Correlation is not causation

Research evidence excerpt

Correlation is not causation

oin in late May. Thirty-two.

Worth $2.5mn. A rounding error for a company sitting on over 847,000 BTC. But the symbolism

was stark. Strategy's entire investment thesis was built on a public promise never to sell. When

they sold — even a minuscule amount — and then announced a new policy framework allowing

further sales for "capital allocation purposes," it was a significant hit to sentiment. CEO Le

further stated that “We will sell bitcoin when it's advantageous to the company" and said the

company would actively manage its balance sheet than hold Bitcoin forever.

Analysts flagged the change immediately, warning that the shift introduced selling risk into a

market that had priced Strategy as a permanent buyer. And the ripple effects went well beyond

one company. Bitcoin ETFs saw record outflows in June. Institutional investors had already cut

positions in Q1. Higher yields and a hawkish Fed did the rest. This is a high-beta risk asset (that

routinely sees stomach-churning volatility) losing its single most important narrative — the

"never sell" covenant by a very large buyer — at exactly the moment the macro backdrop has

turned hostile.

The Magnificent 6: asset-light no more

This is the cleanest story of the five – and largely unrelated to Fed expectations. For two

decades, Big Tech's pitch to investors was capital efficiency. Asset-light businesses. Software

margins. Enormous free cash flow with minimal physical infrastructure. That pitch died in 2025

and was buried in 2026. Big Tech will spend over $700bn in AI capex this year, and a trillion more

next year – more than most countries’ budgets. These are the kinds of numbers you associate

with the very largest oil and mining companies, and not with firms that started life selling books