Follow the Money: Where Capital Has Gone in 2026 — and Where It's Running Out of Places to Hide

The story of 2026's first quarter isn't a crash. It's a rotation that ran out of road.

Money left tech in January. It found defensives and gold. Then the Iran conflict sent oil surging and rewrote the rules. The defensives started bleeding. Gold — the ultimate safe haven — posted its worst week since 1983. Bonds offered no shelter with rates on hold. And now, at the end of March, the data shows something uncomfortable: capital is running out of places to hide.

We track flow, momentum, extension, and volatility percentiles across 30 ETFs covering every major sector and asset class. Here's what three months of rotation looks like when you watch the money move in real time.

Phase one: the tech exodus (January)

The year opened with money leaving growth. Tech (XLK) saw flow fall from the 67th percentile on January 6 to the 44th by February 3. The Nasdaq (QQQ) dropped from the 60th to the 50th. Communication services (XLC) followed. Microsoft's earnings on January 28 — massive capex, Azure growth flat — accelerated the exit.

Where did it go? Consumer staples (XLP) absorbed the first wave. Flow surged from the 71st percentile to the 97th by January 20, then hit 99 by early February. The extension percentile tells the same story — XLP went from the 34th to the 99th in two weeks. Money was piling into Procter & Gamble, Coca-Cola, Walmart at an extraordinary rate.

Utilities (XLU) caught the second wave. Flow climbed from the 62nd to the 87th by mid-February. Extension went from the 29th to the 93rd. Gold (GLD) joined the trade — extension hit the 95th percentile on January 20 as the metal approached its all-time high above $490.

This was a textbook risk-off rotation. Money leaving high-beta, high-multiple growth stocks and flowing into low-volatility, dividend-paying defensives and precious metals. The rotation was orderly. Volatility across the market was low — the S&P 500's volatility percentile was at the 12th on January 6, and even by February 3 it was only at the 18th. The exit from tech was calm, methodical, and rewarded the traditional hiding places.

Phase two: the oil shock (late February – early March)

Then the Strait of Hormuz happened.

Oil (USO) had been quietly strengthening since mid-January — flow climbing from the 46th to the 91st by February 3. When the US-Iran conflict escalated, crude went vertical. Extension hit the 99th percentile by early March. Flow was at the 94th. Energy (XLE) followed — extension at the 99th, flow at the 88th, trend score at 90.

This wasn't a rotation. It was a regime change. Oil above $100 per barrel rewrote the macro picture overnight. The Fed couldn't cut rates because energy-driven inflation was reigniting. The dollar strengthened (UUP's extension went from the 49th to the 94th by March 3). Real yields rose. And the assets that had been absorbing the risk-off flow suddenly had a problem.

Gold, the classic inflation hedge and safe haven, started falling. Not gently. Extension went from the 55th in early March to the 26th by March 17 to the 2nd by late March. Flow collapsed from the 37th to the 4th. An 18.5% drawdown from the January peak. The worst week since 1983. The safe haven broke.

Why? Because the mechanism that should drive gold higher (geopolitical crisis, inflation) was overwhelmed by the mechanism that drives it lower (dollar strength, rising real yields, leveraged liquidation). Futures traders on margin got squeezed. The paper market overwhelmed the physical market.

Phase three: nowhere to hide (mid to late March)

This is where the rotation story becomes genuinely concerning.

The defensives that absorbed January's risk-off flow have given it all back. Consumer staples (XLP) flow has collapsed from the 99th percentile in early February to the 30th now. Extension went from the 99th to the 23rd. Utilities (XLU) flow dropped from the 86th to the 36th. Healthcare (XLV) went from the 61st to the 32nd.

These weren't just modest pullbacks. The defensives that seemed like safe shelter three months ago are now falling as hard as the growth names they were supposed to protect against. Consumer staples is down 8.4% in the past month. Healthcare is down 7.3%. Utilities is down 4.5%.

Meanwhile, the tech selloff has continued but with a notably different character. XLK's flow has stabilised around the 44th percentile — not strong, but not collapsing. Its volatility percentile is at the 22nd — one of the calmest readings in the entire market. Tech has been selling off, but in an orderly, low-volatility fashion. This matters — our research across individual stocks shows that calm oversold conditions in quality companies tend to resolve upward, while volatile oversold conditions are more ambiguous.

The fixed income market offers no sanctuary either. Long-term treasuries (TLT) had a brief moment in February — flow surged to the 93rd percentile on February 17 as the initial risk-off trade played out. But as rate cut expectations evaporated, the bond rally reversed. TLT's flow is back to the 41st. TIPS (inflation-protected bonds) are at the 2nd percentile extension — deeply compressed with flow at the 32nd.

The current state of the market

Here's the snapshot across every major asset class as of late March 2026, ranked by flow percentile:

Three things stand out immediately.

Finding one: energy is the only story with strong flow AND extension

Energy (XLE) and oil (USO) are the only assets where flow, extension, and momentum are all elevated. Energy's flow at the 85th percentile is the highest of any sector. Extension at the 93rd and momentum at the 85th confirm the trend is running hot. And crucially, energy's volatility is at the 30th — calm. This is a trend in control, not a blow-off top.

Every other sector with positive year-to-date returns is seeing deteriorating flow. Utilities is up 6% YTD but flow has collapsed from 87 to 36. Consumer staples is up 4.9% but flow has gone from 99 to 30. These are sectors living off past momentum — the money that pushed them higher has left, and the price hasn't fully adjusted yet. When extension and flow diverge like this, the price typically catches down to the flow.

Finding two: the defensive rotation is exhausted

The January to February risk-off trade into defensives has completely unwound. This is visible in the flow trajectory:

Consumer staples flow went 71 → 98 → 99 → 87 → 51 → 33 → 30 over the quarter. A spike to the 99th, then a persistent decline back to below where it started. Utilities followed a similar arc: 62 → 68 → 71 → 87 → 86 → 68 → 36. Healthcare never fully participated: 80 → 50 → 42 → 61 → 37 → 34 → 32.

The interpretation: the defensive rotation wasn't a structural reallocation. It was a panic trade that reversed once the macro picture (Iran, oil, inflation) made defensives as vulnerable as everything else. When oil is driving inflation higher, consumer staples companies face rising input costs. When rates stay elevated, utilities with their heavy debt loads suffer. The hiding places stopped working.

Finding three: volatility is elevated almost everywhere

This is the most concerning signal in the data. Look at the volatility column. Out of 29 assets tracked, 19 have volatility above the 50th percentile. The S&P 500 is at the 73rd. The Dow is at the 74th. Emerging markets at the 85th. Industrials at the 81st. Silver at the 84th. Gold at the 90th. Oil at the 98th.

The only assets with genuinely low volatility are tech (XLK at the 22nd), crypto (Ethereum at the 41st), the US dollar (UUP at the 44th), and consumer discretionary (XLY at the 46th).

When volatility is elevated across nearly every asset class simultaneously, it tells you the market is in a regime of broad uncertainty — not a rotation between sectors, but a repricing of risk itself. In a healthy rotation, volatility is elevated in the sectors money is leaving and low in the sectors it's entering. Right now, volatility is elevated almost everywhere, which means money isn't confidently rotating anywhere. It's just anxious.

What the data says about what comes next

The percentile framework wasn't designed for macro calls — it describes positioning, not direction. But the current configuration across asset classes does echo patterns from the past.

The closest analogue in our data is the October to December 2022 period. Broad equity indices were compressed (SPY extension below 20). Defensives had been bid up and were rolling over. Volatility was elevated broadly. Energy was the relative leader. The dollar was strong.

What happened next: the S&P 500 bottomed in October 2022 and began a rally that lasted, with interruptions, for over two years. The resolution came when inflation expectations stabilised, the Fed signalled its rate path, and the uncertainty premium began to compress.

The current environment has a similar structure but with one critical difference: there's an active military conflict driving energy prices. The 2022 resolution was driven by clarity on monetary policy. The 2026 resolution likely depends on clarity on the geopolitical situation.

Where the opportunities might form

If you're trying to position for what comes next, the data highlights three areas to watch.

Tech (XLK) is the calmest oversold sector. Extension at the 18th, momentum at the 25th, flow at the 44th — depressed but not collapsed. Volatility at the 22nd — by far the lowest in the equity universe. This is the "calm compression" pattern that our individual stock analysis (on Google, specifically) found precedes recoveries with an 80% hit rate at 63 days. The sector is quietly compressing while everything else is chaotic. If the geopolitical situation stabilises, tech's calm compression could resolve upward sharply.

Gold (GLD) is the most extreme single-asset reading. Extension at the 3rd, momentum at the 1st, flow at the 4th. We covered this in depth earlier this week — the historical pattern is a short-term bounce (83% hit rate at 21 days) but an uncertain 63-day outlook that depends on whether volatility (currently at the 90th) declines. Gold will likely bounce from here. Whether the bounce becomes a new leg up or fades depends on whether the forced selling exhausts itself.

Bonds (TLT, IEF) are in a state of limbo. Flow to mid-term treasuries (IEF) is at the 68th — the strongest reading in the fixed income space. But extension is at the 9th and momentum at the 4th. Money is flowing in, but it hasn't translated to price recovery yet because rate expectations are working against it. If the Iran conflict triggers a genuine economic slowdown (rather than just an inflation spike), bonds could rally hard. The flow is already positioning for that possibility.

What's actually happening

Strip away the narratives about AI spending and software disruption and geopolitical risk, and the money flow data tells a simpler story.

Capital spent January looking for shelter from a tech repricing. It found consumer staples, utilities, and gold. Then the Iran conflict changed the macro regime — oil surged, inflation expectations rose, rate cuts evaporated, the dollar strengthened. The shelter trades reversed because rising energy costs and sticky rates are bad for defensives too. Now capital is sitting in energy (the direct beneficiary of the conflict) and otherwise has nowhere to go with confidence.

The resolution depends on one thing that no percentile framework or technical indicator can predict: what happens next in the Persian Gulf. If the conflict de-escalates, oil drops, rate cut expectations revive, the dollar weakens, and the risk-on trade returns — led by the calmly compressed tech sector. If it escalates further, the current pattern of broad volatility and diminishing hiding places intensifies.

The data can't tell you which scenario unfolds. But it can tell you where the market is positioned for each one. And right now, the positioning is clear: almost everything is oversold, almost everything is volatile, and the only confident flow is into the asset class (energy) that benefits directly from the crisis. When that's the state of the market, the next major move will be driven by the event, not the technicals.

Watch oil. Watch the dollar. Watch tech's volatility. The answers are there.

Data as of March 25, 2026. Analysis based on percentile indicators across 30 ETFs from November 2021 to present. Percentile calculations use a rolling 365-day window. This is not investment advice — it is a description of current positioning across asset classes, based on our indicator framework.