Here’s what makes this different from every other “outlook” you’ve read on the internet: I score myself.

Every prediction from the previous quarter gets a transparent verdict — right, wrong, or still in play.

Since this is issue #1, there’s no scorecard yet. But starting in Q3, you’ll see exactly how this one aged. That’s the whole point — writing under the pressure of knowing we’ll reopen this document in 90 days is what forces honesty.

Before we dive in, one thing I want to make crystal clear:

These are predictions, not investments.

What you’re about to read is my working hypothesis on where I think capital is mispriced and how the next 90 days could unfold. They are starting points for deeper research — not positions I’ve already taken or am recommending you take.

Each of these will need a much deeper dive before becoming a real investment thesis: position sizing, entry zones, hedging, correlation mapping, scenario stress tests, liquidity analysis. That’s the work that comes afterthis document, and it’s what I’ll cover for premium subscribers in dedicated deep-dives over the coming weeks.

Think of this issue as the board on which I’ll be playing — not the moves themselves.

Okay, let’s get into it.

The setup for Q2 2026 is unusually rich.

Hyperscaler AI capex is set to cross $715B this year.

The dollar just broke under 97 for the first time in four years. Copper and uranium are flashing structural deficit signals at the same time.

Europe is rearming at a pace we haven’t seen since the Cold War. And the Bank of Japan finally hiked to 0.75% — the highest level in three decades.

That’s a lot of plates spinning. And historically, when you have multiple regime shifts happening simultaneously, you get extreme dispersion — which is exactly the environment where bold, well-defined predictions either land hard or fail hard. Both are useful.

I’ve ranked the 12 predictions below by risk/reward, from most aggressive to most defensive.

The top of the list contains my highest-conviction asymmetric bets — high potential payoff, but with real probability of being wrong.

The bottom of the list contains ideas I’m more confident in but where the upside is more measured.

That way, you can calibrate based on your own risk tolerance and time horizon.

TIER 1 — High Risk, High Reward

The asymmetric bets. Lower hit rate, but if I’m right the payoff is significant. These are the ones I’ll be watching most carefully.

1. A violent correction (-25% to -40%) on AI hyperscalers during Q2

Let me put some numbers on the table first. Combined capex for Microsoft, Google, Amazon, Meta, and Oracle is projected at roughly $715B in 2026 — that’s about 90% of their combined operating cash flow, up from 65% just last year. To finance this, Morgan Stanley estimates hyperscaler debt issuance will top $400B this year, more than double 2025’s $165B.

Here’s the thing nobody wants to say out loud: capex is now growing roughly 46% faster than AI-related revenue. That gap is wider than the 32% divergence we saw during the 2001 telecom bust. The AI-linked sectors trade at EV/EBITDA multiples near 25x — close to dot-com peak territory. And NVIDIA derives 85% of its revenue from just six customers, with the top four accounting for nearly 60%. That’s binary risk hiding in plain sight.

My thesis: a single hyperscaler cutting 2027 capex guidance — or two consecutive quarters of enterprise AI spending decline — triggers the cascade. The market is pricing perfection. Anything less than perfection is the catalyst.

What would make me wrong: a measurable acceleration in enterprise AI revenue (north of $100B annualized), or four out of five hyperscalers raising 2027 capex guidance in lockstep.

Confidence: 35–45%. Asymmetry: roughly 5:1 if timed correctly.

2. Uranium re-prices structurally toward $130–150/lb

Uranium has already pushed past $100/lb on spot in January, but here’s where it gets interesting: long-term utility contracts are still lagging at $86/lb. Sprott estimates global mine supply will cover less than 75% of future reactor demand. US uranium production fell 44% in Q3 2025. And roughly 70% of post-2027 demand remains uncontracted — the highest level recorded in three decades.

Meanwhile, hyperscalers are signing nuclear PPAs (Microsoft–Constellation, Three Mile Island restart). This isn’t speculative anymore — it’s contracted demand showing up in the data.

The market is still treating uranium as a cyclical commodity. But the structural dependency of AI data centers on carbon-free baseload (90%+ capacity factor, no other source comes close) creates inelastic, multi-decade demand that legacy supply/demand models simply don’t capture.

What would make me wrong: an accelerated restart of Cigar Lake or Inkai combined with a Fukushima-style accident that kills sentiment.

Confidence: 60%. Asymmetry: roughly 3:1.

3. The yen rallies to USD/JPY 140 within 90 days

The BoJ just hiked to 0.75% — the highest since 1995 — and effectively buried Yield Curve Control. Morgan Stanley estimates roughly $2 trillion in yen carry positions are still open globally. As the Fed–BoJ rate gap compresses (Fed median dot at 3.375% by year-end vs BoJ likely 1.25%+), the carry math gets thin.

Here’s what people miss: carry trades don’t unwind because the absolute spread becomes negative. They unwind because path risk becomes intolerable. Volatility on the funding leg goes up, hedging costs go up, and once one big book starts unwinding, everyone else has to follow before the door closes. The August 2024 episode (when BTC dropped 24% in 48 hours on a surprise BoJ move) is the template.

What would make me wrong: BoJ pauses after a weak GDP print, while the Fed holds firm. That keeps the trade alive.

Confidence: 55%. Asymmetry: roughly 4:1 via options structures.

4. A private credit crisis breaks out in mid-market BDCs

Private credit has exploded to over $2T in AUM. Public BDCs trade at NAV discounts that imply real stress, but here’s the part most analysts get wrong: the discount isn’t because of stress, it’s anticipating it. The portfolio composition hides cumulative defaults that monthly NAV markings on private vehicles are smoothing over. The Fed is staying high. The 2026 leveraged loan refinancing wall is real. And the first cliff event is going to be a non-traded BDC gating redemptions.

Once one major non-traded BDC gates, the public BDCs get marked aggressively, and the whole sector reprices.

What would make me wrong: three or four back-to-back Fed cuts that decompress refinancing costs, plus stable EBITDA at LBO-backed portfolio companies.

Confidence: 40%. Asymmetry: roughly 6:1 on a basket short.

Free tier ends here.

  What's behind the wall:

  → Predictions #5, #6, #7, #8, #9, #10, #11, #12 (Tier 2 - Strong Reward, Tier 3 — defensive plays)
  → Specific instruments for each prediction (tickers, options
    structures, futures contracts, ETFs)
  → Entry zones, stop levels, position sizing framework
  → Detailed invalidation scenarios with hard numbers
  → Cross-cutting catalysts to watch (Fed transition, Iran,
    hyperscaler earnings)
  → Bi-weekly mid-quarter updates as theses develop

  → Subscribe to unlock.

TIER 2 — Medium Risk, Strong Reward

Read more