G.L.A INVE$T RE$EARCH - THE INFINITE INNOVATION FAMILY OF NEWSLETTERS- "MUSKONOMY MONTHLY - SPACEX TO $1000? (VIDEO)

G.L.A INVE$T RE$EARCH — THE INFINITE INNOVATION FAMILY OF NEWSLETTERS MUSKONOMY MONTHLY

Supercomputers, Wall Street waking up, and why we think Tesla + SpaceX is a matter of when, not if — Muskonomy Monthly, October 2026

October 8, 2026

Happy Friday G.L.A Family!

Iv’e been on Wall Street for over 25 years, and every so often a company shows up that makes even the experts admit they had it wrong.

SpaceX is that company right now!

This edition we break down a must-watch interview, separate the facts from the hype, and tell you exactly where we stand. Spoiler: we’re long-term bulls, with our eyes open.

Where the stock stands: SPCX closed at $160.57 on Thursday, Oct 8, down 4.2% on the day. That’s about $2 trillion in market cap. The IPO priced at $135 in June. $1,000 a share would be a roughly 6x move from here.

PART 1: WHAT THE VIDEO SAYS

Source: “Every SpaceX Investor Needs To Watch This” on Brighter with Herbert (YouTube, uploaded Oct 8, 2026; watch here). The host is Herbert. His guest is Brian Wang, a futurist and the founder of the NextBigFuture blog. They recorded it on a Wednesday (Oct 7). Everything in this section is their claim, not ours.

The core idea: SpaceX mass-produces AI data centers

Wang’s main point is speed. He says SpaceX can now stand up a roughly 220,000-chip GB300 AI data center (the captions call these units “MiniHard”) in about 5 months. He says everyone else needs 2 to 4 years.

His logic: chips are about 80% of the cost and go in last. That means “the whole capex spend is happening the month before you start making $2 billion a month from it.”

His timeline: the “MacroHarder” cluster was finished in late September, one “MiniHard” lands in November, and another site follows in December. Wang says that by then SpaceX will hold more AI compute than Google, Meta, or Microsoft. He calls the MiniHard “the most valuable product in the world” until Starship and AI in space get going.

Revenue and earnings, per the video

• Anthropic pays $50 to $60 billion per gigawatt per year. Wang cites Deutsche Bank saying the two most recent leases ran $60 to $61 billion per gigawatt per year. In his words, “not only is it not going down, it’s going up.”

• Wang says each GB300 MiniHard adds about $26 billion of annual revenue. A Rubin-generation MiniHard could add about $75 billion.

• Herbert brings up the SpaceX CFO’s statement that the company is on track for a $100 billion run rate by the end of this year. Wang thinks capacity could reach 2.6 GW by year-end, versus the 2 GW guidance.

• Wang’s 2027 bull case: a run rate of $500 billion or more by the end of 2027. He says that would put SpaceX at “Nvidia level.” (If they even come close to those numbers the stock will have a historic run!)

The members renting the supercomputers

These are the customers and users the video names for 2026–2027 capacity:

• Anthropic, the anchor tenant. Wang cites leaked IPO figures of about $84 billion in deals with SpaceX, or roughly $25 billion a year (about $2B a month). He argues Anthropic won’t cancel because “over half” of its ARR depends on SpaceX compute.

• Two other customers, each taking 110,000 GPUs, at $60–61B per GW per year, per Deutsche Bank. The video doesn’t name them.

• SpaceX, Tesla, and xAI themselves, plus Grok Bot. Herbert cites a leaked Q&A, source unclear, saying 60% of the 10 GW goes in-house and 40% gets leased.

• Microsoft and Meta. Wang speculates that by the end of next year they may “bend the knee” and pay SpaceX to build for them.

• Nvidia as the partner. The video says SpaceX gets Rubin chips first.

Elon’s new direction

Herbert reads Elon’s Oct 7 announcement: Grok Bot will now use “the best backend model for any given task, including Claude Opus 5.5, Midjourney, Suno.” In plain English, SpaceX is building the full stack and plugging in whichever model wins, even a rival’s.

Herbert ties the build-out to Elon’s line that “the factory is the product.” This time the factory is the data center.

Analysts waking up

• Herbert cites Barron’s Al Root: “Boy, we were wrong about SpaceX.”

• He says Morgan Stanley’s Adam Jonas asked a room of investors who owned SpaceX, nobody raised a hand, and Jonas concluded the stock is cheap because people “don’t understand it.”

• The $1,000 math: Wang applies Jonas’s 30x P/E on 2028 to his own EPS estimate of $37.29, which assumes a 50% margin. That works out to $1,119 a share. His stated range is “$600 to $1,000.” Herbert’s math, 5x revenue on a stock at ~$170, gives about $850.

• Wang expects a re-rating once SpaceX proves monthly builds, likely around Q1–Q2 2027.

Even Herbert admits: “The numbers are too big for me to agree to right away… we’ll find out in the next six months.”

Fact-check box: what we verified

• Anthropic deal: confirmed. The S-1 shows $1.25B a month through May 2029, about $45B. Reuters later reported up to $84.5B, mostly cancelable on 90 days’ notice (Barchart, MarketWatch).

• One of the 110,000-GPU customers is Google, per public filings: $920M a month from Oct 2026 to June 2029 (CNBC). The video itself didn’t name Google.

• CFO $100B run rate by December 2026: confirmed (Yahoo Finance).

• Grok Bot multi-model routing, Oct 7: confirmed (The Next Web).

• Barron’s “Boy, We Were Wrong About SpaceX” ran Sept 13, 2026 (Barron’s).

• Actual Street targets. No analyst has published $1,000. Morgan Stanley’s Jonas reiterated Overweight, $300 on Oct 5 (Finviz). He also laid out a bear path toward $100 (TheStreet). TD Cowen initiated at Buy, $200 (CNBC). Raymond James holds the Street high at $800 (Motley Fool).

• The “$50B borrowing” mentioned in the video: current reports say about $40B of Apollo-led debt is in early talks (CNBC).

• The $500B run rate and $1,000+ targets are Wang’s own model. They are not company guidance.

PART 2: OUR VIEW (G.L.A)

1. Early innings, not yet profitable, and that’s the point (Long term: 5–10+ years)

Let’s be straight with you. SpaceX is not profitable yet. In Q2 2026 it posted $7.8B of revenue (+92%) and a $541M net loss. The AI segment ran a $1.26B operating loss on $15.8B of AI capex (SpaceX Q2 release).

That’s what a company looks like at the beginning of its mission.

On a long-term risk-versus-reward basis, we believe SpaceX has a high probability of changing the world for the better and making a lot of money for long-term investors. We’re not saying it’s guaranteed. We’re saying the odds favor the patient.

2. Elon totally rebuilt the machine (Now)

We believe Elon has totally revamped SpaceX, xAI, X.com, and Macrohard. He folded AI into SpaceX, turned a Grok training cluster into a rental business that Google and Anthropic pay for, and made Grok Bot model-agnostic. That’s a founder chasing outcomes, not ego.

3. Tesla + SpaceX: when, not if (Our forecast)

Our opinion and forecast, not fact: we believe a SpaceX–Tesla merger is a foregone conclusion. It’s a matter of when, not if.

For the record, no deal has been announced or filed. On Tesla’s July call Elon said any combination must go through “the appropriate process.” On All-In he said, “who can imagine what action one might take” (24/7 Wall St.).

Our forecast: shortly after a merger, the combined company becomes the largest company on Earth, passing Nvidia (about $5.7T today). We believe it remains #1 for the next 50 years. That’s a bold call. It’s our call, and it may not happen.

4. The two companies, side by side

Tesla brings:

• FSD (Full Self-Driving)

• Cybercab

• Optimus

• Battery storage

• The Supercharger network

• Terafab

• And more

SpaceX brings:

• Launch (Falcon and Starship)

• Starlink

• Colossus

• xAI / Grok

• Terafab

• Space data centers

We believe that between the two, they hold the largest and most important projects in the world and the best engineers. They share employees and recruit the top of the class every year.

ARK says (Big Ideas 2026):

• Reusable rockets cut SpaceX launch costs about 95%, from ~$15,600/kg to under ~$1,000/kg. ARK sees a path to $100/kg with Starship (p. 80).

• SpaceX accounts for ~66% of active satellites (p. 79).

• Space-based AI compute could “come to the rescue” of earthly scaling limits (p. 7).

• AI infrastructure investment could exceed $1.4 trillion by 2030 (pp. 21, 24).

5. Late, but he delivers (Long term)

Elon is sometimes late on his predictions. But he usually comes through in the end. He keeps innovating, and for a leader that’s vital.

Apple was the greatest company ever until Steve Jobs died. Then it stopped innovating and started paying dividends. That’s fine. But we want our companies to keep growing, not cater to short-term dividend investors. Elon is the definition of what we want in a CEO and founder.

Warren Buffett said it best:

Elon reinvests every dollar into the next moonshot. That’s the test he’s betting on passing.

6. How we play it: The G.L.A Constitution

• Weighing machine vs. voting machine. In the short run, SPCX gets voted on every day. Look at the 4% drop today. In the long run, it gets weighed on earnings.

• Volatility is a feature. SPCX has traded from about $105 to $225 in its short public life. That’s the price of admission.

• DCA. We add on weakness over time. We don’t chase green candles.

• Never margin. Not on a stock that can swing 10% in a day. Ever.

• Know what you own. That’s why we fact-checked the video for you.

• Mine the risk. See below.

Time horizons:

• Next 6 months: MiniHard delivery cadence, Q3 earnings (late October), Starship Flight 15.

• 2027–2028: the 5–10 GW build-out and earnings that start to show up.

• 2030s and beyond: Starship, orbital compute, Tesla–SpaceX.

RISKS WE’RE WATCHING

• Execution. Five-month builds are a claim until we see it done month after month. Chips and memory are the bottlenecks, even by Wang’s account.

• Valuation. At about $2T and still unprofitable, the stock prices in a lot of perfection. Jonas’s own bear case is ~$100.

• Customer concentration. Anthropic’s contracts are largely cancelable on 90 days’ notice.

• Key-person risk. Everything here runs through Elon.

• Merger, regulatory, and governance risk. Exchange ratios, conflicts of interest, minority shareholders, and antitrust all come into play. A deal may never happen.

• Capex and debt. Tens of billions a quarter, plus a reported ~$40B debt raise.

• Competition. Hyperscalers, custom chips (ARK flags ASICs gaining share, p. 24), and neoclouds.

Thanks for reading, family. Stay patient, stay curious, and keep building toward the mission: “Creating sustainable abundance for all.”

— G.L.A INVE$T RE$EARCH