ARK Invest Recap: SpaceX's $1 Trillion Number, Tesla Autonomy and a Possible Merger

The newest full video on ARK Invest's official YouTube channel, "We Believe The Market Has Not Digested The SpaceX Quarter", was published on November 5, 2026 (about 8:30 AM ET) and runs about 1 hour and 35 minutes. It is a live "call after the calls" panel on SpaceX and Tesla earnings. From ARK: Cathie Wood, Brett Winton, Tasha Keeney, Daniel Maguire and Sam Korus. Guests: Alex Potter (Piper Sandler), Pierre Ferragu (New Street Research) and Aaron Burnett (Mach33). The group walked through the SpaceX revenue and compute numbers that stunned the sell side, Starship and Starlink capacity, how the AI build might be funded, Tesla's full self-driving and driverless progress, ARK's Tesla valuation model, Optimus, and whether SpaceX and Tesla merge.

This is an independent summary written by GNOUP. It is not affiliated with, sponsored by or endorsed by ARK Invest, Piper Sandler, New Street Research, Mach33, WOLF, Public or anyone who appears in the video. The views, forecasts and claims described below belong to the speakers, not to GNOUP. Nothing here is investment, financial or tax advice, and nothing here is a recommendation to buy or sell anything. Forecasts can turn out to be wrong, and past performance does not guarantee future results. Please consult a licensed financial professional before making financial decisions.

Key takeaways

• Sell-side SpaceX models looked far off the mark. Alex Potter of Piper Sandler said Elon's roughly $1 trillion SpaceX revenue figure left consensus for 2028 (he cited about $200 to $220 billion) "roughly five times wrong," and called that "absurd."
• Twenty gigawatts of terrestrial power and cooling was Brett's big number. Brett Winton said SpaceX had lined up about 20 gigawatts of power and cooling for terrestrial compute, versus earlier mental models of maybe 5 to 10 gigawatts installed by the end of 2027, and that this trajectory helps support very large revenue run-rate figures the company discussed.
• Starship cadence of one to two launches a week. Aaron Burnett of Mach33 focused on that weekly cadence as a driver of Starlink bandwidth and as a foundation for Starmind / orbital compute, and said roughly 40 to 60 launches next year would not be crazy relative to industry gut checks that treat those numbers as too high.
• Tesla autonomy numbers impressed the panel. Speakers highlighted about 2 million driverless miles in eight months (Pierre Ferragu argued that pace is about six times faster than Waymo and that vision-only autonomy has cleared proof of concept) and a roughly 75% full self-driving attach rate on new Tesla deliveries in North America before delivery.
• ARK's Tesla base case is $4,400 a share. Tasha Keeney and Daniel Maguire walked through ARK's open-source Tesla model: roughly 80% of enterprise value from robotaxi (about 8 million cars in the fleet by 2031 in their framing), about 14% from Optimus (around 1 million robots cumulatively by 2031 in their forecast), plus energy and AI-compute spend.
• Merger odds were put very high, with caveats. Brett put a SpaceX–Tesla merger at about 90%+ odds within 18 months, with the remaining risk mainly Tesla shareholder pricing and regulators; Alex said better than 50/50 but stressed China and valuation hurdles.
• The bear case was real. Cathie Wood flagged debt-market skepticism (late-1990s scar tissue). Others cited execution over decades, regulation, China risk around a merger, slower robotaxi scale, and the chance the merger never happens.
• These are opinions. Much of the show was sell-side and buy-side reading of earnings calls, models and public statements. Several points were framed as forecasts or hopes, not settled facts.

Newsworthy investment insights

These are opinions, forecasts and claims voiced in the video. GNOUP has not independently checked them. Neither ARK's statements nor ours are a recommendation about any security.

SpaceX's $1 trillion number and the rewrite of sell-side models (opinions)

• Alex Potter said he had never covered a liquid, high-profile company where consensus was so far off: with dozens of firms covering SpaceX, he put 2028 consensus around $200 to $220 billion, versus Elon's much larger 2028 revenue figure, and said analysts were staring at models that needed to be taken "down to the studs."
• Brett tied that upside narrative to terrestrial compute: 20 gigawatts of power and cooling lined up, with talk of coming in around 13 gigawatts and a chance of 20, versus prior guidance-style thinking of 5 to 10 by the end of 2027. He said large run-rate revenue figures (including a $500 billion end-of-2027 style number leading toward a trillion) fall out of driving the terrestrial business that hard.
• Panelists also discussed SpaceX's willingness to test how much compute the market can absorb, with internal uses (Grok / related AI products) as a backstop if external demand softens, and contrasted high marginal rental rates for scarce compute with longer-term neocloud contracts.

Funding the compute build, Starlink and Starmind (opinions and rough estimates)

• Pierre Ferragu and Brett argued that when SpaceX can sell roughly $50 billion a year of data-center capacity that costs on the order of $50 billion to deploy, financing can largely follow the returns; Brett added that SpaceX's shorter deployment cycle and reasonable debt against data centers may mean it does not clearly need equity raises for that strategic compute build.
• Aaron stressed one to two Starship launches per week as the number that unlocks much more Starlink bandwidth than many people expect, and as the foundation for orbital/"Starmind" compute beyond the terrestrial gigawatts. He said his own modeling had been closer to the 40 to 45 launch range, while the company's cadence language points toward 50-plus.
• Daniel Maguire said the next Starship Flight Test 15 aims to catch the upper stage with "chopsticks," which he called the most notable milestone of the program for full, rapid reusability, and bet they would do it. Brett discussed launch costs falling toward the low hundreds of dollars per kilogram with reusability, and potentially toward about $10 per kilogram at true scale (framed as a long-horizon view).
• Tasha said Starlink's opportunity includes enterprise and government, not only consumers, and that the connectivity market SpaceX is talking about is on the order of a trillion dollars across broadband, mobile and related services. Aaron discussed direct-to-cell and how satellite coverage can address large parts of low-density geography without matching every traditional carrier spectrum model.

SpaceX–Tesla merger odds (forecasts, not facts)

• Brett said from a forecast perspective there is a 90%+ chance they merge within the next 18 months, with the remaining roughly 10% hanging on whether Tesla shareholders accept a fair price and whether regulators approve. He argued the companies are already tightly linked (Grok in cars, Starlink on cybercabs, shared fab work, Optimus needing advanced AI).
• Pierre also put eventual merger probability around 90%+, while stressing how painful getting a wide enough Tesla shareholder yes could be. Alex said he would not necessarily say 90%, but better than 50/50, with China regulation and the premium Tesla holders would demand as hard-to-handicap hurdles. He noted Elon's compensation framing implies a very large Tesla valuation in isolation, which can make a modest premium feel insufficient to long-time holders.
• Cathie Wood said she is looking forward to a merger and described SpaceX and Tesla as leading a shift from a long stretch of about 3% real GDP growth toward what she framed as a 7% to 9% real GDP growth world, calling the discontinuity hard for markets to digest. Brett disclosed ARK as longtime Tesla supporters while arguing a combined entity is in both shareholder bases' interest.

Tesla: FSD, driverless miles, robotaxi and ARK's $4,400 model (opinions)

• Tasha said ARK's latest open-source Tesla base case is about $4,400 a share, with robotaxi roughly 80% of enterprise value (about 8 million cars in the fleet by 2031 in their model) and Optimus about 14% (around 1 million robots cumulatively by 2031), with energy making up much of the rest near term.
• Alex highlighted FSD subscription revenue becoming real P&L contribution (he referenced about 1.74 million as a figure that mattered to him in that context) and said it has more than replaced older high-margin credit revenue that skeptics used to dismiss.
• Pierre focused on about 2 million driverless miles in eight months, arguing Tesla is deploying driverless experience about six times faster than Waymo, that the vision-only approach has cleared proof of concept, and that transfer of learning from supervised FSD into unsupervised driving is a major advantage for robotaxi scale.
• Daniel said the call cited roughly a 75% FSD attach rate among North American consumers paying before they get the car (despite a free trial month), which the panel treated as strong product demand. ARK does not put long-term FSD subscriptions at the center of its model because it sees robotaxi as the much larger opportunity. Alex separately models possible future FSD licensing to other automakers as a meaningful line, while putting $0 for Optimus in his formal price target (as a modeling choice, not a claim that Optimus is worthless).

Optimus: overhyped or underhyped? (opinions)

• Several guests said Optimus is not really priced into Street models. Tasha said headlines can look overhyped even while Elon himself warns how hard manufacturing will be; Daniel called the opportunity appropriately hyped but the near-term timeline overhyped; Brett stressed mechanical and software constraints and compared today's humanoids to robotaxis a decade ago—promising demos, hard corner cases everywhere.

Bear case and why the market may not have "digested" the quarter (opinions)

• Cathie Wood said debt markets and analysts scarred by the late 1990s fiber boom may "go on strike" when they see this scale of growth, even though she argues today's shortages of compute and power are a different setup than dark fiber.
• Alex agreed seized-up financing would be bad systemwide, though he argued Elon's companies have historically found capital when others could not. Brett said a lot of misexecution is already priced into SpaceX's current valuation relative to the trillion-run-rate talk, and that markets have not fully internalized Anthropic-like AI cash-flow evidence either.
• On Tesla downside: Alex flagged China / geopolitical risk to a merger; Tasha cited slower robotaxi scale and regulation; Daniel added high-profile accidents and Starship reusability risk for the compute needed for Optimus; Brett said the biggest Tesla risk may be that the merger never happens and Elon's attention stays on SpaceX velocity.

Lighter moments

• Alex joked that covering SpaceX means getting used to "looking dumb" when Elon moves the goalposts.
• Sam quipped that the SpaceX–Tesla merger talk is the "Ross and Rachel" of the day—will they or won't they—and that Brett should record a clip so clients stop asking the same question.
• Cathie and Brett traded the phrase "super duper exponential" from a Big Ideas planning session.

Watch the video

We Believe The Market Has Not Digested The SpaceX Quarter on YouTube (ARK Invest's official channel, published November 5, 2026)

Episode guide

YouTube listed no official chapters for this video, so the topics below are listed in the order they came up on the panel, without timestamps.

• Intro, guests and disclaimer
• SpaceX "number that mattered": $1 trillion revenue and sell-side model rewrite (Alex, Brett, Aaron)
• Super intelligence, Grok / Cursor and compute pricing
• Orbital compute / Starmind and funding the 20-gigawatt build
• Starlink, enterprise connectivity and direct-to-cell
• Capital allocation across terrestrial data centers, Starlink and orbital compute
• Starship Flight Test 15 and launch cadence / launch costs
• SpaceX–Tesla merger odds (Brett, Pierre, Alex, Cathie)
• Tesla "numbers that mattered": ARK's $4,400 model, FSD attach, driverless miles
• How panelists value Tesla (robotaxi, Optimus, FSD, compute)
• Robotaxi scale vs Waymo and the Europe autonomy battleground
• Optimus rapid-fire: overhyped or underhyped?
• Bear case / biggest downside risks for SpaceX and Tesla
• Close

About this summary

GNOUP wrote this summary in its own words from the video's automatic YouTube transcript (via the watch page) and its official description. Short quotations are taken from that transcript or the description. The video had no official YouTube chapters. The transcript had no time markers or reliable speaker labels, so we name speakers (Cathie Wood, Brett Winton, Tasha Keeney, Daniel Maguire, Sam Korus, Alex Potter, Pierre Ferragu, Aaron Burnett) where the official description or clear context makes the speaker obvious, and use "a panelist" or "an ARK speaker" otherwise. Automatic transcripts can contain errors, so some names, numbers and phrases may differ from what was said; we preferred the official description's spellings and the headline figures it highlights, and we left out passing claims we could not confirm (including any guest briefly named in the live intro who does not appear in the official description). We did not check any figures, forecasts or claims made in the video. The video's own disclosure says the information is for informational purposes only and does not constitute an offer, solicitation or recommendation regarding any investment; references to particular companies or securities are not endorsements or recommendations by ARK to buy, sell or hold any security; ARK and its clients or related persons may but do not necessarily have financial interests in securities or issuers discussed; historical results are not indications of future results; and forward-looking statements involve risks and uncertainties. GNOUP has no connection to ARK, Piper Sandler, New Street Research, Mach33, WOLF, Public or their products or offers and does not endorse them. This article is general information only. It is not investment, financial or tax advice, it does not recommend buying or selling any security or asset, and it is not a substitute for advice from a licensed financial professional.