Michael Hartnett, chief investment strategist at Bank of America Securities, has laid out three specific market thresholds in his latest "Flow Show" report that he believes would trigger broad short-covering across asset classes: the Mag7 ETF falling below $60, USD/JPY dropping under 110, and a reinversion of the yield curve. None of those conditions have been met yet. But the signals building underneath - net outflows of $8.5 billion from U.S. equity funds, the first since March, following a historic $119.2 billion inflow - suggest investors are not waiting around to find out.
That rotation is already visible in where the money is going. Liquidity is moving out of mega-cap tech and into cyclicals: semiconductors, mid-caps, housing, and REITs. Hartnett frames this as a preemptive repositioning, with markets anticipating a policy pivot toward affordability concerns rather than growth maximization. For operators in regulated retail - including cannabis - this kind of macro shift deserves attention. A Michigan dispensary POS platform tracking daily transaction volumes, for instance, can surface exactly the kind of consumer spending compression that follows tightening financial conditions at the regional level, where the macro numbers eventually show up in individual basket sizes and traffic counts.
The underlying stress in tech is harder to ignore than the headline numbers suggest. Hartnett's central question - how much further do cloud equities have to fall before markets start pricing in a capex pullback - points to a genuine tension in AI infrastructure spending. Apple raising MacBook prices, Microsoft increasing Xbox pricing: both moves trace directly back to rising memory costs. Vera Rubin rack memory pricing has increased cumulatively by 435% according to the context Hartnett is working from, and Goldman Sachs projects AI capital expenditure could reach $14 trillion by 2027. That's an enormous number sitting on top of an equity market that is already rotating out of the companies expected to absorb it.
What the Bond Market Is Telling Operators and Investors
Since Fed Chair Powell's confirmation on May 22, U.S. Treasuries have returned 3.2% while equities have declined 1.6%. That divergence is the bond market asserting itself - and Hartnett is listening. His positioning framework reflects it directly: he views long-duration Treasuries as the single most contrarian long trade available right now, gold below $4,000 as still carrying strong allocation value, and the U.S. dollar as a short-term holding only, structurally weak against a bullish backdrop for emerging markets.
For cannabis businesses, this macro context is not abstract. The industry operates almost entirely outside conventional capital markets - no institutional equity support, no access to standard commercial lending, no Federal Reserve backstop when credit tightens. When financial conditions shift and investors rotate defensively, the ripple effects hit cannabis operators through tighter vendor credit terms, constrained wholesale pricing, and slower capital availability for store builds or license acquisitions. The companies that have managed to secure working capital lines or structured debt arrangements find themselves in better shape precisely because they prepared during the window when capital was easier to access.
AI Capex, Memory Costs, and the Retail Technology Budget
There's a more direct operational implication buried in Hartnett's AI capex argument - one that cannabis retail technology buyers should track. The sustained underperformance of mega-cap cloud stocks relative to chip equities reflects a market debate about whether the infrastructure spending underpinning software-as-a-service platforms can hold at current levels. Most dispensary operators rely on cloud-hosted POS systems, compliance reporting platforms, seed-to-sale integrations, and inventory management tools. Those services run on the same infrastructure cost curves that are now under pressure.
The thing is, software vendors in cannabis tech have historically priced modestly to win market share in a capital-constrained industry. If memory and compute costs continue rising - driven by AI infrastructure buildout competing for the same hardware - that pricing pressure eventually passes downstream. Not immediately, and not dramatically in the short term. But operators who are mid-contract or approaching renewal windows should be paying attention to what their vendors' underlying cost structures look like, and whether their contracts include provisions for pass-through cost increases.
Three Thresholds, and the Broader Bet on What Comes Next
Hartnett's three triggers function as a discipline, not a forecast. They tell institutional investors what conditions would justify a major change in positioning, without demanding a prediction about whether those conditions arrive. That's a useful framework for cannabis operators too - not because dispensary owners are trading Mag7 ETFs, but because the discipline of identifying specific thresholds before acting is exactly what separates operators who respond well to market shifts from those who react after the fact.
For a licensed cannabis business, the equivalent thresholds might look different: a specific drop in average transaction value, a margin compression on wholesale flower that crosses a floor, a state excise tax increase that pushes effective tax burden past a threshold that makes certain SKUs unviable. The mechanism is the same. Define the level that changes your decision, watch for it, and have the operational response ready before it arrives - rather than scrambling after the fact when cash flow is already tighter than the P&L suggested it would be.
None of Hartnett's three conditions have triggered yet. But the money is already moving. That's usually how it works.
