Welcome back to“MUSKONOMY MONTHLY” where we report on the greatest innovator of our time ELON MUSK.
G.L.A. INVE$T RE$EARCH THE INFINITE INNOVATION FAMILY OF NEWSLETTERS MUSKONOMY MONTHLY Issue: October 2026 | Coach Kyle
"Stop being patient and start asking yourself, how do I accomplish my 10-year plan in 6 months? You will probably fail but you will be a lot further ahead of the person who simply accepted it was going to take 10 years." — Elon Musk
THE STACK IS NO LONGER A COLLECTION OF COMPANIES. IT IS ONE MACHINE.
Subscribers, this is the issue where the Musk stack stops being a set of separate bets and starts reading like one industrial system.SpaceX is public. xAI already sits inside it. Starship has reached orbit. Terafab just got a hard ownership line drawn in public. And the orbital computer — Starmind — is no longer a podcast line. It is an FCC-accepted filing for up to one million satellites.
We are here to keep you in the know with everything Elon and is band of mission driven engineers do!
SCOREBOARD — AS OF OCTOBER 7, 2026
SpaceX (SPCX):IPO Listed June 12, 2026 at $135. Record $75 billion raised. Offer valuation about $1.77 trillion. Shares recently cleared $170 again, with reports of a market value above $2.3 trillion. June high was $225.64. Morgan Stanley target $300. Goldman $230. Morningstar has argued the listing was already rich versus its intrinsic estimate. Lockup is the quiet risk: more supply can hit after third-quarter results, and the full lockup opens December 8, 2026.Revenue estimates have been raised for 2027 and 2028 recently.
Tesla (TSLA): Trading near $375, roughly a $1.5 trillion public market cap, about 16% lower on the year. Third-quarter deliveries beat expectations. Cybercab robotaxi count in Austin has moved from the mid-40s in early September to roughly 150-plus registered units.Earnings October 21.
xAI: Acquired by SpaceX on February 2, 2026, after the March 2025 combination of X and xAI under X.AI Holdings. The AI lab is no longer a cousin. It is a subsidiary.
Starlink: About 11,150 satellites working in orbit. Flight 14 put the first Starlink V3 birds up from Starship.
THE ORBITAL PRINT — SPACEX AND STARLINK
Flight 14, September 28, 2026, is the fact that matters. Starship reached orbit for the first time, deployed 26 Starlink V3 satellites, and splashed down north of Hawaii.
Flight 15 is the next catalyst, penciled for late October or early November. Morgan Stanley has flagged an upper-stage catch attempt —if it happens, and if it works — as the largest upside event for the stock since the IPO.SpaceX has not confirmed the catch.
Under the hood, Q2 revenue was about $7.8 billion, up roughly 92% year over year. Connectivity is the profit engine. Space and AI are still the investment engines.
Starlink at ~11,150 working satellites is a real utility.The filed ambition — up to 100,000 Starlink satellites — is a different company. Direct-to-cell, defense (Starshield), and V3 throughput are the bridges between those two numbers.
TESLA — THE GROUND NODE
Tesla is the customer, the sensor fleet, and the robot factory of the stack. The car business had a hard couple of years but is starting to perk up. The multiple no longer lives there. It lives in robotaxi and Optimus, and neither is at scale……YET!
What is real:Cybercab is in Austin, the registered fleet has more than tripled in about a month, and Q3 deliveries beat. What is not real yet: a robotaxi business that shows up in the income statement, or an Optimus line that changes factory economics. Earnings on October 21 is the next hard print. Listen for fleet size, city count, unsupervised miles, and any Terafab capital language.
TERAFAB — THE WAFER IS THE MOAT
This morning Elon Musk closed the ownership question in public. Responding to speculation that TSMC would own and run the Texas complex, he wrote: "No, we will build and run the fab. Let there be ZERO doubt about that." TSMC, in his framing, can sublease a slice. Nothing more. Hours earlier, Intel CEO Lip-Bu Tan told Bloomberg that Intel stays involved.
Site:Grimes County, Texas, about an hour from Houston. Confirmed August 2026. Phase-one commitment: $16.8 billion.Longer-range talk — 100 million square feet, a campus Musk has called the largest and most valuable building on Earth, investment figures cited as high as $119 billion. Pilot line talk is about 3,000 wafers a month. Silicon production is being discussed for 2029, with some filings pointing at a construction window through the end of 2028. Physical progress is still early.
What the wafers are for:Tesla inference (FSD, Cybercab, Optimus), xAI training and inference, and the space-based computer. That is why Terafab sits at the center of every merger conversation. A foundry that serves three of his companies is not a side project. It is the shared factory.
Coach Kyle’s read:Terrafab will change the game, Elon could be late with this but my money is that he completes it and revolutionizes fab production.
THE DATACENTER IN SPACE — STARMIND
On October 6, SpaceX published a Space Safety page that put numbers on the orbital computer.Plans: up to 100,000 Starlink satellites and up to 1 million Starmind AI-compute satellites. The FCC accepted the million-satellite filing in February. Musk confirmed the Starmind name in June. In August, SpaceX said Nvidia is working on the compute payload.
Architecture, per SpaceX VP Michael Nicolls: clusters of about 10 satellites flying in tight formation, linked by on the order of 10 terabits of bandwidth, each satellite roughly a 250-kilowatt rack of compute, tied back into the wider constellation. Ground compute does not go away. Memphis — Colossus, including a facility described at over 2 gigawatts — is the training floor. Orbit is the inference and scale floor, sold on the idea that power, cooling, and permitting get easier above the atmosphere.
Related ground pieces:a Gigasat-class factory discussed for Bastrop to build the orbital compute satellites, and Starship as the truck. Internal targets floated in the ecosystem — prototypes early 2027, volume late 2027, a gigawatt-class annual add, then a steep ramp — should be treated as targets, not guidance.
XAI — INSIDE THE PUBLIC COMPANY
The February 2, 2026 acquisition folded xAI, and with it X, into SpaceX.That is why SPCX is not a launch stock. Connectivity, launch, and an AI lab with a social distribution pipe are in one ticker. Grok is the model. Colossus is the cluster. Starmind is the proposed off-planet extension. X is the messy, lower-multiple media asset inside the same box.
Valuation argument, stripped of romance: Morgan Stanley is effectively saying a slice of today's price is the AI computer, and that slice could be worth far more if orbital compute works. Skeptics (Morningstar and others) say you are paying for that outcome at a low probability. Both can be right at $171. The late-October quarter is the next time the market has to separate Starlink cash from AI ambition.IMHO anyone who doubts musk is mentally slow.
THE BORING COMPANY — THE QUIET UTILITY
Still private. Still the least priced part of the stack. Current tape: a push toward four tunnel-boring machines operating in Nashville, and a reported ambition to build on the order of 150 kilometers of tunnels in the UAE. The Loop is a niche transit product until a city actually runs serious volume through it. We watch it because tunneling is the unglamorous cousin of the same idea as Starship: move mass where the surface is congested. It is not a merger driver. It is optional infrastructure.
NEURALINK — THE LONG DURATION BET
Still private, still independent, but has massive longterm upside.
Fresh Signal:the company has been training on tens of thousands of hours of brain data and has posted a cursor-control record around 11 bits per second. Blindsight, the vision implant, is aimed at a first patient trial in 2026 and has carried FDA breakthrough-device designation. The commercial path Musk has described — higher-volume implant production, a more automated surgery — is a 2026 intention, not a product launch.
THE MERGER — ALREADY HALF DONE
Stop saying "if he combines the companies" as if nothing has combined.
Done:X and xAI into one holding structure (March 2025). That structure into SpaceX (February 2, 2026). Starlink was never separate. The public vehicle already contains launch, satellite broadband, the social network, and the AI lab.
Not done:Tesla. On the July earnings call and again at the All-In Summit in September, Musk declined to close the door. His line, paraphrased across both: there is more overlap every quarter, and a combination cannot be discussed on an earnings call — it has to go through the appropriate process. Wedbush's Dan Ives has put the odds at 80 to 90 percent and pointed at the first half of 2027. Gene Munster has said 90 percent over the next few years. Jefferies has sketched a nil-premium structure in which Musk could still hold roughly 55 percent of the vote. Gwynne Shotwell has said folding things together might make management simpler.
Why the logic exists: Terafab serves both. Starlink connectivity in Cybercab and Optimus is obvious. Tesla energy storage at Starbase and at ground compute sites is obvious. Optimus labor inside a fab and a launch site is the speculative version. A stock deal is cleaner now that SpaceX has a ticker.
Why it is not free: Tesla shareholders have to approve a combination of this size. Dual-class control at SpaceX (Musk kept overwhelming voting control at the IPO)makes the negotiation asymmetric. A nil-premium deal favors the story over the Tesla multiple. Regulatory review of a combined auto, robotaxi, launch, satellite, and AI company would not be a rubber stamp. And a merger does not fix Raptor reliability, robotaxi utilization, or fab yield.
Our stance:underwrite the overlap, do not underwrite the closing date. The economic merger is already happening through Terafab, energy, connectivity, and compute. The legal merger is a 2027 question with a real process in front of it.
FIVE COMPANIES THAT BENEFIT FROM A MERGER BETWEEN TSLA and SPCX
These are not recommendations. They are transmission belts. If the stack combines — or even if it only keeps collaborating — demand shows up on other people's order books. Sized as suppliers, not as Musk substitutes.
1. Intel (INTC). The named process partner. Lip-Bu Tan confirmed on October 7 that Intel stays in Terafab, on the 14A node. A combined Tesla-SpaceX buyer is a larger, stickier customer than either alone. Risk: 14A has to yield, and Musk has reserved the right to run the building himself. Intel can be essential and still not control the outcome.
2. Nvidia (NVDA). Near-term beneficiary, long-term tension. SpaceX has said Nvidia is on the Starmind compute payload. Colossus is still a GPU cluster. Every year custom Terafab silicon slips, Nvidia ships. Every year custom silicon works, a slice of inference leaves the merchant GPU. The thesis is the ramp, not permanent exclusivity.
3. Applied Materials (AMAT). Wafer-fab equipment. A campus discussed at 100 million square feet does not get built without tool orders years before the first good die. Phase one at $16.8 billion is the number that can turn into purchase orders. The $119 billion figure is not.
4. Lam Research (LRCX). Etch and deposition, the other half of the tool duopoly that gets paid when a new logic-and-memory fab is real. Same timing logic as Applied: orders lead wafers by years. Same kill-switch: if Terafab stays a site plan, the orders never print.
5. Quanta Services (PWR). The unsexy one. Terafab, a multi-gigawatt Memphis cluster, Starbase, and a Bastrop satellite factory are all electrical-infrastructure projects before they are technology projects. Grid interconnect, substations, and high-voltage work do not care whose name is on the rocket. They care that the capex is funded. SpaceX's post-IPO cash pile is what makes this one less theoretical than the tool names.
Honorable mention, not in the five: battery and lithium exposure (Panasonic, Albemarle) if Optimus, Cybercab, and Megapack at launch sites scale together. We left them off because the merger does not change their thesis as cleanly as it changes Intel's, the toolmakers', and the grid contractors'.
WHAT WE WATCH INTO NOVEMBER
- Starship Flight 15: date, catch attempt or not, engine performance versus Flight 14.
- SpaceX late-October quarter: connectivity margin versus AI and launch spend. Then the lockup window.
- Tesla October 21: Cybercab fleet, cities, unsupervised miles, any shared capex language with SpaceX.
- Terafab: a real ground-break, a real tool order, or another post. Only the first two count.
- Starmind: any hardware photo, any spectrum update, any slip past the "scaling next year" line.
- Moon: orbital refueling schedule. Not dome renderings.
Close
The greatest innovator of the era is not a stock. He is a capital-allocation system that now has two public tickers, one private neuro company, one private tunnel company, a foundry that does not yet make wafers, and a filing for a million computers in orbit. Muskonomy is the work of separating the machine that exists from the machine on the slide.
Stay in the know. Compound on purpose.
Coach Kyle G.L.A. INVE$T RE$EARCH The Infinite Innovation Family of Newsletters
Disclaimer: G.L.A. INVE$T Research Consulting is not a licensed advisor or broker. This is independent research and opinion for subscribers. It is not a solicitation to buy or sell any security. Do your own work.