The Number Everyone Is Talking About

Year-to-date, retail investors have purchased $22.5 billion of US semiconductor ETFs. That figure has surged more than 1,000% since early April 2026. Last month alone saw $12 billion flow in, the highest single-month total on record.

The narrative writes itself: AI demand is real, chips are the picks-and-shovels play, and the pullback from the highs is a gift. Millions of retail investors are acting on that thesis right now.

The question worth asking is not whether AI demand is real. It is. The question is whether that demand justifies buying into a market where the price structure is telling a very different story.

What Is a Bear Market, Really?

Most people have heard the rule: a stock or index is in a bear market when it falls 20% from its most recent high. That is a useful shorthand, but it only tells you how far something has fallen. It does not tell you whether the fall is finished.

A more useful concept is market regime. Think of it as the personality of the market at a given moment. There are broadly two:

• Markup: Institutions are accumulating. Price trends upward on expanding volume. Each dip gets bought quickly. This is when buying dips works reliably.
• Markdown: Institutions are distributing. Price trends downward. Rallies get sold into. Each bounce is a trap for late buyers. This is when buying dips punishes you.

The critical thing to understand is that the same 20% drop can mean completely different things depending on the regime. A 20% drop in a markup regime is a buying opportunity. A 20% drop in a markdown regime is often just the beginning.

The Tech Bear Nobody Is Mentioning

While the conversation has centred on semiconductors, the broader tech universe contains some extraordinary drawdowns from recent highs. These are not obscure small-caps. They are household names.

These stocks are down between 48% and 69% from their highs. A buy-and-hold investor who entered at peak enthusiasm would need the stock to double or more just to break even. That is the maths of deep drawdowns that most retail participants do not properly account for.

Buying the Dip vs. Catching a Falling Knife

These two phrases describe very different situations, but they can look identical in the moment.

Buying a genuine dip means entering a stock that is in a healthy uptrend after a temporary pullback, usually caused by profit-taking or short-term sentiment. The underlying trend remains intact. Institutions are still buyers on weakness. Price tends to recover quickly.

Catching a falling knife means entering a stock that is in a genuine downtrend, one where the dominant players are sellers. Every bounce feels like a recovery. Every recovery fails. You buy each dip expecting a reversal, and each dip becomes a lower low.

The painful truth is that a stock down 57% feels cheaper than it did at the high. That feeling is real. But “cheaper” is only relevant if the stock actually recovers. In a markdown regime, cheap can become cheaper for months or years.

The difference between the two scenarios is not obvious from price alone. It requires understanding what the underlying regime is doing and, crucially, what the players with the most information are actually doing with their money.

How Regime Analysis Changes the Conversation

Rather than relying on gut feel about whether something “looks cheap”, regime analysis applies a structured framework to every instrument. The goal is to classify each stock as operating in a bullish (markup) or bearish (markdown) phase based on objective criteria, not opinion.

Our screener currently shows 18.2% of stocks in markdown regime. That means nearly one in five stocks in our universe is showing bearish structural characteristics right now. Not just pulling back, but structurally trending lower.

When you overlay that against the retail buying data, the picture becomes clearer. A significant portion of the capital flowing into chip ETFs is going into instruments that the structural data classifies as being in distribution.

You can check the current regime classification for any stock in our coverage universe on the Convergence Screener. The regime read is one of the core layers within the convergence framework.

Why Insider Positioning Matters More Than Retail Flow

There is a meaningful difference between retail money flowing into an asset class and sophisticated institutional money doing the same thing. One is driven by narrative and recency bias. The other is driven by structural analysis, proprietary information, and long-term conviction.

A notable current example: Michael Burry, the investor who became famous for shorting the US housing market before the 2008 financial crisis, is reportedly buying Microsoft 2028 LEAPS. LEAPS are long-dated options with expiry dates two or more years out. They require genuine multi-year conviction, because the premium paid is lost if the thesis does not play out.

What is interesting about this is the timeframe. Burry is not betting on a near-term bounce. He is taking a position that only pays off if Microsoft is meaningfully higher two years from now. That is a very different type of trade from retail capital chasing a momentum story in semiconductor ETFs.

Similarly, clusters of insider buying in beaten-down names, where company executives and directors are purchasing shares with their own money, tend to be a more reliable forward indicator than aggregate retail flow. Insiders buy for one reason: they believe the stock is worth more than its current price. They rarely buy into genuine structural downtrends in their own companies.

Our Insider Tracker monitors these transactions as part of the broader convergence framework, flagging when insider positioning aligns with other bullish structural signals.

How to Apply This Before You Buy Anything

There is a simple three-step process that dramatically improves decision quality before entering any dip buy.

This framework applies whether you are buying a semiconductor ETF, an individual tech stock, or anything else. The process does not change. The data inputs do.

Check Any Stock’s Regime Before You Commit

Our Convergence Screener applies a multi-layer framework to the stocks in our universe, including regime classification, institutional positioning signals, and multi-factor convergence scoring. You can see exactly where any covered stock sits within that framework before making a decision.

Also available: Ethical Screener | Insider Tracker | All Foundry Articles

The Bottom Line

The $22.5 billion flowing into chip stocks may turn out to be the right call. AI infrastructure spending is real, and the long-term secular story remains intact. But capital allocation based on narrative alone, without checking the structural regime, without checking what informed investors are actually doing, and without a clear plan for when the thesis is wrong, is not investing. It is a form of hope dressed up in economic logic.

The stocks in the table above were also compelling stories at their highs. Oracle was an AI infrastructure play. ServiceNow was a productivity revolution. Palantir was the future of defence and intelligence analytics. The stories were real. The prices got ahead of the reality, and the regime turned, and retail investors who did not check those signals are sitting on losses of 48% to 69%.

Check the regime. Check the smart money. Define your exit. That sequence costs you nothing and protects you from the most common and most expensive mistake in markets.