmoil.

International Equities

International stock markets continued their powerhouse 2025 performances in the third quarter, retaining a comfortable lead over U.S. markets so far this year. With a calming in trade rhetoric, China led emerging markets higher, and countries, such as South Korea, that reached trade agreements with the U.S. fared particularly well. A July trade deal with the EU helped developed market stocks post a strong quarter as well, and Japan finally seems to be emerging from more than three decades of economic nightmare.

International stock markets continued their powerhouse 2025 performances in the third quarter, retaining a comfortable lead over U.S. markets so far this year.

International Equities

Source: Morningstar

What about tariffs, inflation, and the Fed?

In a highly anticipated and well-telegraphed move, the Fed resumed cutting short-term interest rates in September. While there have been months of political calls for the Fed to cut more aggressively, the markets have largely agreed with the Fed’s more patient path. In fact, with so much uncertainty regarding whether one-time tariff-based price increases could lead to a resumption of higher consumer inflation expectations, there was significant risk that earlier rate cuts could have fueled the bond market’s inflation concerns. This would have likely resulted in higher long-term rates (and mortgage rates) despite lower short-term rates. With inflation still seemingly on a controlled path, bond markets rallied around the rate cut.

Fixed Income

Source: Morningstar

What does all this mean for client portfolios?

What does this mean for investors who find themselves at this crossroads? Should they keep riding the “A.I. wave” as the world becomes increasingly convinced that this time is different and there are endless profits to be had with this latest transformational technology? Should they sell exposure to Mag 7 stocks and large caps fearing the signs of market hysteria are getting too strong?

At Montis, clients have seen us preparing for this dilemma for the past few years. As the stocks at the very top of the S&P 500 keep getting increasingly expensive, investors must remember that the price you pay will impact your returns, and when you pay historically high multiples on each dollar of earnings, you simply can’t expect the returns going forward that you saw as those multiples grew and got stretched. However, it should not be an “all or none” decision, and we want to invest for the range of outcomes, not place a bet on a best or worst case. Many clients have seen us preserve our exposure to further upside, but trim some of it to invest in sectors where we see better value and some ability to mitigate any severe market decline. Whether it is shifting some S&P 500 exposure to the “non-Mag 7” stocks in the index, or midcaps, which have more attractive valuation and have actually outperformed large caps over their history, or small caps, which, in addition to lower valuation, are better positioned in a world of reduced global trade.

…the price you pay will impact your returns…

Beyond U.S. equities, we have found ways to invest in A.I. infrastructure through the data centers themselves and their long-term contracts to receive cash flows from the cash-rich Mag 7 companies. Also, we have invested in other private markets where disciplined underwriting standards determine the prices of transactions, not the “buy at any price” manias that occur in the public markets. With U.S. equities so expensive in recent years, we’ve had much more focus on international markets, which finally seem to be coming into their own this year. We have retained our potential to profit handsomely from a continued A.I. boom, but we have also found other potential sources of compelling returns to mitigate the A.I.-related risk, providing us with improved protection should that technology boom turn into another transformational technology investment bust.

Our objective is to build “durable portfolios” for clients – portfolios that find a broader range of attractive risk/return opportunities than traditional stock- and bond-only portfolios, but do so in a way that does not just blindly reduce risk to protect against loss, but rather reallocates risk to different sources of return so that the overall portfolio is positioned to generate a steadier stream of compelling returns as clients march toward their investment objectives.

Sources cited in this commentary

1. JPMorgan, Guide to the Markets, Sept 30, 2025
2. Bloomberg, October 2, 2025
3. www.slickcharts.com/sp500, October 3, 2025
4. Nvidia press release, September 22, 2025
5. Wall Street Journal, September 29, 2025