3Q Investment Commentary: Exploring a Possible AI Bubble

By Paul A. Bodnar
CM Wealth CHIEF Investment OFFICER, Partner

Global equity markets advanced in the third quarter, with the S&P 500 gaining 8.1% to reach new all-time highs. The rally was supported by resilient corporate earnings and economic data suggesting the labor market, while still strong, is cooling enough to give the Federal Reserve policy flexibility. Notably, market leadership broadened, with the ‘Magnificent 7’ underperforming the broader index for the first time in recent memory as small-cap value stocks outperformed.

We believe the current environment is not a speculative bubble but the initial phase of a multi-decade buildout around AI. However, there are signs of excess. The path forward will not just move up and to the right; there will be bumps in the market along the way.

The value created by AI will be twofold: first for the AI-native "Enabler" companies, and second, for the established "Adopter" companies in traditional industries that successfully integrate the technology to create durable competitive advantages. The latter needs to be significantly larger than the former to justify the massive infrastructure investments and to support further investment.

JP Morgan estimates AI contributed 1.1% to GDP growth in the first half of 2025, which represents ~80% of total GDP growth, so any softening in AI demand could carry wider economic impacts. The near-term risk to the markets and AI stocks is an external shock, like a recession, which would tighten financial conditions and trigger a painful market shakeout.

CM Wealth is actively investing across the AI ecosystem in both public and private markets. There were two noteworthy events on this front in the quarter. Our December 2024 investment in Databricks was marked up 63%. We are also committed to participating in the spinoff of Verily from Alphabet. We would view a market dislocation opportunistically and remain focused on investing for the long term.

With Valuations So High, Are We in an AI Bubble?

By any traditional measure, many of the technology stocks leading the market appear expensive. The S&P 500 is trading at over 22x forward earnings, a level significantly above its historical average (5-year average: 19.9x; 10-year average: 18.6x). Valuations are even more stretched for companies at the center of the AI theme. It is companies that trade at these rich valuations in public and private markets that have some calling it a bubble. For example, a company like Palantir trades at over 600x trailing earnings and more than 200x forward earnings. While Palantir is truly a special company, these unprecedented multiples price in years of flawless execution and growth.

This has understandably led to widespread concern about a speculative bubble, fueled by staggering capital spending plans. For instance, OpenAI’s CEO Sam Altman announced a $500 billion AI infrastructure plan known as Stargate and has since suggested the company could spend “trillions” on AI. To finance this, some companies are using unconventional arrangements that have raised eyebrows. In September, Nvidia announced a deal to invest up to $100 billion in OpenAI’s data center buildout, leading some analysts to question whether the chipmaker is propping up its own customers. This concern is not new; Nvidia has backed dozens of companies that then use the capital to buy its expensive chips. The sheer scale of capital required is immense; OpenAI alone expects to burn through $115 billion in cash by 2029.

While these dynamics create pockets of significant froth and warrant caution, we do not believe this is a repeat of 1999. We are in the early innings of AI investment and adoption that will power the next era of economic growth. Unlike the dot-com era, today’s buildout is led by many of the world's most profitable corporations with formidable balance sheets and cash flows. While the valuation debate will continue, we believe the more critical question for long-term investors is identifying where the value from this technological shift will ultimately accrue.

The Two-Sided AI Opportunity

The long-term value created by AI will not be confined to just technology companies. We see a powerful, two-sided opportunity that will unfold over the coming decade.

First are the enablers: the companies designing the chips, building the models, and providing the cloud infrastructure. They are the essential "picks and shovels" of this technological gold rush and have been the market's initial focus. Second, and perhaps more importantly, are the adopters: companies that leverage AI to create profound competitive advantages, often in traditional industries. This is where we believe the most durable, long-term economic value will be created. Leading analysts now estimate that the economic impact of AI will be measured in trillions of dollars, with much of that value captured not by tech companies, but by the traditional industries that successfully utilize it in their operations. To date, results are mixed, but there are early successes, including:

• Banking: JPMorgan Chase's $2 billion annual AI investment is already generating a matching $2 billion in annual benefits, according to a recent interview with Jamie Dimon.

• Healthcare: AI is being deployed at an unprecedented speed, with a recent Bain survey finding that 54% of healthcare organizations see a material return on investment within the first year. One healthcare provider told us there was an 80%+ reduction in staff in revenue cycle operations after implementing AI-based solutions.

• Logistics: Companies like UPS are using AI to optimize their complex delivery networks. Its ORION system uses advanced algorithms to create the most efficient routes for its drivers, saving the company an estimated 100 million miles and 10 million gallons of fuel each year, representing hundreds of millions of dollars in annual savings.

CM Wealth Portfolio Update - AI

At CM Wealth, our strategy is to invest across this entire spectrum to capture the full breadth of the opportunity. We believe this combination of public and private exposure is the most effective way to participate in the theme, capturing both the stability of established leaders and the high-growth potential of emerging disruptors.

In public markets, this includes positions our managers hold in companies like Nvidia, as well as indirect investments, such as uranium miners that will help provide the power. In private markets, we made significant investments in digital infrastructure required for AI's expansion. We invested in Tract during 2023, a company that develops master-planned data center parks to meet the massive power and land requirements of next-generation computing. We also have exposure to next-generation semiconductors. Through our long-standing partnership with Primary VC, we are investors in EtchedAI, an ambitious startup developing specialized AI chips that could be significantly more efficient than current GPUs for specific tasks.

Additionally, in our venture program, we have exposure to companies like SandboxAQ and ReflectionAI. We have also invested in the large, late-stage private companies in our public equity book. For example, in December 2024, we invested in Databricks, which was recently marked up by 63%. We also committed to invest in Verily, a spinout from Alphabet, which seeks to be a key platform for data and AI in healthcare.

Outlook: Navigating the Bumps in the Road

The path to this AI-driven future will not be linear. The investment landscape is shaped by a complex macroeconomic environment, where persistent inflation may keep interest rates elevated, increasing the risk of an economic slowdown. Geopolitical tensions, specifically the ongoing US-China rivalry over advanced semiconductor technology and export controls, present a direct risk to the hardware supply chain essential for the AI buildout. Furthermore, a robust debate is underway about whether these productivity gains will translate into broad economic growth or primarily benefit a narrow set of companies and capital owners, a theme we are monitoring as it has long-term societal and market implications.

The most significant near-term risk remains an external shock or a policy misstep that leads to a recession. An economic downturn would tighten financial conditions and force a pullback in capital spending, causing the funding that fuels much of the innovation to dry up. This would undoubtedly trigger a painful shake-out, punishing companies with high cash burn rates and unproven business models. However, we view this risk not as a threat to the long-term thesis, but as a temporary and even healthy market correction. For the well-capitalized technology leaders and the traditional companies already reaping the rewards of AI adoption, a downturn would be a bump in the road from which they could emerge even stronger.