To The Moon

I buy the supply-gap thesis, but not blindly at any price. $MU’s alpha window still looks open because AI servers need more HBM, DRAM, and high-performance storage, while 2026 HBM supply is reportedly sold out and pricing remains tight. That supports the “locked-in” thesis. But after the huge rally and trillion-dollar narrative, the easy re-rating may already be partly priced. The risk is not demand collapse, but cycle ceiling + expectation risk: if Samsung/SK Hynix add supply faster, or hyperscalers slow capex, MU can derate sharply. Optical likely gets the next rotation, especially CPO/800G/1.6T networking names, because compute clusters need faster, lower-power interconnects. Power is the deeper bottleneck: if electricity becomes the constraint, investors may rotate into power, cooling

The civilisation story is powerful, but it is not a valuation anchor. With Tesla, investors are paying for optionality, not just EV earnings. A three-digit P/E only works if autonomy or robotics unlock step-change profits. Without that, the multiple is stretched. The “Elon premium” tied to Elon Musk reflects execution history, but also amplifies downside when timelines slip. The Mars-linked compensation signals commitment, especially alongside SpaceX, but it is symbolic, not a near-term revenue driver. So it comes down to belief vs timing. If autonomy lands, today’s price may still be early. If it delays, valuation compresses fast. Sensible stance: respect the vision, but wait for proof before chasing.

The trillion-dollar narrative around Micron Technology is no longer pure fantasy. The market is finally treating memory not as a cyclical commodity business, but as critical AI infrastructure. That is the key shift. In the old cycle, DRAM and NAND were brutally commoditised. In the AI cycle, HBM has become a strategic bottleneck. Nvidia GPUs are useless without ultra-fast memory feeding them. That changes pricing power, margins, and investor perception. Micron’s own guidance and analyst commentary point to sustained tight HBM supply into 2027.  But a trillion-dollar valuation still requires another leg higher from already aggressive levels. At that size, institutions will demand evidence that: HBM margins remain durable, supply discipline holds, Samsung and SK Hynix do not flood the m

The rally is real in the sense that government money materially reduces survival risk for a sector that has been burning cash for years. But the market is probably pricing in “quantum dominance” far faster than the technology itself can mature. The gap between capital inflow and commercial utility is still enormous.  The next milestone is not “a quantum computer exists”. That already happened in limited forms. The market now needs proof of economic usefulness. Specifically: Demonstrable quantum advantage on commercially relevant workloads Error correction that scales reliably Higher qubit counts with lower error rates Real enterprise revenue beyond pilot projects Integration into AI, pharma, defence, optimisation, or cryptography stacks Right now, most listed names are still effective

The print from Nvidia is not weak. It is too strong for expectations already priced in. 85% growth with ~75% margins tells you demand is still supply-constrained, not fading. The muted reaction signals positioning fatigue, not a broken thesis. What matters is the second-order move: Advanced Micro Devices +8% Arm Holdings +15% Micron Technology +5% This is classic cycle broadening. When the leader stops accelerating, capital rotates into laggards and suppliers. So which is it? Not a top yet, but no longer early. Nvidia: transitioning from hyper-growth leader → “infrastructure anchor” Market: shifting from single-stock trade → ecosystem trade (compute, memory, networking, power) On $220: Bull case: still early in inference + sovereign AI + enterprise adoption → higher base Bear case: expecta

The “Elon premium” tied to Elon Musk is not about current earnings, but future dominance. Markets are pricing outcomes where Tesla cracks autonomy or robotics, or SpaceX extends its economic moat. At a three-digit P/E, this is not a valuation call. It is a probability bet. If even one moonshot scales, it works. If not, downside is severe. The Mars-linked compensation is strong signalling, not practical alignment. It reinforces mission, filters believers, and anchors Musk’s long-term narrative, but it is too distant to anchor financial value. Bottom line: you are buying execution at extreme scale, not cash flow.

The market is no longer debating whether AI demand is real. It is now debating who captures the next dollar of that demand. Let’s separate signal from noise. 1. Nvidia itself NVIDIA is no longer a “growth discovery” story. It is a scale + expectations story. An 85% YoY growth on that base, with 75% margins, is exceptional. But the tepid guidance reaction tells you something important: The market has already priced continued perfection Incremental upside now depends on beating extremely stretched expectations So at ~$220, Nvidia is not “cheap early-cycle” anymore. It is closer to a high-quality compounder with limited room for narrative expansion unless: Blackwell ramps faster than expected, or Hyperscaler capex surprises meaningfully again Otherwise, you get more “good results, muted price

I would not follow this blindly. At +535% YTD, you are no longer early. You are deciding whether to pay for peak narrative plus tightening supply. Let’s separate signal from noise. 1) Tepper buying: meaningful, but not a green light David Tepper tends to lean into macro dislocations, not chase retail momentum. His entry tells you one thing: he believes the cycle still has legs. It does not tell you the entry price is attractive. He can absorb volatility. Most cannot. 2) The real driver: memory cycle turning + AI demand The move in SanDisk is tied to: AI infrastructure pulling forward NAND demand Supply discipline after years of underinvestment Spillover from HBM strength (even though NAND is a different segment) Add Seagate Technology supply warnings, and you get a classic scarcity premium

Short answer: it is far more likely an opening chapter than an ending. But it changes how you should think about Rocket Lab. 1) SpaceX S-1 is not bearish for the sector If SpaceX is genuinely moving toward public markets, it does two things immediately: Forces institutional capital to price the entire space economy properly Validates that launch, satellites, and data infrastructure are no longer speculative niches That is typically bullish for listed peers, not destructive. 2) But it is bearish for lazy RKLB theses Let’s be direct. Many RKLB bulls relied on a “next SpaceX proxy” narrative. That breaks the moment SpaceX becomes investable. Capital that chased RKLB for scarcity may rotate. So RKLB must now stand on fundamentals, not comparison. 3) Where RKLB still has a real edge RKLB is not

This is less about Nvidia alone and more about where we are in the AI cycle. Start with the uncomfortable truth: This did not look like a blow-off top. If anything, it looked like maturing leadership. 1) Nvidia itself: not cheap, but not exhausted An 85% YoY growth rate at this scale, with ~75% gross margin, is not normal late-cycle behaviour. The muted reaction despite strong numbers suggests positioning was crowded, not that the story is broken. The real signal is this: buybacks + dividend + “$200B TAM expansion”. That is a company preparing for durability, not just peak hype. $220 is not a “starting point” in the traditional sense. It is more like a transition zone where expectations are already high, so upside depends on execution staying near-perfect. 2) The more important signal: bre