Mega IPOs & AI

Financial headlines about the IPO of SpaceX can be summarized as: It took off like a rocket on its first couple days (making Elon Musk the world’s first trillionaire) and then gravity began pulling the stock back toward Earth.

Nobody really knows how this will ultimately play out, but the hype around how an IPO could be an extraordinary investment creates FOMO (fear of missing out). The data, however, suggests the wiser choice is to avoid hot IPOs.

A recent Wall Street Journal article noted the average IPO in the US realizes a first-day gain of about 19%: “The average historical return for buying an IPO at the end of its first day of trading and holding it for three years is about 21% lower than what a value-weighted market index would have returned.”¹ This analysis from Jay Ritter, University of Florida, covered more than 9,000 US-listed IPOs from 1980 to 2024.

Research by Dimensional Fund Advisors corroborates this analysis, showing the underperformance is systematic. IPOs tend to be smaller, growth-oriented companies with low profitability and high rates of investment. Decades of research into asset-pricing theory have shown that this combination of characteristics in a company means lower expected returns. At Forum, we talk a lot about “factor investing,” which is fundamental to our investment approach, emphasizing those factors that have historically shown to have higher expected returns. With IPOs, we see the other side of the factors — the side with lower expected returns!

These long-term views are complicated by the market contortions in those first few days after a stock goes public. Employees and early investors in IPOs are typically restricted from selling most of their shares immediately following the IPO — they are forced to hold most of their shares for months, creating restricted supply when demand for shares is often exuberant. The limited access to shares can cause a spike that is hard to arbitrage due to the extreme volatility and limited trading volumes in its early days of trading.

That is not to say that IPOs never win. A handful do go on to deliver extraordinary longer-term returns. The problem is that it is only a handful — and how to know in advance which ones those are.

Dimensional expects to consistently apply their portfolio management rules to upcoming IPOs. They will add these new stocks only after enough of the stock is publicly traded (generally about a year post-IPO) and only if the stock meets the factor-based inclusion rules of the fund. If an IPO stock does not meet the criteria, that stock will not be added.

Another hot topic this year has been the AI revolution, which is being compared to the industrial revolution or the advent of the internet. At the highest level, we do think AI will result in a foundational change in how work gets done. Forum has been exploring AI tools internally on a limited, controlled basis, with an emphasis on data security and thoughtful oversight of what these tools can access. Our current use is focused on internal efficiency, supporting internal software development, drafting presentations, and streamlining internal processes. We are approaching adoption deliberately, and this caution is especially warranted when it comes to financial planning.

Specifically speaking of AI’s use in financial planning, we have found it to be dangerously convincing when used by someone without deep contextual expertise. Asking AI a seemingly simple question, like “Should I convert my IRA to a Roth IRA?” often gives a clear and concise answer, but it may be the wrong answer for the person asking. Properly answering that question for any individual or couple requires a host of considerations, including joint life expectancy, future tax rates, IRMAA (Medicare’s “Income-Related Monthly Adjustment Amount”) surcharges, estate taxes, inheritors’ tax rates, and charitable intent. Asking complex questions in a simple way seems easy but can lead to less-than-correct answers.

There are also many big, unanswerable questions around AI. How do workforces adapt? What do entry-level jobs and career paths look like? Who in society benefits, and is that okay? Which is faster: new job creation or unemployment? Which stocks will benefit: the big owners of the tech or the effective users of it? How do I use it personally, and where are the lines? We can’t answer many of these questions. We will answer the narrow one about investing in AI.

Are we invested in AI stocks? The short answer is yes. We can start by looking at the largest public companies in the market. Some are building chips. Some are building infrastructure. Some are building models. Some are embedding AI into software, services, advertising, phones, research processes, and manufacturing systems.

In other words, AI is practically everywhere. It is woven through the largest companies you already own, and it’s being used by the smallest. The largest names in the stock market themselves have an outsized exposure:

• Nvidia is a central player building chips used to both train AI models and in data centers bringing AI to consumers.
• Microsoft is heavily invested in OpenAI and is incorporating AI into its broad software ecosystem across Microsoft Office, Azure, GitHub, and more.
• Alphabet, Google’s parent company, is investing heavily in Gemini, AI search, custom chips, and advertising tools.
• Amazon is invested in Anthropic and is building AI into its AWS offerings, shopping algorithms, and its own AI chips.
• Meta is investing aggressively in AI throughout its social media and advertising platforms.
• Apple is improving Siri and putting AI onto its devices and in its operating system.

Simply by holding a broad market portfolio, you are swimming in AI and innovation. The harder question is whether we should intentionally tilt even more toward AI or “transformational technology” companies. That is where we would counsel for prudence and discipline.

One of the tricky things about investing in innovation is that even when the story is real, the prices often reflect what we already know. As such, the investment outcome can disappoint even if the prognostications come to pass.

The internet changed the world, but many internet stocks did terribly as investments when bought at too high a price. Electric vehicles are real. Streaming is real. Clean energy is real. Biotech is real. It’s important to acknowledge that not every exciting company or idea rewards investors.

AI might transform the economy over the next decade. But we do not know in advance which companies will capture the profits. For those largest US technology companies we mentioned, the prices per unit of earnings are much higher than historical averages, reflecting that there are already extraordinary expectations priced in.

So, we do not think of your portfolio as underexposed to innovation. We think of it as broadly exposed, without making a concentrated bet on any one specific company, technology, or IPO. That is portfolio intelligence!

We all want to benefit from AI and other innovations that reshape our world. But we do not want the portfolio to become overly dependent on today’s most exciting story and whether it plays out exactly as expected. We continue to invest in a diversified manner across the global market, tilting toward the factors of higher expected returns, and across the tens of thousands of companies that are all trying to profit from new technology and human innovation!