Tycho Arete Macro Fund

Overview

Tycho Arete Macro Fund aims to deliver competitive risk-adjusted returns while maintaining a low correlation with all major asset classes. It strives to construct and update macro-analytical frameworks that incorporate the rapidly changing macroeconomic conditions around the world, as well as the significant idiosyncrasies of large global actors such as China and Japan.

Strategy & Manager

Performance

Commentary

Investment Manager’s Commentary – June 2026

June saw volatility return to the AI trade with a vengeance. After a stretch in which every dip was a buying opportunity, the semiconductor and memory complex buckled under a series of sharp, circuit-breaker-tripping selloffs, the most violent of them in Asia. The triggers were several — unease over how the hyperscaler build-out is financed now that capex has outrun cash flow, Chair Warsh’s hawkish debut and the discount rate it implied, and a wave of fresh AI supply into a nervous market. None of it, in our reading, fundamentally changes the underlying demand for compute, but it was more than enough to unsettle a market positioned for one-way travel.

The month’s defining feature was divergence. The damage concentrated precisely where positioning was most crowded and most levered — Korea above all, where two memory champions dominate the index and single-stock leverage magnifies every move. China’s onshore complex – most notably China’s semi and hard-tech index, STAR50 – proved a notable haven: more domestically funded, less hostage to foreign flows, and outside the leveraged machinery that amplified the selling elsewhere.

Our PnL reflected that barbell. The largest contributor was short gold, which gained as Warsh firmed the dollar and lifted real-yield expectations — a clean expression of the thesis set out below. The insulated East did much of the rest: our China AI hardware and A-share longs, STAR 50 foremost, advanced even as the regional tape sold off, while our long in global AI hardware held its ground — a reminder that the demand story and the funding story are not the same thing.
On the other side of the ledger, several H-share sector baskets – such as China financials and internet, being more sensitive to weakening internal demand – gave back ground even as the broad onshore indices rose, a reminder that “China” is rarely a single trade.

Current Outlook

June reminded us that conviction and humility are appropriate bedfellows in volatile times. Sharp moves punctuated the month throughout and we expect this action to continue as leverage and concentration continue to increase across markets. Yet volatility of this kind tends to obscure rather than alter the structural picture. We examine, in turn, the month’s key developments — Warsh’s debut and the turbulence across AI — and whether they changed the big picture.

The Fed Under Warsh: A Stylistic Shift

Kevin Warsh’s first meeting as Fed Chair marked a new chapter, and markets approached it preoccupied, as ever, with the path of the policy rate. We think the exercise is largely fruitless. Warsh has been explicit that policy is data-dependent and that reducing forward guidance is central to his philosophy, which means the path is only as legible as the data — and the data has rarely been noisier. Backward-looking prints still carry the tariff impulse; forward-looking expectations point the other way, toward disinflation, with sharply lower oil draining much of the second-round risk. The more consequential development was stylistic — the deliberate retreat from guidance, and the beginning of a different kind of Fed.

Consider the balance sheet. Little was offered explicitly, but Warsh’s advocacy for a structurally smaller footprint is well established. This matters less for liquidity mechanics than for narrative: a disciplined central bank is a powerful antidote to the debt-unsustainability anxiety that formed one of the central pillars of last year’s dollar-bear thesis.

The second shift is towards saying less. Counter-intuitively, a quieter Fed may prove better in communications. Policymakers have always treated the dot plot and their speeches as conditional on the data; to the extent markets receive them as commitments, the gap becomes a source of miscommunication. A Fed that says less also leaves less room for the miscommunication gap, reinforcing the very credibility and independence whose perceived erosion was the other pillar of the dollar bear. Remove both pillars, and the bear case for the dollar is left standing on rather little.

This dovetails neatly with what remains our most out-of-consensus position: short gold. Our case rests on the G2 framework. A world investing more — across both dominant poles, the US and China — is a world raising total factor productivity in the two economies that matter most. Higher TFP lifts real yields, lifts both the dollar and the renminbi, and steadily erodes the appeal of a non-yielding store of value. A more credible, more disciplined Fed accelerates every link in that chain.

AI Hardware: Is the Trade Over?

The semiconductor and memory complex lurched through a series of violent swings, and the consensus reaction was swift and familiar: has the AI trade finally run its course? To answer, begin with what actually moved — and why.
The cause was technical — a story about how the build-out is financed, not the degree to which it is in demand. Stripped to first principles and put simply: where does the money come from? For most of this cycle the answer was internal cash flow; that is no longer true. Hyperscaler capex has now caught up with operating cash flow, so the next leg must be financed — and a meaningful share of end demand rests on a handful of still-unprofitable model labs whose ability to keep raising capital is itself the swing variable. A build funded from cash flow is indifferent to the cost of capital; one funded by issuance and by faith in the next funding round is acutely sensitive to both. That is the thread tying this month’s air pockets back to Warsh.

Positioning did the rest. The sharpest moves originated in Korea, where two memory champions make up half the index, single-stock leveraged products magnify momentum both ways, and reversing foreign flows turned an orderly pullback into a rout. That is a statement about market structure, not demand. China’s complex — more domestically funded and largely outside that machinery — sat notably apart: China’s STAR50 – the Shanghai-listed index largely comprised of semiconductor and tech hardware – was up 26% in June.
So — is the trade over? The most relevant evidence points the other way: demand is the foundation, not the fault line. Today’s chipflation is fundamentally demand-led, not a supply shock — HBM is effectively sold out, the DRAM shortage is the most acute in over a decade, and pricing power now extends visibly into 2027. Inelastic demand meeting inflexible supply is what turns a cyclical upswing into a structural reset, and the June tape did not lay a finger on any of it.

And yet it would be complacent to call the trade unchanged. It now rests on a belief — that scaling laws hold and model capability keeps advancing — because that progress, and only that, keeps the unprofitable end of the demand base fundable. Where could the belief crack? The physical build is slower and more stubborn than the order books imply, constrained less by ambition than by power, skilled labour, and local opposition. The model layer has yet to demonstrate a durable moat — leadership has changed hands repeatedly within months, and switching costs are near zero. And token economics may deflate faster than volume compounds. The tell is simple: while price and volume rise together, the cycle is healthy; the day it turns to falling price on rising volume — or the frontier stops astonishing — the belief, and the funding chain beneath it, begins to waver.

We are not there yet, and the longer arc may prove benign. As the unit cost of reasoning falls — through added capacity or competition — the addressable market expands rather than contracts, and for East and West alike. This is the Jevons paradox applied to compute, and the channel through which China, advancing frontier capability at a fraction of the Western cost (GLM-5.2 the latest marker), becomes a beneficiary rather than a casualty. The caveat: this holds only if demand proves elastic enough to absorb the cheaper unit, and in the interim cheaper intelligence can open air pockets in hardware even as the long-run pie grows. In short, a trade we still want to own — but no longer one we will own unhedged.

Where Does This Lead Us?

On balance, the month leaves our structural convictions intact and, in places, strengthened. We hold increased conviction in our short gold, and maintain our short in US Treasuries and our long investment thesis across AI and renewable energy. What has changed is not the thesis but the terrain beneath it. We are adding hedges to our longs — not because we doubt the demand, which has if anything been reaffirmed, but because the architecture funding it has grown more leveraged, more circular, and more crowded, held increasingly in vehicles that amplify rather than absorb shocks. The view is unchanged; the environment simply demands we carry it with more ballast.

Documents