What's happening with SpaceX price action? Here's what AI says...

A few years ago I sat on the other side of the trade everyone is piling into right now.

Not in equities. In crypto.

Some of the most hyped token launches of the last cycle were engineered.

A sliver of the supply hits the market, the fully diluted value looks enormous, and with almost nothing to actually sell, the price has nowhere to go but up. I worked with some of the top market makers in the space to design exactly these setups.

The mechanic is simple.

When only a few percent of the supply can trade, even modest buying overwhelms the order book and walks the price up faster than it should.

Shorts cannot borrow enough to lean against it, so the float does the work.

So when $SPCX IPO'd on June 12, I recognized the machine immediately.

SpaceX floated ~5% of its shares. It opened at 150, ran to 225 intraday by its third session, and now everyone I talk to is asking the same three questions.

• Is it a buy?
• Is it a short?
• What actually happens next?

I don't do price targets. But I have worked with this setup before, and I wanted to leverage AI to build a research report on how this has played out in similar "low float high valuation" setups in the past.

A powerful thing that these AI tools can do is identify historical corollaries .... and we'll be working with this this week.

Here's our plan:

• Find every comparable low-float listing, stocks and tokens
• Pull the exact lockup and vesting dates from the filings
• Measure when they top and what times the drawdown
• Map all of it back onto SpaceX

Download the full report

I'm recapping the top research findings below but if you'd like to read the full report for yourself check it out here: Link​

What I had Claude Code build...

The "low float" cohort came out to 61 listings. Thirty-four fresh IPOs and de-SPACs, and 27 low-float crypto tokens, going back to 2019.

Between them I tracked 129 supply-release events. Every price series, every lockup expiry, every token vesting cliff, pulled and checked against the original filings and tokenomics docs.

Here is every one of them, stacked on top of each other.

Each faint line is one listing. The melt-up is real. So is everything after it.

4 lessons came out of the data, plus one that cuts the other way....

Lesson #1: The top comes fast

The first thing that jumped out was the timing.

Median time to the top was 24 trading days after listing (~5 weeks)

And 71% of these names never traded higher again. That first blow-off was the high-water mark for the entire history I could see.

I watched this happen in real time with the 2021 class.

Rivian topped inside its first week and gave back more than 90%. Lucid and QuantumScape ran the same shape.

The melt-up is violent, and it is brief.

Lesson #2: The drawdown is timed to supply, not valuation

This is the one I actually use.

The big declines are not random, and they are not really about the stock getting expensive.

They are timed to the supply hitting the market.

That chart is the average move around every supply release, lined up so zero is the release date.

Prices are already soft heading in. Across the whole set, the ten trading days into a lockup expiry run about negative 6.1% of market-adjusted return.

Then the two asset classes split, and the split is the whole story.

• Equities mostly stop falling. The weakness front-runs the lockup, the supply gets priced in ahead of the date, and the stock steadies.
• Tokens keep bleeding. The ten days after a token cliff average negative 8.2%, and it keeps drifting from there. That held even when I kept just one cliff per token.

The date is on the calendar months in advance. The reaction is not.

Lesson #3: Who is locked up matters more than how much

Here is the pattern that matters most for SpaceX.

Who holds the locked shares changes the outcome.

When the supply is spread across early investors and employees, they sell, and the stock bleeds after the date.

But when a single controlling founder holds the block and legally cannot sell, the market prices that overhang in ahead of time instead of getting dumped on.

Look at the difference. Diffuse holders are weak going in and keep falling after the date. The controller names take almost all their pain before the date, and then the move after is basically flat.

The overhang still gets priced. It just does not get sold into the tape.

Lesson #4: It is a supply structure, not a chart pattern

I care about the misses as much as the hits, because they tell you what this actually is.

It was not a technical signal.

Overbought RSI did not mark these tops in any reliable way.

And a bigger melt-up did not lead to a deeper crash. The collapse was close to universal no matter how high the thing ran first.

There is no clean line on that chart. The drawdowns cluster near the floor across the whole range of run-ups.

This is not a pattern you trade off a screen. It is a supply schedule you read off a filing.

And if you have ever tried to short one of these early, you know the other half.

Low float cuts both ways.

There are not enough borrowable shares to press, so the squeezes are vicious, and a short who is right on direction still gets carried out first.

So if you're brave enough to short you must watch the calendar like a hawk...

Lesson #5: It does not always round-trip

This is the lesson that complicates the story, and it is the one that matters most for SpaceX.

Not everything cratered.

The shallowest drawdowns in the whole set were all equities, and they shared two traits.

A controlling holder who could not sell, like Arm, Birkenstock, and Saronic.

Or a real, fast-growing business, like Reddit, Astera Labs, and CoreWeave.

Often both at once.

And 17 of the 58 names later made a new high.

Arm ran from its IPO to more than 400 dollars. Reddit, Astera, Rocket Lab, and AST SpaceMobile all re-rated to fresh highs in a later cycle.

But read the fine print on that chart. Even these survivors fell a third to two-thirds first.

“Worked out” meant living through a deep drawdown, not skipping it.

The setup is not destiny!

So where does that leave SpaceX?

I will lay out the setup and let you draw your own line.

Its release schedule, straight from the prospectus, is the densest I found in the entire cohort.

An earnings-linked release of 20% of the locked pool, plus another 10% if the stock holds 175.50.

Five more tranches of 7% each between August and late October.

A 28% release after the third-quarter print. Then the full cliff on December 8.

All of it lands in the back half of this year, sitting under a float of a few percent.

Now weigh the other side.

SpaceX has both traits that defined the survivors in Lesson #5. Musk’s roughly 6.4 billion shares are locked for 366 days with no early-release provision, so the controlling block reads as a known June 2027 event rather than an imminent seller.

And the underlying business is not a meme....it's literally creating an entire new category of frontier science....

So you have the densest supply calendar in the set running into year-end, against the float scarcity and controller lock that have historically softened the fall. That is the tension for you to make a decision on. The schedule is printed in advance, which is more than you can usually say.

The part that matters for you

I am not a market-structure desk. I built this in a day, by myself (using Claude Code).

I have the methodology in the raw download of the full research report above .... but if you're keen on learning the nuances of how I made this in Claude Code feel free to hit the reply button and I'm happy to give more details.

Personal

As mentioned last week, I'm doing a free live online stream with Andrew Walker from the Yet Another Value Podcast and Ben Collins from AlphaSense on how I think about designing the buyside AI stack.

If you prefer to get a recording of the event after, feel free to sign up and you'll automatically get a recording of it.

See you there!

-Dave