Special Coverage: Greenwich Economic Forum 2025

When global capital markets are immersed in the intertwining narratives of AI frenzy, soaring gold prices, and rate cut bets, Steve Sosnick, Chief Market Strategist at Interactive Brokers, delivered a sobering reminder at the Greenwich Economic Forum on October 8, 2025.

He pointed out that while the U.S. government shutdown has "almost no impact" on markets in the short term, since the tech stocks dominating major indices are unaffected by government shutdowns, the real risks are spreading elsewhere. These include extreme concentration in tech stocks, circular funding in the AI industry, and investors blindly following momentum and greed.

"AI is heading in the right direction, but the pace is too fast," he said. "This reminds me of the 'bandwidth bubble' of the late 1990s. The world indeed needed bandwidth then, but those companies laying fiber optic cables ultimately didn't make money. I'm not suggesting AI companies are engaging in fraudulent behavior, but rather issuing a warning. Even when the direction is correct, capital allocation can become severely imbalanced in the short term."

Interview Transcript:

Q: In your panel discussion, you mentioned that the U.S. government shutdown has "almost no impact" on markets. Could you elaborate? Under what circumstances might such a shutdown cause greater shock to U.S. markets?

Steve Sosnick: The context of my statement was "almost no short-term impact." The reason is simple. A government shutdown doesn't immediately affect most people's daily lives. The real substantial impact occurs when government employees don't receive their paychecks on the 15th, or when TSA personnel don't show up for work, causing flight delays. That's when the public truly feels the consequences of a "government shutdown."

However, current market attention is completely occupied by tech stock performance and Federal Reserve rate cut expectations. Even if data releases are temporarily interrupted, the Fed might choose to wait and see, but overall, U.S. corporate activity and investment sentiment continue to operate "as usual," which is why stock markets haven't reacted significantly.

Q: You also mentioned that U.S. stock markets are highly concentrated in tech giants, such as the so-called "Magnificent Seven." Does this concentration make markets more fragile? Are current valuations already too high?

Steve Sosnick: I believe risks are indeed rising. Here's an analogy: if everyone crowds in front of one exit, and when evacuation is needed, everyone wants to rush out simultaneously, problems will definitely arise. Markets work the same way. When all capital is concentrated in one group of tech stocks, if selling signals appear, you'll find no one willing to buy.

We saw a preview of this in April. During market volatility triggered by tariff news, the "Magnificent Seven" fell the hardest, followed by the Nasdaq 100, then the S&P 500. Sectors that rise the most often fall the hardest. Although markets quickly rebounded afterward, this concentration risk still exists.

The market's most panicked moment occurred because President Trump once showed an attitude of "not caring about the stock market," which made investors very uneasy. They originally thought he would always be "pro-market." When he returned to a "market-friendly" stance, the rally quickly resumed. This shows that whether it's stock markets or subway cars, being too crowded is problematic.

Q: Returning to AI. On stage, you compared the current AI boom to the "internet bubble" and even mentioned "Enron-style circular trading." Can you help us understand the current funding logic in the AI industry?

Steve Sosnick: This is exactly what concerns me. There's a lot of "circular trading" in the AI industry. For example, OpenAI signed a trillion-dollar chip procurement contract with NVIDIA, but the funding comes from NVIDIA's investment in them. This makes the accounting data look attractive, but the cash flow isn't real.

Looking at OpenAI's partnerships with AMD and Oracle, the same issues exist. Where do these companies get hundreds of billions of dollars to pay for computing power? According to estimates I've seen, OpenAI's current annualized revenue is about $12 billion, while cash burn is close to this level. To support such capital expenditure, they either need to grow revenue to over $60 billion in the coming years or continue raising funds and issuing debt.

Some institutions predict that capital expenditure in the AI industry over the next few years could reach $500 billion to $1 trillion. These are staggering numbers, meaning global "idle liquidity" might be absorbed by the AI industry.

More troubling is that this hardware construction might quickly become obsolete. We might see "overcapacity" in the short term. This reminds me of the "bandwidth bubble" of the late 1990s. Bandwidth was indeed needed then, but those companies laying fiber optic cables didn't make money, and many went bankrupt, like Global Crossing.

I'm not suggesting AI companies are engaging in fraudulent behavior, but rather issuing a warning. Even when the direction is correct, capital allocation can become severely imbalanced in the short term.

Q: For investors, considering the current situation of highly concentrated tech stocks, elevated valuations, and lurking risks, how should they allocate assets?

Steve Sosnick: This is exactly why gold prices have surged recently. People are seeking safe-haven assets. Some funds have also flowed into cryptocurrency markets. While I still question their practical utility, as long as markets believe they can "preserve value," they can serve a psychological hedging function.

My advice to investors is: "Don't fight the market, but insure against it."

This means you can't predict when markets will peak, and it's difficult to profit from shorting because trends might last longer. But you can reduce risk through hedging or diversification. For example, allocate to companies with abundant cash flow that can genuinely support dividends.

Of course, these companies are usually "boring," without stories like AI, but they're stable. Investing, like diet, requires balance.

Two emotions are prevalent in current markets: "MOMO" (Momentum trading) and "FOMO" (Fear of Missing Out).

Trend investing is indeed simple and effective, but if you only look at prices without fundamentals, losses will be greater when turning points arrive. Sometimes, investors need to learn to "eat vegetables" instead of only chasing desserts.