Modern
Weimar.
Hyper
Gamble.

Modern Weimar runs on the printer. M2 expansion is no longer a crisis tool — it is policy. Every fiscal regime in the developed world is structurally short discipline and structurally long obligations. Something has to give, and historically what gives is the unit of account.

Bitcoin is the cleanest expression of that trade. 21M cap, no issuer, no jurisdiction. ETF rails opened it to institutional balance sheets in 2024. Sovereigns are quietly accumulating. Microstrategy has shown corporates how to lever it. The asymmetry remains: a small allocation that goes to zero is survivable; a small allocation that 5x's is portfolio-defining.

Position is core, sized to be held through 60% drawdowns. Not a trade. A regime hedge.

Saylor figured out something most CEOs missed: in a debasement world, the highest-return use of corporate capital is not buybacks or capex but converting fiat into hard assets. MSTR is the public market's only real way to express that institutionally — a levered, optionable proxy on BTC with a perpetual ATM.

The convertible structure is the magic. Issue paper at low coupons, buy BTC, watch NAV expand, repeat. The premium to NAV oscillates wildly — the trade is to add on compression and trim on euphoria.

Risk: dilution overhang and the eventual day BTC has a 70% drawdown while debt is rolling. Sized accordingly.

If BTC is digital gold, SOL is the casino floor. It runs the meme economy, it hosts the on-chain prediction markets, it processes more transactions than every other L1 combined. In a world where attention is the only durable asset, SOL is closest to where attention transacts.

The thesis isn't just throughput — it's that consumer crypto is a power law and SOL has won this cycle's distribution war. Every retail-facing app that matters in the next 24 months ships on Solana first.

Sized smaller than BTC because the failure modes are real: chain halts, validator centralization, and the perennial possibility that consumer crypto is a rolling Ponzi. But the convexity if it works is 5-10x.

The entire premise of this fund is that markets have changed: 24/7 access, near-zero friction, retail liquidity at sovereign scale. HOOD is the cleanest public expression of that secular shift. Every quarter that passes, more financial life moves onto phones, and HOOD takes a vig.

The PFOF model gets criticized but it's also how this whole apparatus stays free at point-of-use. Add prediction markets, retirement accounts, crypto, options expansion, and 24-hour trading — HOOD is no longer a meme broker, it's the operating system for the new retail.

Risk is regulatory. But every administration that has tried to kill PFOF has discovered that the constituency is now 25 million voters.

The S&P is now seven stocks in a trenchcoat, and those seven stocks are the QQQ. TQQQ gives 3x daily exposure to that concentration with no margin call risk and no overnight wipeout below zero. Volatility decay is real but overstated when the underlying is in a sustained uptrend.

This is a momentum vehicle, not a buy-and-hold. The discipline is exit-on-trend-break, not riding through 2022-style drawdowns. Sized to be cut hard if QQQ breaks the 200dma.

Why not options? Options bleed theta on flat tape. TQQQ holds you in the trade and lets the AI capex cycle compound.

The AI capex super-cycle is real and underestimated. Hyperscaler 2025 capex is $300B+, almost all of it routes through TSMC's leading-edge nodes, which routes through ASML's EUV machines, which routes through SOXL's basket. The picks-and-shovels case has rarely been this clean.

3x leverage compresses time. If the cycle plays out over 18 months, SOXL captures the slope; if it stalls, decay punishes you. Same discipline as TQQQ — momentum vehicle, not core.

The hedge is implicit: when AI capex disappoints, MSTR and crypto names usually rip on the easing that follows.

Two simultaneous tailwinds rarely line up this cleanly. (1) Western governments are rearming after a 30-year holiday from history, and PLTR is the software vendor of choice for the Five Eyes complex. (2) AIP is the only enterprise GenAI product showing actual ROI in production, with land-and-expand metrics that look like Snowflake circa 2020.

Valuation is religious — the multiple only makes sense if you believe Karp's framing that this is a category-defining company. The market increasingly does. The margin of safety isn't price; it's the structural moat of being already-deployed in 50+ classified environments.

Sized to ride a multi-year compound, with the recognition that a 50% drawdown is the tax for owning it.

ETH has been the disappointing cousin to BTC and SOL through 2024-25, and that's exactly when the asymmetry sets up. The ETF is approved. The fee burn keeps net issuance negative under any meaningful activity. L2 throughput is finally credible.

The trade is mean reversion against BTC dominance plus secular adoption of programmable money. Catalysts include a credible institutional staking product, restaking primitives going mainstream, and the eventual realization that "ultra-sound money" wasn't just a meme.

Lower conviction than BTC but higher beta if the regime trade plays out.