The Frontier File · Issue #4 · June 24 2026

I Almost Sold
The Steak.

One retail account, one frontier-tech thesis, documented in public. This week I nearly trimmed my biggest winner days before it printed the blowout of the year. The trade that didn't happen taught me more than the ones that did.

+58%
On capital · Day 85
▲ new all-time high
$8,360
Book value · from $5,270
20 / 20
Positions green
Cold Open

The trade that didn't happen

Last Friday, Micron popped 5% to a fresh record high and I came within a click of selling my entire position. The logic was real: it was my largest holding, it had run from the low-$300s to over $1,100, and a concentrated position that size makes me twitchy. I'd talked myself into it. Then the order didn't complete, the weekend came, and Monday a global memory-chip rout hammered MU 13% in a single session. For about 48 hours, not selling looked like a mistake.

Then Wednesday night happened. Micron reported revenue of $41.5 billion — up 346% year over year — beat EPS by 24%, posted a record 84.9% gross margin, and guided next quarter to $50 billion against a $43 billion estimate. It signed sixteen multi-year customer agreements and a memory-supply deal with Anthropic, and the CEO said the supply tightness is "locked in to persist beyond 2027." The stock ripped 12–15% after hours toward $1,180 — a new all-time high.

I held every share. My book crossed +58% and a new high tonight. And the lesson isn't "diamond hands always win" — it's the opposite of cocky. It's: I nearly let position-size anxiety override a thesis I had no fundamental reason to abandon. That's the trade I need to learn from, even though it worked out.

The Thesis Board

Where my head is at

The map before the positions. firing   mixed/watching   next-wave seed   headwind.

AI Infrastructure / Memory MU · STX · NVDA · ARM
Micron just answered the "AI spending won't pay off" panic with a $50B quarterly guide, 85% margins, and binding multi-year contracts. The single loudest data point of the whole AI buildout. This is the foundation, confirmed.
Concentration Risk MU now ~22% of book
NEW this issue. After the pop, my biggest winner is a fifth of the whole account. The thesis is intact, but the sizing isn't. This is the thing I'm managing next — carefully, after the euphoria settles, not at the gap-up.
Rates / The Fed still the headwind
A hawkish dot plot last week, BofA flagging a hike, yields elevated. May PCE drops tomorrow — the macro referee. MU's blowout helps sentiment; it doesn't lower the discount rate.
Space Economy SPCX · RKLB · PL · ASTS · MDA
SPCX cratered 26% off its peak after a surprise bond sale — exactly why I cut it to one share near the top. Proxies still strong (RKLB +62%, PL +54%). The IPO froth is deflating on schedule.
Small-Cap / Rate-Sensitive ACHR · JOBY · SOFI
The Russell tagged 3,000 for the first time ever this week before pulling back. My rate-sensitive names held up — they bounce hardest whenever the hike fear eases.
Next-Wave Seeds SERV · OKLO · QBTS
Planted small, behaving. Micron's Anthropic deal is one more proof point that the physical-AI and infrastructure layer keeps getting more real, not less.
The Week in One Tape

Just the numbers

MarkerLevelRead
MU after-hours the blowout
~$1,180
+12–15%
Nasdaq chip rout Tue
25,587
-2.21%
S&P 500 tech-dragged
7,365
-1.44%
Russell 2000 tagged 3,000 first time
~2,960
milestone
SPCX off the peak
~$155
-26%
WTI Crude peace roadmap
~$73
easing

It was a two-act week. Act one: a global memory rout, sparked by SK Hynix slowing advanced-chip output and AI-spending skepticism, dragged the whole semi complex down — Micron fell 13% Tuesday alone. Act two: Micron's print blew the skepticism apart after the bell Wednesday and dragged the sector right back up. The fear and the answer arrived in the same five days.

The Portfolio Report

The whole book, no hiding

MU
+266%
Blowout · ~22% of book
RKLB
+62%
Launch proxy
PL
+54%
Sat imagery
STX
+46%
Memory tailwind
CENX
+45%
Aluminum
AEVA
+39%
LiDAR
BWXT
+38%
Nuclear defense
GOOGL
+29%
Joins Dow Mon
MDA
+29%
Space systems
ASTS
+29%
Sat connect
JOBY
+28%
eVTOL
ACHR
+27%
eVTOL
NVDA
+24%
Bounced on MU
AMZN
+18%
Cloud · robots
SPCX
+15%
Cut to 1 share ✓
SOFI
+10%
Fintech rails
QBTS
+7%
Quantum
OKLO
+6%
Nuclear AI
SERV
+4%
Delivery bots
ARM
+2%
Chip IP

This issue's moves

Held (barely): the full MU position through a near-sale and a 13% rout — into a blowout. Trimmed: SPCX to one share near the highs; it's since fallen 26% off its peak, so the call holds. Next: right-size MU back toward ~15% of the book once the post-earnings euphoria settles — not at the gap-up. Same discipline that trimmed it at +77%, applied at +266%.

+58%
On invested capital
$8,360
Book value
20/20
Positions green
+266%
MU, the steak
~4×
vs S&P 500
The Catalyst Calendar

We position around dates

Thu Jun 25
May PCE inflation + Q1 GDP final — the Fed's referee on the hike question
High
Mon Jun 29
GOOGL joins the Dow Jones Industrial Average — index-inclusion tailwind
Tailwind
Days ahead
Right-size MU after the pop settles — manage the new ~22% concentration
My move
~90–180d
SPCX lock-up expiration — the known risk on the remaining share
Watch
Q3 2026
Honeywell Aerospace (HONA) spin-off — the GE-style unlock
On deck
Late Jul
SOFI Q2 earnings — the coiled spring gets re-tested
High
The Hot Take

"The chip rout said AI spending would never pay off. Two days later Micron guided to fifty billion dollars a quarter. The bears didn't get a thesis — they got a tantrum."

For 48 hours this week the narrative was that the AI trade was over: SK Hynix slowing output, hyperscaler capex with no returns, a global memory selloff. Then the largest pure-play memory company on earth posted 346% revenue growth, record 85% margins, sixteen binding multi-year contracts, and a guide that obliterated estimates by seven billion dollars. The CEO said tightness is locked in "beyond 2027."

That's the difference between a valuation wobble and a demand problem. The rout was the former dressed up as the latter. The earnings are real, the contracts are signed, the margins are at all-time highs — and a hawkish Fed doesn't change any of that, it just changes the multiple you pay for it. I'll keep saying it: I get cautious when the orders get cancelled. This week the orders got locked in for five years. The tantrum is not the thesis.

Buddy, Real Talk

It worked — and that's exactly why I have to be careful

Here's the honest part, and it cuts against the win. I almost sold my whole Micron position out of pure position-size anxiety — not because anything in the thesis had broken. It hadn't completed, the stock dipped, then it blew the doors off and I look like a genius for holding. But I didn't hold on conviction. I held by accident. If that order had gone through Friday, I'd have written a very different issue this week, and I'd have been telling you why selling the top was disciplined.

That's the trap of a good outcome: it can launder a sloppy process into a "great call." So I'm marking it plainly — the result was lucky, the near-decision was emotional, and the actual lesson is that I let a 22%-of-book position get that big without a plan, then nearly panic-trimmed it at the worst possible moment. Now MU is even more oversized after the pop, and the grown-up move is to right-size it deliberately once the euphoria cools — not to high-five myself and let it ride to 30% of the account.

You get the win and the self-critique in the same breath. That's the only way this stays honest.

The Bottom Line

Hold the thesis, manage the size

Eighty-five days. $5,270 deployed, ~$8,360 on the board, a new all-time high at +58%, and a Micron position that just printed the blowout of the year — $41.5B in revenue, an $50B guide, 85% margins, contracts locked beyond 2027.

But the issue isn't the win — it's the wobble behind it. I nearly sold the steak right before the feast, out of anxiety rather than analysis, and now that same position is a fifth of my whole book. The thesis has never been stronger; the sizing has never been more lopsided. So the plan is simple and unglamorous: hold the conviction, trim the concentration deliberately after the pop settles, watch tomorrow's PCE print, and keep the SPCX lesson close — froth deflates, fundamentals don't.

Issue #3 was the dip that got bought. Issue #4 is the trade I almost made and the discipline I almost skipped. The market rewarded me anyway. I'd rather tell you it was luck than pretend it was genius.

See you next Sunday, buddy. — Alex