I can drop the text and all 8 pictures into the Chrome X draft. First: Chrome menu View → Developer → Allow JavaScript from Apple Events, and have the article draft open. Cover still goes in the preview slot.
Open the X draft in Chrome, allow JavaScript from Apple Events, then push.
1
Tuesday session
2
Semis got the rates punch. SpaceX got it too, then found a bid.
3
Micron lost the $1,000 reclaim. That bounce was the trap.
4
Unlock calendar
5
Falcon 9 is on launch 100 of the year. That will not save Thursday's unlock.
6
Who owns it
7
The map for Thursday
8
The SMH put print sat under the memory unwind.
X Article · Tuesday 18 Aug 2026
SpaceX bought the dip. Thursday is the test.
$SPCX closed $143.34, down 2%. Nasdaq was worse. Memory was much worse.
SpaceX (SPCX) closed Tuesday at $143.34, down 1.98%. That is the story only if you ignore everything around it.
It opened $140.74, tagged $139.41, ran to $146.46, and finished still above the $135 IPO on 83.6 million shares.
Nasdaq was about -1.3%. The semiconductor index was about -5%. Memory was -7% to -9%.
SpaceX ate a macro punch, bought most of it back, and still looks like the bid in a tape that had no bid for the crowded AI hardware names.
Think of it like this. Two kinds of expensive growth stocks got hit by the same 5.33% 30-year yield.
One group (Micron, SanDisk) had never been washed out. The other (SpaceX) already went IPO $135 to $226 to $105, survived a 912 million share unlock, and squeezed shorts from 34% of the float toward 11%.
Tuesday was the market treating those two groups as different animals.
Not investment advice.
The last 24 hours on the tape
Monday close was $146.23, up 4.5%. That session pushed the stock further above the $135 IPO after a two-week rebound from about $105.
The driver was not a new launch. It was the first real look at who owns the company after 13F-type disclosures, plus the afterglow of the 6 August unlock squeeze.
Overnight into Tuesday, futures and long yields did the damage. The 30-year Treasury tagged about 5.33%, a 19-year high. Brent sat near $91.
That is a rates-and-oil tape. SpaceX is a long-duration, high-capex, still-unprofitable mega-cap. On paper it should have been toast.
Print
Previous close
$146.23
Open
$140.74 (-3.7%)
Low
$139.41 (-4.7%)
High
$146.46
Close
$143.34 (-1.98%)
Volume
83.6 million
1Tuesday session2Semis got the rates punch. SpaceX got it too, then found a bid.3Micron lost the $1,000 reclaim. That bounce was the trap.
The shape matters more than the close. This was not a grind lower. It was a gap-and-go dump that found buyers under $141, a push back toward $146 that failed at the $150 wall again, then a settle in the mid-$140s.
Relative strength is the actual 24-hour fact. Memory never recovered. Micron -7.3%. SanDisk -9.3%. SK Hynix -8.4%. Nvidia about -2.2%. AMD about -5%. Intel about -7%.
SpaceX -2% after a 4.7% peak-to-trough dip. If you only read the red close, you missed it.
Why it dipped. Why it did not stay down.
Three things hit at once. Only one of them is SpaceX-specific.
1. Macro, not SpaceX
This was a rates day. Anything whose cash flows are years away got a higher discount rate.
Q2 already showed capex of about $18.4 billion, of which about $15.8 billion was AI, running more than 2x quarterly revenue of $7.81 billion. The open to $140 was index selling, not a new filing.
2. Thursday unlock, two days away
About 319 million shares, roughly $45 billion at Tuesday's price, become eligible Thursday. That is about 3.3 days of average volume.
It is smaller than the 6 August dump of 912 million, which the market absorbed and then squeezed. The difference: 6 August hit a stock near $105. This one hits a stock that just rallied about 40% off the low.
Early holders who refused to sell at $108 are much more willing at $143. That is the real supply risk.
Date
Shares
What it means
6 Aug
~912 million
Done. Stock went up.
20 Aug (Thu)
~319 million / ~$45B
This week. 40% off the low.
September
>700 million
Next wave.
October
>650 million
Float still expanding.
By year-end
~4.9 billion tradable
~70% of non-Musk stock.
Elon ~6.4B / ~48%
Locked to June 2027
He cannot be the seller.
4Unlock calendar
3. Galloway's $10 to $30 call
On a podcast released Monday, Scott Galloway said SpaceX is still crazy overvalued. At $146, his $30 cap is about 79% below the market.
His argument: the price was a market-structure story (tiny float at IPO plus Nasdaq-100 demand), not a cash-flow model. Street average target is still about $226.
He is not a forensic hole in Starlink. He is a valuation polemic. It still matters because it gives underweight managers language on a day they already wanted to sell long-dated growth.
What did not happen: no earnings, no guidance cut, no failed launch, no Musk lockup. Launches are the business. Unlocks are the stock.
What the company actually did
While the stock argued about float, the factory kept running. Starlink Group 17-50 from Vandenberg. 24 satellites. Falcon 9 booster B1097 on its 12th flight. SpaceX's 100th launch of 2026.
5Falcon 9 is on launch 100 of the year. That will not save Thursday's unlock.
Starlink is already about 11,000 satellites on orbit and still about 70% of revenue (about $4.29 billion of the $7.81 billion quarter). This kind of launch used to be a stock event. It is now background noise.
Q2 is still the overlay every seller cites. Revenue $7.81 billion, +92% year on year. Earnings -$0.09 versus about -$0.16 to -$0.26 expected. Capex about $18.4 billion.
Musk: memory is the limiting factor, demand +200% versus supply +20%. Street 2026 revenue about $44 billion. 2027 about $99 billion. That doubling is what a $1.8 to $2 trillion market cap requires.
The stock sold off on the beat two weeks ago because capex was more than 2x sales. The last 24 hours are the market still metabolizing that, not a new fundamental break.
Who owns it
The last 24 hours are also the hangover from the first clean public cap table.
Holder
Stake
Why it matters
Elon Musk
~6.4B / ~48%
Locked to June 2027.
Alphabet
~551M / ~$77B
Largest non-Musk holder.
Fidelity
~303M / ~$42B
Long-only bid.
Gigafund
~172M
Unlock-sensitive.
Saudi PIF
~154M
Sovereign, long horizon.
Nvidia
~123M / ~$17-21B
Came via xAI.
Harvard
~13M / ~$2.2B
Endowment stake.
Norway
~$1.22B
Largest SWF.
6Who owns it
Retail flipped. First net selling since the 12 June IPO. Average retail cost basis about $147. At $143 they are roughly flat to underwater, which is exactly when retail stops being a dip buyer and starts being supply.
The handoff is retail selling the round-trip to IPO, institutions absorbing it. Tuesday's close above $143 on 84 million shares is consistent with that, not with a vacuum.
Shorts had a bad two weeks. Short interest peaked near 34% of the tradable float, then collapsed toward about 11%. The old squeeze is mostly done. A new one is harder with a bigger float.
The map for Thursday
Level
Role
$150
The wall. Close through it on volume = trend.
$146
Monday close / Tuesday high.
$139
Tuesday demand. Lose it, dip-buyers done.
$135
IPO line. Lose it on volume = squeeze.
$105
Washout low.
7The map for Thursday
The acute unlock window is this week through early September. If $135 to $140 holds on 80 to 100 million shares, the 6 August template repeats at a higher price. If Thursday loses $135 on exploding volume, the $90 to $120 bear case is in play.
What social actually argued
The argument is not whether SpaceX is a real company. It is whether $143 is the right number two days before 319 million shares unlock, on a 5.3% 30-year.
Camp A. It held.
They are reading relative strength, not calling $226 tomorrow.
scary to be short… too many institutions behind the name that will not sell at these prices if ever
Stocktwits
such a shit day it did very well
Stocktwits close
Camp B. $150 is a double top.
TraderTV Neal, Tuesday morning: double top at 150, look short unless 150 breaks. Tactically right on the high ($146.46 failed). Wrong on a collapse. Street $226 versus Galloway $30 is the most violent disagreement in the name.
Camp C. Own Micron, not SpaceX.
That relative-value argument had a historically bad 24 hours. Micron -7%. SpaceX -2%. The market paid "already washed out platform with a bid," not "cheap bottleneck."
8The SMH put print sat under the memory unwind.
Do not mix up the stocks with the chips. If memory stays scarce, SpaceX AI capex stays expensive. If memory stocks are sniffing a price peak, SpaceX capex gets easier later. Tuesday did not resolve that.
My view
I would not treat -2% as a failed rebound. I would treat it as digestion of Monday's +4.5% into a macro dump, with the IPO line still intact.
Lose $135 on Thursday's unlock, and the last two weeks were a squeeze. Hold $135 to $140, and the 6 August template gets a second data point. The more important one, because it happens 40% higher.
Galloway is directionally right about structure and wrong about the business at $10. The honest bear case is $90 to $120 on unlock plus rates, not $10. The honest bull case near-term is $150 then $175, not $226, until the float stops expanding every other Thursday.
At about $1.9 trillion you are paying for Musk's $100 billion run-rate / $1 trillion 2030 story. Q2 capex says the company is spending like that story is real. The P&L says it is not earning it yet.
Tonight's 100th launch of 2026 will not save Thursday. If you are going to be wrong, be wrong on the float event, not on whether Falcon 9 can put 24 Starlinks in orbit. It can.
What to watch
$139. Tagged and held. Lose it on unlock, dip-buyers are done.
$135. The IPO. That line decides if the rebound was real.
$150. A close through it on rising volume is the first trend signal.
Thursday volume versus the 319 million. Hold on 80 to 100 million shares = absorbed. Walk through $135 on exploding volume = Galloway gets a vote.
Oil and the 30-year. Brent through $100 does more damage than any lockup table.
Full PDF: digest.panke.app/day/2026-08-18 · Not financial advice.
SpaceX bought the dip. Thursday is the test.
$SPCX closed $143.34, down 2%. Nasdaq was worse. Memory was much worse.
SpaceX (SPCX) closed Tuesday at $143.34, down 1.98%. That is the story only if you ignore everything around it.
It opened $140.74, tagged $139.41, ran to $146.46, and finished still above the $135 IPO on 83.6 million shares.
Nasdaq was about -1.3%. The semiconductor index was about -5%. Memory was -7% to -9%.
SpaceX ate a macro punch, bought most of it back, and still looks like the bid in a tape that had no bid for the crowded AI hardware names.
Think of it like this. Two kinds of expensive growth stocks got hit by the same 5.33% 30-year yield.
One group (Micron, SanDisk) had never been washed out. The other (SpaceX) already went IPO $135 to $226 to $105, survived a 912 million share unlock, and squeezed shorts from 34% of the float toward 11%.
Tuesday was the market treating those two groups as different animals.
The 24-hour tape was constructive, not bullish. Constructive means a bid at $139 on an ugly day. Bullish would have been a close through $150. It did not. Thursday's 319 million share unlock is the test.
Not investment advice.
The last 24 hours on the tape
Monday close was $146.23, up 4.5%. That session pushed the stock further above the $135 IPO after a two-week rebound from about $105.
The driver was not a new launch. It was the first real look at who owns the company after 13F-type disclosures, plus the afterglow of the 6 August unlock squeeze.
Overnight into Tuesday, futures and long yields did the damage. The 30-year Treasury tagged about 5.33%, a 19-year high. Brent sat near $91.
That is a rates-and-oil tape. SpaceX is a long-duration, high-capex, still-unprofitable mega-cap. On paper it should have been toast.
Print
Previous close
$146.23
Open
$140.74 (-3.7%)
Low
$139.41 (-4.7%)
High
$146.46
Close
$143.34 (-1.98%)
Volume
83.6 million
Tuesday session
Semis got the rates punch. SpaceX got it too, then found a bid.
Micron lost the $1,000 reclaim. That bounce was the trap.
The shape matters more than the close. This was not a grind lower. It was a gap-and-go dump that found buyers under $141, a push back toward $146 that failed at the $150 wall again, then a settle in the mid-$140s.
Relative strength is the actual 24-hour fact. Memory never recovered. Micron -7.3%. SanDisk -9.3%. SK Hynix -8.4%. Nvidia about -2.2%. AMD about -5%. Intel about -7%.
SpaceX -2% after a 4.7% peak-to-trough dip. If you only read the red close, you missed it.
Why it dipped. Why it did not stay down.
Three things hit at once. Only one of them is SpaceX-specific.
1. Macro, not SpaceX
This was a rates day. Anything whose cash flows are years away got a higher discount rate.
Q2 already showed capex of about $18.4 billion, of which about $15.8 billion was AI, running more than 2x quarterly revenue of $7.81 billion. The open to $140 was index selling, not a new filing.
2. Thursday unlock, two days away
About 319 million shares, roughly $45 billion at Tuesday's price, become eligible Thursday. That is about 3.3 days of average volume.
It is smaller than the 6 August dump of 912 million, which the market absorbed and then squeezed. The difference: 6 August hit a stock near $105. This one hits a stock that just rallied about 40% off the low.
Early holders who refused to sell at $108 are much more willing at $143. That is the real supply risk.
Date
Shares
What it means
6 Aug
~912 million
Done. Stock went up.
20 Aug (Thu)
~319 million / ~$45B
This week. 40% off the low.
September
>700 million
Next wave.
October
>650 million
Float still expanding.
By year-end
~4.9 billion tradable
~70% of non-Musk stock.
Elon ~6.4B / ~48%
Locked to June 2027
He cannot be the seller.
Unlock calendar
3. Galloway's $10 to $30 call
On a podcast released Monday, Scott Galloway said SpaceX is still crazy overvalued. At $146, his $30 cap is about 79% below the market.
His argument: the price was a market-structure story (tiny float at IPO plus Nasdaq-100 demand), not a cash-flow model. Street average target is still about $226.
He is not a forensic hole in Starlink. He is a valuation polemic. It still matters because it gives underweight managers language on a day they already wanted to sell long-dated growth.
What did not happen: no earnings, no guidance cut, no failed launch, no Musk lockup. Launches are the business. Unlocks are the stock.
What the company actually did
While the stock argued about float, the factory kept running. Starlink Group 17-50 from Vandenberg. 24 satellites. Falcon 9 booster B1097 on its 12th flight. SpaceX's 100th launch of 2026.
Falcon 9 is on launch 100 of the year. That will not save Thursday's unlock.
Starlink is already about 11,000 satellites on orbit and still about 70% of revenue (about $4.29 billion of the $7.81 billion quarter). This kind of launch used to be a stock event. It is now background noise.
Q2 is still the overlay every seller cites. Revenue $7.81 billion, +92% year on year. Earnings -$0.09 versus about -$0.16 to -$0.26 expected. Capex about $18.4 billion.
Musk: memory is the limiting factor, demand +200% versus supply +20%. Street 2026 revenue about $44 billion. 2027 about $99 billion. That doubling is what a $1.8 to $2 trillion market cap requires.
The stock sold off on the beat two weeks ago because capex was more than 2x sales. The last 24 hours are the market still metabolizing that, not a new fundamental break.
Who owns it
The last 24 hours are also the hangover from the first clean public cap table.
Holder
Stake
Why it matters
Elon Musk
~6.4B / ~48%
Locked to June 2027.
Alphabet
~551M / ~$77B
Largest non-Musk holder.
Fidelity
~303M / ~$42B
Long-only bid.
Gigafund
~172M
Unlock-sensitive.
Saudi PIF
~154M
Sovereign, long horizon.
Nvidia
~123M / ~$17-21B
Came via xAI.
Harvard
~13M / ~$2.2B
Endowment stake.
Norway
~$1.22B
Largest SWF.
Who owns it
Retail flipped. First net selling since the 12 June IPO. Average retail cost basis about $147. At $143 they are roughly flat to underwater, which is exactly when retail stops being a dip buyer and starts being supply.
The handoff is retail selling the round-trip to IPO, institutions absorbing it. Tuesday's close above $143 on 84 million shares is consistent with that, not with a vacuum.
Shorts had a bad two weeks. Short interest peaked near 34% of the tradable float, then collapsed toward about 11%. The old squeeze is mostly done. A new one is harder with a bigger float.
The map for Thursday
Level
Role
$150
The wall. Close through it on volume = trend.
$146
Monday close / Tuesday high.
$139
Tuesday demand. Lose it, dip-buyers done.
$135
IPO line. Lose it on volume = squeeze.
$105
Washout low.
The map for Thursday
The acute unlock window is this week through early September. If $135 to $140 holds on 80 to 100 million shares, the 6 August template repeats at a higher price. If Thursday loses $135 on exploding volume, the $90 to $120 bear case is in play.
What social actually argued
The argument is not whether SpaceX is a real company. It is whether $143 is the right number two days before 319 million shares unlock, on a 5.3% 30-year.
Camp A. It held.
They are reading relative strength, not calling $226 tomorrow.
scary to be short… too many institutions behind the name that will not sell at these prices if ever
Stocktwits
such a shit day it did very well
Stocktwits close
Camp B. $150 is a double top.
TraderTV Neal, Tuesday morning: double top at 150, look short unless 150 breaks. Tactically right on the high ($146.46 failed). Wrong on a collapse. Street $226 versus Galloway $30 is the most violent disagreement in the name.
Camp C. Own Micron, not SpaceX.
That relative-value argument had a historically bad 24 hours. Micron -7%. SpaceX -2%. The market paid "already washed out platform with a bid," not "cheap bottleneck."
The SMH put print sat under the memory unwind.
Do not mix up the stocks with the chips. If memory stays scarce, SpaceX AI capex stays expensive. If memory stocks are sniffing a price peak, SpaceX capex gets easier later. Tuesday did not resolve that.
My view
I would not treat -2% as a failed rebound. I would treat it as digestion of Monday's +4.5% into a macro dump, with the IPO line still intact.
Lose $135 on Thursday's unlock, and the last two weeks were a squeeze. Hold $135 to $140, and the 6 August template gets a second data point. The more important one, because it happens 40% higher.
Galloway is directionally right about structure and wrong about the business at $10. The honest bear case is $90 to $120 on unlock plus rates, not $10. The honest bull case near-term is $150 then $175, not $226, until the float stops expanding every other Thursday.
At about $1.9 trillion you are paying for Musk's $100 billion run-rate / $1 trillion 2030 story. Q2 capex says the company is spending like that story is real. The P&L says it is not earning it yet.
Tonight's 100th launch of 2026 will not save Thursday. If you are going to be wrong, be wrong on the float event, not on whether Falcon 9 can put 24 Starlinks in orbit. It can.
What to watch
$139. Tagged and held. Lose it on unlock, dip-buyers are done.
$135. The IPO. That line decides if the rebound was real.
$150. A close through it on rising volume is the first trend signal.
Thursday volume versus the 319 million. Hold on 80 to 100 million shares = absorbed. Walk through $135 on exploding volume = Galloway gets a vote.
Oil and the 30-year. Brent through $100 does more damage than any lockup table.
SpaceX did not almost recover the dip and then fail. It did recover the dip on an ugly tape, then gave back a normal amount of Monday's melt-up into a known supply event. The next 48 hours decide which camp had a process and which camp had a two-week squeeze.
Full PDF: digest.panke.app/day/2026-08-18 · Not financial advice.
Cover goes in the preview slot, not the body. If a paste still drops pictures, use Copy under each one.
