This Isn’t 2008—and AI Is Just Getting Started

Why Today’s Housing Market Isn’t 2008—and Why AI Is More Than a Dot-Com Rerun

I keep hearing the same two arguments: today’s housing market is another 2007–2008 waiting to happen, and artificial intelligence is simply the dot-com bubble all over again.

I think both comparisons overlook major changes in the financial system and the technology itself.

History matters. But if you are going to use an earlier crisis to predict the next one, you need to compare the underlying conditions. A familiar-looking chart is not enough.

Housing: Look at What Actually Caused the Last Crisis

The housing crisis developed alongside an extraordinary expansion of risky mortgage credit. Weak income verification, subprime lending, and loans that became unaffordable after introductory rates expired helped create the problem. Those risks spread through mortgage securities, connecting problems in individual households to the wider financial system.

That is the lending environment people remember when they talk about NINJA loans and the excesses of the housing boom.

Today’s mortgage market has a different foundation. The Federal Reserve’s May 2026 Financial Stability Report described mortgage credit risk as low, housing leverage as well below previous peaks, and overall mortgage delinquencies as low, supported by substantial homeowner equity and stronger underwriting. Subprime lending still exists, and some recent borrowers are under pressure, but that does not establish a repeat of the earlier credit structure.

There is also a major difference between the financing costs facing a new buyer and those facing an existing homeowner. The Fed’s July 2026 report noted that a majority of outstanding mortgages carried rates below 4%, helping explain why many owners were reluctant to sell and take on a more expensive loan.

Housing can be expensive, transactions can slow, and prices can decline without reproducing the same chain of mortgage defaults, securities losses, and banking failures that defined 2008.

An affordability problem and a systemic mortgage-credit crisis require different analysis.

The Fed’s Crisis Response Has Changed

The other major difference is the intervention playbook.

The Fed entered the financial crisis without the experience it now has using large-scale asset purchases and targeted emergency facilities. QE and the Term Asset-Backed Securities Loan Facility, or TALF, emerged during that crisis: the initial agency-debt and mortgage-security purchase program and TALF were both announced in November 2008.

Those responses established precedents for later intervention.

Today, liquidity analysis has to account for multiple operations. Repo lending temporarily adds reserves to the banking system. Reverse repos temporarily absorb reserves and help manage short-term interest rates. These tools serve different purposes; treating every operation as another form of QE misses how the system works.

The Bank Term Funding Program provides a more recent example. Introduced during the 2023 banking turmoil, it supplied funding to eligible banks to help them meet depositors’ needs. It has since ended, with the final loans repaid in March 2025, but it demonstrates the ability to establish a facility aimed at a particular source of financial stress.

We did have a recession in 2020. The business cycle has not disappeared, and it would be inaccurate to say there have been no recessions since the financial crisis.

My argument is that policy intervention now deserves a central place in market analysis. I expect policymakers to fight aggressively against another cascading breakdown in the financial system, using the tools and authority available to them.

That does not give the Fed unlimited control. Liquidity support cannot erase every credit loss, guarantee asset prices, or remove the constraints created by inflation. But anyone forecasting an exact replay of 2008 needs to explain how the policy response fits into that forecast.

AI Already Has Real Customers and Real Uses

The second comparison I question is the idea that AI is simply another dot-com story.

The internet itself was real during the dot-com boom. It produced genuine technological and productivity advances. The problem was that investment and market expectations could run far ahead of what businesses could deliver. A technology can transform the economy while investors still pay too much for it.

That distinction matters for AI.

We already have evidence of commercial demand. In its fiscal 2026 fourth-quarter earnings release, Microsoft reported more than 30 million paid Microsoft 365 Copilot seats. Those are paying customers using an actual product.

There is also evidence of useful work getting done. A study of 5,172 customer-support agents found that access to an AI assistant increased issues resolved per hour by 15% on average, although the benefits varied considerably across workers. That is a measurable result in a specific workplace setting.

My view is that we are still early in discovering how widely these capabilities can be applied. Businesses need time to change their processes, integrate the tools, and determine which applications actually improve results.

Some companies will disappoint. Some investments will prove excessive. Some stocks may already reflect years of success that have not happened yet.

AI can be early in its development while individual AI stocks are expensive.

That is why I do not think “dot-com 2.0” is a sufficient explanation for what is happening. We need to examine adoption, revenues, costs, competitive advantages, and valuations separately.

Uranus in Gemini: My Cycle Perspective

There is also an astrological cycle perspective behind my outlook.

Uranus first entered Gemini in July 2025 and returned on 04/25/26, following a temporary move back into Taurus. Its previous passage through Gemini occurred, with interruptions, between 1941 and 1949.

That earlier period overlapped an extraordinary acceleration in technology. World War II brought major advances in radar, nuclear research, and computing. Jet aircraft entered combat, while machines such as Colossus helped accelerate the deciphering of German communications. Many of the underlying discoveries began earlier; what stands out is the speed of development and deployment during the war and its aftermath.

In my cycle work, I associate Uranus in Gemini with disruption and innovation in communication, information, and technology. AI fits those themes closely.

That is my astrological interpretation, rather than proof of cause and effect or a guarantee of investment returns. Through that framework, I see the current passage as consistent with a technology cycle that still has considerable room to develop.

Study the Conditions in Front of You

I believe today’s mortgage structure, the Fed’s intervention experience, and AI’s commercial development deserve more attention than simple comparisons with the last two major market crashes.

There will still be corrections. There will still be failed businesses and bad investments. Stronger mortgage underwriting does not prevent every housing decline, and powerful technology does not justify every stock price.

But I do not believe we are watching an automatic replay of 2008 in housing or an empty technological story in AI.

My view is that AI is still in the early stages of a much larger transformation—and that investors who focus entirely on the last crisis risk overlooking the conditions shaping the next cycle.