Are We in a Bubble?
Summary
- The continued, rapid price appreciation and market dominance of leading technology companies since the 2022 release of ChatGPT has created the popular perception of an "AI bubble," echoing the frenzied speculation seen in previous technological breakthroughs.
- Metrics like the Shiller CAPE Ratio suggest U.S. equities are historically expensive (second only to the late-90s tech boom). However, the implied U.S. equity risk premium is only slightly below average, not signaling the "risk blindness" typical of a mania peak.
- Current AI-related capital expenditure, though rising, is estimated to be less than 1% of GDP. This is substantially more restrained than the 2%-5% of GDP seen during previous defining technology cycles (e.g., the 1990s tech buildup).
- Unlike previous bubbles where companies used excessive debt and corporate margins peaked years before the collapse, today's AI investment is largely funded by strong free cash flow, and profit margins remain steady.
- While some bubble characteristics exist, conditions haven't reached past danger levels. Since a true bubble is only clear after the fact, the best defense is to avoid outsized exposure to any single theme (including AI) and maintain broad diversification.
“History doesn’t repeat itself, but it often rhymes.” – Mark Twain
The continued ascent of the U.S. stock market, particularly the rise in the leading technology companies, have led many to conclude that we’re in an AI bubble. To many, this has represented a classic case of hope over reality.
Every few decades, a new technology promises to change everything. Railroads. The internet. Crypto. Now AI. But before we decide whether AI fits that pattern, it’s worth stepping back to look at what a bubble actually is, how past bubbles formed, and whether the same signs are showing up today.
What is a Bubble?
The success of the major technology companies since the release of ChatGPT in 2022 is obvious to anyone paying attention. Their rise makes intuitive sense. But rapid price appreciation alone doesn’t automatically mean we’re in a bubble or that the market is setting up for a dramatic collapse.
Defining a bubble isn’t easy because each one looks different. Still, most share a common pattern: frenzied speculation tied to a new engine of growth. Whether it’s railroads, the internet, or AI, a breakthrough technology captures investors’ imagination and draws in a wave of capital from both established players and eager newcomers.
As enthusiasm builds, asset prices surge in anticipation of future profits. Companies race to invest and expand, capital floods into the sector, and the story feeds on itself. Eventually, though, the fundamentals stop keeping up. When prices finally break, spending and investment dry up, and the unwinding can spill into the broader economy.
At its core, a bubble forms when the combined market value of the companies involved far exceeds the future cash flow they can realistically produce.
History gives us plenty of examples. A few of the most famous include:1
- 1630s – Tulip Mania in Holland
- 1790s – The Canal Mania in the UK
- 1840s – The Railway Bubble in the UK
- 1873 – The Railway Bubble in the US
- 1920s – The Stock Market Boom in the US
- 1980s – The Land and Stock Bubble in Japan
- 1990s – The Technology Bubble, Global
- 2007 – The Housing / Banking Bubble in US and Europe
It’s true that today’s wave of innovation shares some traits with the early stages of past bubbles. And while there’s no universally agreed-upon definition of a financial bubble, the same signs tend to repeat themselves: rapidly rising asset prices, stretched valuations, and growing systemic risks driven by rising leverage and falling profitability.
Prices have clearly surged. Now let’s turn to how this environment stacks up on the three other hallmarks of bubble behavior—valuations, leverage, and profitability.
Valuations
Valuation can be an important lens though which to explore whether there is true exuberance building. By many measures, U.S. equities are as expensive as they’ve ever been outside of the tech and telecom bubble of the late 1990s.
One such measure many look at is the Shiller CAPE Ratio. This ratio takes the current price of the S&P 500 Index and divides it by the average of the last 10 years of earnings (adjusted for inflation). A high CAPE suggests the market is rich relative to its long-term earnings power; a low CAPE suggests the opposite. The metric isn’t perfect, but historically, elevated readings have lined up reasonably well with lower future returns.
Beyond the CAPE metric, we like to look at the implied U.S. equity risk premium, which measures the extra return investors demand for holding equities instead of U.S. Treasuries.
According to Professor Aswath Damodaran, who calculates this measure each month, the premium currently sits around 3.95%.2 With this measure, lower readings point to lower future expected returns. Today’s level is slightly below its long-term average of roughly 4.25%, but it’s not remotely signaling the kind of risk blindness we saw during the late-90s mania.
Source: Adjusted CAPE Ratio (earlyretirementnow.com); Implied Equity Risk Premium (Aswath Damodaran); Nova R Wealth. Adjusted CAPE Ratio data as of 11/13/2025; Implied Equity Risk Premium Data as of 11/1/2025.
On balance, valuations in the U.S. are looking stretched relative to history, but they haven’t reached the extremes that typically appear before a bubble breaks.
Investment, Financing, and Profitability
Today’s AI investment is undeniable; capital investment from companies like Amazon, Meta, Google, MSFT, and ORCL is set to double from where it stood when ChatGPT first hit the scene. Announcements of even bigger outlays have sparked debate about whether the pace is sustainable. But when you view this cycle through a historical lens, the scale still looks modest. Stack the current surge against the late-1990s tech buildup and you’ll see a far more restrained and grounded investment landscape.
Source: Goldman Sachs Global Investment Research; Bureau of Economic Analysis.
Even with 2025 almost behind us, estimates show that AI-related investment (<1%) remains meaningfully smaller as a share of GDP than in prior defining technology cycles (2%-5%).3 In other words, spending is rising, but we’re nowhere near the runaway acceleration that defined earlier bubbles.
Another crucial difference lies in leverage and financing behavior. In the 1990s, companies leaned hard on debt to fuel their expansion. Corporate borrowing climbed sharply, and balance-sheet health deteriorated well before the stock market finally cracked. In contrast, the financing behind today’s AI build-out looks much stronger. So far, large firms are funding most of their investment out of free cash flow. While debt issuance has ticked up recently, it’s rising from historically strong levels and remains below the danger zones of the past booms.
Profitability tells a similar story. In the late 1990s, corporate profit margins peaked years before the bubble burst. That breakdown was one of the earliest signs the foundation was cracking. Right now, we’re not seeing that same erosion. Margins have remained steady, and productivity growth is only beginning to capture the benefits of AI adoption. Put simply, the economic engine behind AI is still running smoothly, not straining under its own weight.
Source: Bureau of Economic Analysis; Bloomberg; Nova R Wealth. Data as of 6/30/2025.
Significance and Action
All of this matters because when fundamentals finally break down, the consequences can be brutal.
After Japan’s asset bubble burst in 1989–1990, the Nikkei 225 Index lost more than 80% of its value and took over 30 years to recover. Similarly, when the tech bubble cracked in 2000, the NASDAQ plunged nearly 80% from its highs, wiping out years of gains and forcing investors into a long, griding recovery. We highlight these examples because they show just how severe the fallout can be when valuations, leverage, and investment imbalances finally reverse.
To be clear, while we’re seeing some of the hallmarks of a bubble such as high valuations, rising investment relative to GDP, and an uptick in leverage, we don’t believe conditions have reached the levels typical of past episodes before they burst.
If we’re still in the early innings of AI, pullbacks along the way are to be expected on the road to more potential gains. The more immediate risk is that earnings disappoint and investors begin to question the sustainability of current growth rates. That scenario could spark a meaningful correction, though it’s less likely to trigger the kind of broad collapse we’ve seen in previous bubbles.
Looking back, past bubbles always seem obvious in hindsight. But in real time, they rarely are. We believe identifying a true bubble is only clear after the fact. And if that’s the reality, then the best protection for investors is simple: reassess your portfolio in light of your goals, avoid outsized exposure to a single theme (AI included), and diversify across asset classes, geographies, sectors, factors, and risks.
At the end of the day, the greatest danger in any bubble isn’t owning the hot theme; it’s owning it exclusively and with leverage.
(1) Goldman Sachs Global Investment Research. Global Strategy Paper: Why We Are Not in a Bubble… Yet. October 8, 2025.
(2) Damodaran, Aswath. Home Page for Aswath Damodaran - NYU Stern. New York University, 17 Nov. 2025, pages.stern.nyu.edu/~adamodar/.
(3) Goldman Sachs Global Investment Research. Global Economics Analyst: The AI Spending Boom Is Not Too Big. October 15, 2025.
IMPORTANT DISCLAIMERS AND DISCLOSURES
This presentation is for educational and illustrative purposes only. It is not intended to offer or deliver investment advice in any way. Past performance is not indicative of future results. The information contained in this presentation has been gathered from sources we believe to be reliable, but we do not guarantee the accuracy or completeness of such information, and we assume no liability for damages resulting from or arising out of the use of such information. Any projections, market outlooks, or estimates are based upon certain assumptions and should not be construed as indicative of actual events that will occur. The performance numbers displayed herein may have been adversely or favorably impacted by events and economic conditions that will not prevail in the future. Any index presented does not incur management fees, transaction costs or other expenses associated with investable products. It is not possible to directly invest in an index.
Nova R Wealth, Inc. is a registered investment advisor with the U.S. Securities and Exchange Commission. SEC registration does not constitute an endorsement of the firm by the Commission, nor does it indicate that the adviser or investment adviser representative has attained a particular level of skill or ability. Nova R provides investment advisory and related services for clients nationally. This material is proprietary and may not be reproduced, transferred, or distributed in any form without prior written permission from Nova R Wealth, who reserves the right at any time and without notice to change, amend, or cease publication of the information contained herein.