The Latticework Monday Morning Briefing is a research-based, data-driven slide presentation sent on a separate mailing list (complimentary to members). If you do not wish to receive it, opt out here.
Over the next couple of weeks the discipline of value investing runs headlong into the loudest event the public markets have seen in a very long time, if ever. SpaceX has filed for the largest IPO in history, with a listing reportedly targeted for June 12. OpenAI is expected to follow in September, and Anthropic is eyeing October. Together the three are seeking something on the order of $3.5+ trillion in combined market value. Three pieces shaped my thinking on what a long-term investor should make of this, and each comes at it from a different altitude.
Trung Phan reframes the SpaceX S-1 as an AI infrastructure story wearing a rocket costume. His word count of the filing tells the tale: “compute,” “xAI,” and “connectivity” now crowd the document alongside “launch.” The pitch is no longer rockets and Starlink alone. It is a $26.5 trillion AI total addressable market sitting beside a $1.6 trillion connectivity number and a $370 billion space number, with SpaceX positioning to own the compute. Phan is candid about the absurdity, invoking Matt Levine’s “Elon Markets Hypothesis,” the notion that some assets are priced not on their cash flows but on their proximity to Elon Musk. At $18.7 billion of 2025 revenue, which he pegs at roughly 93 times sales, that hypothesis is carrying an enormous load. What a value investor should notice is the structure of what is being sold: a narrative priced as though the option has already paid off.
Aurelion Research does the arithmetic the narrative would prefer you skip. They give the reframe its sharpest name, the AI Trojan Horse, a compute business carried to market inside a rocket company. Their sum-of-the-parts is the most useful work I read this week. Strip away the story, and Starlink, the AI segment, and the launch business support a valuation near $700 billion (still generous, in my opinion). To bridge from there to $1.75 or $2 trillion, you have to assign close to a trillion dollars to space data centers that do not yet exist and to a Mars optionality that may never produce revenue. Their cross-IPO comparison is equally clarifying. Of the three, OpenAI carries the worst loss profile and the deepest dependence on someone else’s compute, while Anthropic shows the fastest revenue growth, the earliest path to break-even, and the cheapest forward multiple. Aurelion states plainly that they will own none of the three, and that their own book offers better risk and reward. That is the right instinct. Relative value is real, but the least expensive of three expensive things is a ranking, not a margin of safety.
Ben Thompson lands on the most philosophically honest framing. He says outright that the filing cannot be justified by any financial model, calls the numbers absurd, and notes that growth slowed even as the xAI acquisition tipped the company into a $4.9 billion loss on $5.1 billion of AI research expense, money spent building a model that sits in fifth place. And yet he argues that orbital data centers are plausible, that SpaceX could become the dominant supplier of the world’s marginal compute, and that this alone might be enough. The rumored IPO valuation embeds extraordinary risk.
The three pieces agree on the one fact that matters and disagree only on whether it should trouble you. Phan, Aurelion, and Thompson all conclude that current financials cannot support the proposed valuations, with SpaceX being the most egregious and Anthropic being the most palatable. Phan and Thompson lean on the Musk reality-distortion field, the Elon Markets Hypothesis, and the Tesla precedent, to argue the price may hold regardless. Aurelion does the work to show exactly how much hope is embedded in the number, and then declines to participate.
This is the capital cycle in its most public form, and the frames from last week apply here without modification. The AI buildout is cresting, the spending curve is steep, and the most celebrated names are being floated at the very top of it. Paying up at that moment, on margins and multiples that are stretched at the same time, is the error the historical record warns against most consistently. None of this makes the businesses bad. Starlink is one of the finest assets to reach the public markets in a decade, and SpaceX’s launch economics are a genuine moat. It means the edge for a long-term investor lies in the willingness to decline mispriced hope, to read this IPO wave as a sentiment marker rather than a shopping list, and to hunt instead in the corners where fear, not hope, is the thing being priced in.
A bit of feedback
“Most of what I monitor, all in one place. Great value add.” —Brad Lummis
“Loving these Monday briefings!” —Jon Bartel
“Tightly presented and easy to digest. I just spent 20 minutes going through it, and it’s helped to level set me for the week ahead.” —Michael Loftis
“A great piece and thoughtfully assembled.” —Brian Wolf
“I don’t think I have ever seen more valuable content in one place.” —Bill Coleman
“Worth its weight in gold.” —Shree Viswanathan
A few words on the format
The Briefing is designed to answer a deceptively simple question. If you were sitting down before the weekly market open, as an investor rather than a trader, what would you want in front of you?
Each week, the Briefing walks through four parts.
Weekly Review & Outlook covers equity performance, sector moves, the earnings just reported, and the earnings coming up, alongside curated editorial highlights from our Weekly Inspiration newsletter.
Idea Generation surfaces candidates from screens we run: biggest decliners, names near 52-week lows, low multiples, high FCF yields, spinoffs, activist situations, buybacks, short interest, and more.
Market Valuation & Positioning steps back to the index level: the Buffett Indicator, aggregate multiples versus history, S&P 500 concentration, equal-weight versus cap-weight, and long-run factor returns.
Macro & Fixed Income rounds out the picture with rates, credit spreads, the Fed balance sheet, the dollar, labor, regional PMIs, and housing.
The new issue is attached below. We welcome your suggestions for improvement as we refine the format week to week.
