DarkStone Capital Research

The defining flow story of 2026 is a great rotation — capital migrating from the most concentrated, growth-rich pockets of the U.S. equity market into broader, cyclical, and increasingly international assets. The cap-weighted S&P 500 has gone roughly sideways year-to-date while the equal-weighted version has been touching new highs, a divergence that maps cleanly onto a multi-pronged reallocation across sectors, asset classes, and geographies. For three years, owning the index meant owning seven names. In 2026, that trade is being unwound — slowly, deliberately, and with a clear destination.

Out of Mega-Cap Tech, Into the Real Economy

The most visible flow is sectoral. Technology, which carried 2025's gains on the back of the AI infrastructure build, has stumbled: the Nasdaq 100 is down roughly 6% year-to-date, while real-economy sectors are leading by wide margins. Industrials are up more than 16%, consumer defensives 13%, and energy more than 22% YTD. Caterpillar, Walmart, Costco, Exxon, and Chevron have been the largest single-name contributors to index returns — a roster that would have looked unfashionable just twelve months ago.

The catalyst stack is unusually clean. The "One Big Beautiful Bill Act," signed into law in January, makes permanent the 2017 corporate tax cuts and is projected to drop corporate tax liabilities by roughly $129 billion across 2026 and 2027. That fiscal impulse landed simultaneously with a geopolitical shock: the Iran conflict that escalated in late February pushed crude through $65 in the spring and re-introduced an inflation-via-energy channel that markets had largely written off. Fiscal stimulus plus a commodity tailwind, applied to a tape priced for tech exceptionalism, was always going to bias flows toward cyclicals. It has.

The reallocation extends down the cap structure. Small caps are outpacing large caps for the first time in years, helped by lower funding costs and a closing earnings gap between the Magnificent Seven and the broader market. Within tech itself, leadership inside the AI trade has shifted from end-user software toward infrastructure beneficiaries — meaning the AI dollars are still flowing, but they're moving down the stack into power generation, copper, cables, transformers, and the industrial supply chain that physically builds out compute capacity. If 2024 and 2025 were about the code, 2026 is about the copper, the cables, and the kilowatts.

Inside Fixed Income: The Defensive, Short-Duration Bid

Bond flows tell a more cautious story than equity flows, and they're worth reading carefully because they're often the first place institutional positioning shifts.

In the first two weeks of March, fixed income ETFs represented over 75% of total ETF flows as cross-asset volatility picked up. The duration mix was telling: roughly $28 billion into short-term government exposures against $2.7 billion of outflows from long-term governments, with another $1.3 billion into inflation-linked bonds and roughly $5 billion of outflows from growth-sensitive credit, including high yield and EM debt. The message is consistent across each line of that profile. Investors want to be paid to wait at the front end of the curve while the long end struggles with sticky inflation, rising term premium, and deficit concerns. The rolling three-month flow into long-duration Treasuries has fallen to one of the lowest readings on record.

This is the unglamorous half of the rotation, but arguably the most important. A persistent bid for T-bills tells you the marginal allocator is positioning for optionality, not duration. That's a regime where capital is rotating not just betweenequity sectors but out of long-duration assets entirely — including the longest-duration equity exposures, which is to say mega-cap growth.

The Hard-Money Decoupling: Gold vs. Bitcoin

Perhaps the most analytically interesting flow story is inside hard money. Gold was the unambiguous crisis-hedge winner of the prior cycle, rallying from roughly $2,600 to over $5,500 per ounce between early 2025 and early 2026 — a more than 100% move in twelve months. That run is now experiencing classic profit-taking on top of dollar-strength and elevated-real-yield headwinds.

Bitcoin has moved in the opposite direction. It is down roughly 20% year-to-date, falling from approximately $93,000 at the start of the year to around $74,000 by mid-April, even with CPI at 3.3% YoY in March. The result is a Bitcoin–gold correlation coefficient near -0.88, one of the most negative readings in years.

This has quietly reframed institutional positioning. Gold is increasingly being treated as systemic tail insurance — the asset central banks and sovereign wealth funds buy. Bitcoin is being treated as a high-beta liquidity asset that trades alongside the Nasdaq, with its own halving-cycle and ETF-flow dynamics, but firmly inside the risk-on bucket. They are no longer competing for the same dollar. Gold ETFs saw over $6 billion in net Q1 outflows against $12 billion of inflows in the year-ago period; Bitcoin ETPs also saw Q1 outflows, a notable slowdown from last year's pace. The barbell strategy — gold for systemic protection, Bitcoin for debasement and liquidity exposure — remains intact, but capital is rebalancing inside it rather than between asset classes.

International: The Quiet Rebalance Out of U.S. Concentration

Cross-border flows reinforce the same broadening narrative. Emerging and frontier markets returned roughly 30% and 41% in USD terms over the first eleven months of 2025, and the MSCI EM Index is up 7% year-to-date in 2026 with consensus EM earnings growth expected at 29% versus 14% for the U.S.

The driver isn't a wholesale exit from American assets — it's a marginal reweighting at the edges. Dollar weakness over the past year has made non-U.S. assets cheaper in home-currency terms, and European and UK allocators in particular have grown vocal about U.S. concentration risk after years of being underweight. Inside EM itself, the flows are concentrated in Asian semiconductor exposure (Korea and Taiwan, where the AI hardware earnings impulse runs directly), privately owned Chinese internet and AI platforms building a ring-fenced domestic ecosystem, and select Latin American beneficiaries of the commodity bid.

For an asset class that represents roughly 40% of global GDP but only 11% of the MSCI ACWI, even a small normalization in positioning is a meaningful flow. That's the trade.

The Unifying Thesis

Stitch the four strands together and a clean picture emerges. Capital is leaving the most expensive, most concentrated, longest-duration parts of the market — mega-cap U.S. tech, long Treasuries, recently-extended gold — and migrating into shorter-duration, more cyclical, more globally distributed assets: real-economy sectors, small caps, T-bills, EM equities, and the AI supply chain rather than the AI brand names.

2026 isn't the year the AI trade dies. It's the year the market stops paying for the idea of AI and starts paying for its physical inputs. Hyperscaler capex is now tracking toward $650 billion or more for 2026, up from $520 billion at the start of earnings season. That spend has to land somewhere, and increasingly it's landing in industrials, energy, materials, and the EM semiconductor complex rather than in the multiples of a handful of mega-cap names.

The risk to this thesis is symmetric. Energy, materials, industrials, and staples are now stretched well above their 50-day moving averages and into overbought territory, which means the rotation could exhaust quickly if inflation cools or the front end of the curve reprices lower. And the geopolitical tape — Iran, Venezuela, dollar dynamics — is doing as much work as the fundamentals. None of that is durable in the way structural earnings convergence would be.

Which leaves the question worth holding open: is this a regime change in market leadership, or a tactical unwind of a single crowded trade dressed up as a structural rebalance? The answer probably arrives in Q2 earnings — specifically, whether non-tech earnings growth converges with the Magnificent Seven on a sustained basis, or whether the cyclical bid fades the way every reflation trade since 2010 eventually has.

For now, the flows are unambiguous about what the marginal dollar is doing. The harder question is whether the marginal dollar is right.