By Dr. Chris Kacher
MU earnings
A key AI company, Micron, reported blowout fiscal Q2 2026 earnings:
Key Results
- Revenue: $23.86 billion (beat estimates of ~$19.7–20.1B; +196% YoY, +75% QoQ)
- Non-GAAP EPS: $12.20 (beat estimates of ~$8.8–9.00; massive beat)
- GAAP Gross Margin: 74.4% (huge expansion)
- Adjusted Free Cash Flow: $6.9 billion
- Q3 Guidance: ~$33.5B revenue (well above consensus ~$24.3B) and strong EPS outlook
- Other moves: 30% dividend increase, raised full-year capex by $5B to chase demand
Steelman Pros (Bull Case)
- Explosive AI-driven growth — Revenue nearly tripled year-over-year. Tight industry supply + massive demand for HBM and AI server DRAM created strong pricing power and record profitability. Micron is one of the few companies directly benefiting from the AI infrastructure tsunami.
- Outstanding execution and margins — Gross margins at 74.4% and record EPS show excellent cost control and mix shift toward high-value AI memory. They are "more than sold out" on certain HBM products.
- Strong forward momentum — Q3 guidance of ~$33.5B is a monster beat. Management remains very bullish on continued AI demand through 2026 and beyond.
- Shareholder returns improving — Raised dividend 30%, strong free cash flow ($6.9B), and ability to fund heavy capex internally.
- Strategic positioning — As a leading US memory maker, Micron is gaining share in HBM (critical for Nvidia/AMD GPUs) while competitors face constraints.
Steelman Cons (Bear Case / Risks)
- Extremely high expectations baked in — Stock is up ~357% in the past 12 months and ~63% YTD in 2026. Even a huge beat led to an initial sell-off because the bar was set so high — classic "buy the rumor, sell the news."
- Heavy capex commitment — Raising spending by another $5B signals confidence but increases risk if AI demand slows or a memory downturn returns. Memory is a notoriously cyclical industry.
- Valuation and pullback risk — After such a massive run, the stock trades at premium multiples. Any hint of softening AI orders, delayed HBM ramps, or broader macro weakness (e.g., higher interest rates, delayed hyperscaler spending) could trigger a sharp correction.
- Cyclical nature remains — While AI currently masks it, traditional DRAM/NAND markets can still swing violently. If AI hype moderates or supply catches up faster than expected, margins could compress quickly.
- Geopolitical and supply-chain risks — Heavy reliance on Asia for manufacturing and exposure to US-China tensions (especially around advanced memory) remain ongoing concerns.
Bottom Line
This was one of the strongest earnings reports of the AI cycle so far; a massive beat-and-raise driven by real, structural demand for AI memory. The fundamentals look excellent in the near term. However, the stock's enormous run means investors are paying a very high price for continued perfection. The reaction (initial dip despite blowout numbers) highlights how frothy sentiment has become.
In the days and weeks ahead, if we see enough positive price/volume entry signals in leading names, this could suggest a major bottom. But until then and as always, let the market tell you how to position your portfolio, and at present, we are in a choppy downtrend.
AI tsunami
The AI tsunami isn't coming—it's supersonic and already here. Revenues are exploding: Anthropic just hit $14B+ run-rate (from $1B not long ago), OpenAI at $25B annualized, no historical parallel in software.
Jensen's math holds: $50B to build a gigawatt-scale AI factory, pulling ~$10B/year in revenue. That's not hype; it's the new economic foundation. Hyperscalers are deploying trillions by 2030 to turn intelligence into a utility like electricity: cheap, abundant, everywhere.
We are seeing dematerialization, demonetization, and democratization followed by disruption in action.
Neuromorphic chips just solved complex physics simulations at 1,000x better energy efficiency than supercomputers. When compute gets that cheap, you don’t just do the same things faster. You do entirely new things that were economically impossible before. Drug discovery moves from weeks on supercomputer clusters to hours on desktop chips. Climate modeling that required national labs runs on university hardware.
The next 6 months? Capability jumps, labor gets displaced at scale, compute clusters multiply, energy constraints start cracking. Nuclear power, solar power, and fusion demos advancing fast (private money pouring in, but still unproven at scale). This isn't incremental tech; it's interlocking exponentials rewriting reality: free intelligence + unlimited energy + robotic labor = abundance economy. Ride the wave or get drowned.That said, bottlenecks could delay the wave:
- Power grids — Gigawatt campuses strain utilities; many projects face delays/cancellations due to grid connection queues, transformer shortages, or regulatory hurdles. Ratepayers could subsidize upgrades, sparking backlash.
- Supply chain & talent — Building at this scale requires unprecedented coordination (chips, cooling, construction); historical mega-projects show overruns.
- Paradigm shifts — Emerging architectures (neuromorphic, sparse computing) could disrupt GPU dominance, stranding investments optimized for today's dense matrix multiplies.
Ray Dalio has written on how the last several hundred years of long-term debt cycles have resulted in collapse and major regime change. He did acknowledge, however, that productivity from AI is a wildcard that could delay or even eliminate such a cataclysmic event. Still, history shows that tech bubbles tend to overspend then burst, so the question is whether this time is different with AI. The numbers so far show AI is setting records but will it be enough to overcome the bottlenecks. Meanwhile, more capital is being put towards solving these bottlenecks than ever, thus the concern of capex overspend.
In terms of addressing the power bottleneck, Elon who has been wrong before (right on direction (reusable rockets, EVs), wrong on timelines (always "next year") where a technology often takes longer to materialize than his estimates) has doubled down hard against Earth-based fusion, calling it "super dumb" because the Sun is already a massive, free fusion reactor overhead; harnessing it beats building tiny ones here. (He even quipped that even torching four Jupiters wouldn't dent the Sun's dominance in solar-system energy.) On the project side, he's targeting insane scale: Tesla (and separately SpaceX) aiming for 100 GW/year of U.S. solar cell production this decade to directly fuel AI data centers (and space-based ones). Recent moves include scouting big new solar factories, expanding Buffalo, and tying it to the AI power crunch; Musk sees solar + storage (Megapacks surging in 2026 deployments) as the practical path to unlimited energy, outpacing fusion timelines for the massive grids AI demands. Fusion's cool for experiments, but solar's the real winner in his view. Aligns perfectly with the abundance thesis: bet on what's already working at planetary scale.
Elon is also creating a chip fabrication facility comparable in scale to TSMC’s largest plants. During Tesla’s January 2026 earnings call, he confirmed the company would “have to build a Tesla TeraFab: a very big fab that includes logic, memory and packaging, domestically” to avoid hitting a hard ceiling on chip supply in three to four years.
The energy bottleneck that threatened to constrain AI is being attacked from every direction simultaneously from physics breakthroughs, nuclear power plant revivals, and vertical integration of the chip supply chain. When challenges get big enough, as Ayn Rand always said, the solutions scale to match. Ultimately, this bodes well for stocks but nothing goes up in a straight line. Let price/volume action guide the way.
How many rate cuts?
As for rate cuts, should the next Fed Chair become Kevin Warsh, markets are betting that he would cut rates. Odds right now suggest no cuts are in the offing for the remainder of 2026. Warsh believes AI is structurally disinflationary which suggests that the Powell dot plot will be a less reliable roadmap going forward. If supply side deflation takes hold, that would give the Fed room to cut. But the question is if and when. That said, the Iran-fueled oil shock gives the Fed more reasons to hold on account of inflationary pressures, but the labor market continues to deteriorate giving the Fed reason to cut. Various cross currents cloud the picture:
### The Current Tug-of-War (Why Cuts Are on Hold)
Two big forces are pulling in opposite directions:
1. **Iran-fueled oil shock (inflationary pressure)**
Oil above $90–100 and potential Strait of Hormuz disruptions are pushing headline inflation higher. This is classic stagflation risk. The Fed’s reaction function is simple here: **hold rates** (or even consider a pause longer than planned) to avoid letting energy costs feed into broader prices.
2. **Deteriorating labor market (dovish pressure)**
Cooling jobs data, rising unemployment claims, and softening wage growth give the Fed reason to ease. A weaker labor market is the classic trigger for cuts.
Add in the longer-term **AI disinflation tailwind** that Warsh highlights, and you get messy crosscurrents. The net result right now: the Fed is on hold until the picture clarifies.
### Realistic Scenarios for 2026
- **Base case (most likely right now)**: **0–1 cut total**, starting no earlier than September/October. Oil shock dominates in the first half; any cut comes only if labor data weakens significantly.
- **Warsh scenario (if he takes over mid-year)**: Markets reprice for **1–2 cuts** (possibly 50bp total) by year-end 2026 as the Fed starts giving more weight to AI productivity.
- **Upside dovish surprise**: If oil prices stabilize quickly and labor softens more, we could see **2 cuts**.
- **Hawkish surprise**: Prolonged energy shock or sticky inflation → **zero cuts** all year (or even a hike discussion such as being discussed by Bank of England and European Central Bank).
Bottom line: The path is clouded. The oil shock has killed near-term cut odds, but a weakening economy and the AI productivity story (especially under a Warsh Fed) keep the door open for easing later in 2026. Right now the market is betting “none until fall at the earliest.” Watch oil prices, unemployment claims, and the next couple of CPI prints. Those will decide whether we get any cuts at all this year.