"Permission Slip"
Week of 6/14
Good morning and welcome back, everyone.
First, congratulations on a strong first half, because this one was earned. The story of 2026 so far has been a leadership change, and the people who saw it early got paid. A new class of leaders carried this market, names out of the semiconductor ecosystem that spent years as afterthoughts and came roaring back. INTC tripled off its lows and printed its first all-time high in a generation. DELL more than doubled. MRVL doubled and tagged fresh records. ARM kept compounding. The whole silicon complex re-rated in front of our eyes. And while the new guard did the heavy lifting, the prior leaders did not break, they built. The megacaps that led the last leg held near their highs and consolidated, digesting the run instead of giving it back. That is what a healthy market looks like, old leadership basing while new leadership runs.
That backdrop is the whole reason we keep our screens pointed where we point them. The semi ecosystem is still where the money is moving, and the flow this week says it has not stopped. Now we protect what we built and we press where the edge is.
This is a heavier read than usual because the week earns it. Fed Wednesday, a fresh AI policy curveball over the weekend, and a tape that wants higher but needs permission. Let’s get into it.
Weekend Headlines
• Anthropic pulled offline. The US Commerce Department issued an emergency export-control directive Friday evening barring foreign-national access to Anthropic’s two most advanced models, Fable 5 and Mythos 5. Unable to verify nationality in real time, Anthropic disabled both globally. First time Washington has effectively recalled a commercial frontier model. Full breakdown below, it is the headline that matters most for our book.
• Iran peace deal looks close. Trump confirmed planned strikes were called off and a final document could be signed in Europe as soon as this weekend, reopening the Strait of Hormuz and lifting oil sanctions. Crude fell more than 3 percent Friday to an eight-week low in the mid-80s.
• SpaceX debuts and rips. SPCX priced at 135, opened at 150, closed its first session around 161, up roughly 19 percent. Largest IPO in history at a 1.77 trillion valuation. Options listing is the next catalyst and the flow will be loud when it lands.
• Consumer sentiment ticked up. June Michigan came in at 48.9 from 44.8, with year-ahead inflation expectations falling to 4.6 from 4.8. Still near record lows, but the direction is right into a Fed week.
• Oracle got punished. ORCL fell 8.5 percent Thursday after guiding fiscal 2027 capex above estimates. The market is rewarding AI spend discipline and punishing the names that show it spiraling. Remember that everywhere this week.
The Anthropic Story & What Regulations Means
This is the one everybody will be talking about Monday, so this is where we keep our heads while the timeline loses theirs.
Here is what happened. Friday evening, Commerce invoked national security authorities and ordered Anthropic to cut off all foreign-national access to Fable 5 and Mythos 5, including foreign nationals inside the US and even Anthropic’s own non-citizen employees. The trigger was a report that the model could be used to find software vulnerabilities, reportedly surfaced to the administration by Amazon. Anthropic says the capability is relatively basic, exists in other public models, and that the whole thing is a misunderstanding they expect to resolve. White House AI adviser David Sacks said the restriction was issued reluctantly and that he hopes it gets fixed so the models return to general release. Their other models including Opus 4.8 were untouched.
Now the bear case making the rounds, and credit where it is due, it is a smart argument that deserves a real answer. The thesis: this bull market is built on AI, the trade prices in continued rapid model advancement plus fast commercialization, and export controls on models plus restrictions on foreign nationals working on them introduce friction to exactly the layer the market assumes moves fastest. A large share of frontier researchers are foreign nationals. Gate the talent and you slow iteration, and if AI progress becomes policy-bound instead of engineering-bound, the multiple expansion behind this whole market gets harder to justify.
Here is my answer, and it is the same thing I said on X this weekend. Regulation is a familiarity in nearly every legal business aspect of the world. It is not a foreign object that breaks the machine, it is a constant the machine has always run inside of. What separates winners from losers is how companies adapt to and benefit from the rules, and that adaptation is exactly how you gauge who carries the next leg. There is always a bull market somewhere. But to the bear’s point, plenty of names that fail to adapt are going to get run over, and that part is true.
The reason I am not flipping defensive on this specific headline comes down to one word: scope. This is a one-company, one-model enforcement action, contested by the company, called reluctant by the White House, framed by everyone involved as something to be resolved rather than a permanent regime. It is not a broad export ban on all frontier AI and it is not a blanket restriction on foreign researchers across the industry. The layer we actually trade, compute and infrastructure, is not where this bites. If anything, a world where access to centralized frontier models becomes a policy football strengthens the case for the picks and shovels underneath, the chips, the memory, the power, the grid, because demand for the buildout does not care which lab is cleared to serve which user this week.
So the read is this. Watch for a Monday open that wants to fade the headline. If the AI complex gaps down and the infrastructure names get dragged with it, that is the dislocation we want, not the breakdown we fear. The risk to respect is real: if this expands from one company into a framework, or the resolution drags, the friction argument starts earning its multiple compression. We watch the scope. Scope is everything here.
The Setup
This is a Fed week, and not a routine one. The FOMC decision lands Wednesday at 2pm and the market is pricing essentially no chance of a move, rates hold at 3.50 to 3.75. That part is settled. Everything around the decision is not. This is Kevin Warsh’s first meeting in the chair, and the reporting going in is that he may start dismantling the forward guidance machinery we have all traded around for years, possibly dropping the dot plot’s rate forecast and stripping the bias language out of the statement. That is the real event Wednesday. Not the rate. The framework.
The backdrop sharpens it. CPI is running near 4.2 percent and the repricing this year has gone one direction: fed funds futures now lean toward a hike as the more likely next move, not a cut. We came into 2026 expecting cuts. That conversation has mostly closed. The risk Wednesday is a hawkish surprise dressed up as a process change. The offset is energy. Oil in the mid-80s and falling on the Iran de-escalation takes pressure off the exact inflation prints keeping the Fed boxed in, and Friday’s Michigan data leaned the right way. So we walk in with a real tailwind and a real two-sided risk sitting on top of it.
If you are reading this in the Substack chat, the Discord is where the real-time version of all of this lives. The chat here is fine, but the Discord moves all day, every day, live flow, live reads, levels called as price hits them. That is the room you want to be in. Paid subscribers, come find us.
And Volux is almost here. Options flow, breadth, AI research, screeners, creator tools, all of it in one terminal built the way I have always wanted to trade. The waitlist is open right now and launch is imminent. If you have followed this newsletter for the edge, Volux is that edge built into a platform. Get on the list before the doors open.
The Indices
The whole board tells one story right now and it is a constructive one. We took the hit on June 5, we found the low, and the tape has spent the last week putting in supportive closes right where it needed to. What we want from here is the only thing that matters: hold these closes, carve a higher low, then take out a higher high. That is how a flush becomes a launchpad. One index is already doing it.
