The Call After The Calls: SpaceX And Tesla

Tesla

Across the shows: 24 speakers in other episodes, 18 bullish and 19 bearish calls, past 30 days

• Brett Winton · ARK InvestBullish

  SpaceX-Tesla merger is 90%+ likely within 18 months and a win-win for both shareholder bases.

  Brett puts the merger at 90%+ odds within 18 months. He sees it as a win-win because the companies are already cross-integrated: Grok is distributed through the cars, Starlink is appearing on Cybercabs, they are building TerraFab together, and Optimus depends on SpaceX's superintelligence. Cybercab cash flow will flow onto Tesla's balance sheet. The remaining 10% depends on whether Tesla shareholders accept the price and regulators approve. He argues a fair premium that deals in both shareholder bases is the market-clearing price, and that by 2031 two-thirds of combined revenue and more than two-thirds of profit will be SpaceX line items. Brett is a long-time Tesla holder.

  “we think there's a 90 plus percent chance that they merge within the next eighteen months”at 39:24

• Pierre Ferragu · New Street ResearchBullish

  SpaceX-Tesla merger is about 90% likely eventually; the open question is timing and shareholder acceptance.

  Pierre sees the merger as roughly 90% or more likely to happen eventually. The real questions are how long it takes and how much Tesla gets. The only thing standing between the two companies is an offer that satisfies a wide enough group of Tesla shareholders, which may be painful. He says he has rarely seen two companies with as many operational reasons to combine.

  “So if the question is will they eventually merge, [I] think that probably the [probability] is like 90% or 90 plus percent.”at 40:54

• Alex Potter · Piper SandlerUncertain

  Merger odds are above 50/50 but a 20-30% premium may not satisfy Tesla holders.

  Alex won't go as high as 90% but thinks the merger is better than 50/50 because Elon clearly wants it and you don't bet against what Elon wants. The China regulatory angle is hard for him to handicap. The main hurdle is price. Elon's compensation package implies he values Tesla on its own at about $8.5 trillion. Tesla holders remember the stock going from a $50 billion market cap to $650 billion, so a 20-30% premium does not sound lucrative enough to them.

  “I don't know if I would say 90%, but I certainly hope that that's true. I think definitely better than fifty fifty.”at 41:31

• Tasha Keeney · ARK InvestBullish

  Robotaxi should be about 80% of Tesla's value, with an 8 million car fleet by 2031.

  ARK's updated Tesla model assumes robotaxi will make up roughly 80% of enterprise value, based on an assumed fleet of about 8 million cars by 2031. The update also has Optimus ramping faster than the prior model, to about 14% of enterprise value and about 1 million robots. The majority of value therefore depends on the robotaxi business scaling.

  “we think that the majority of the enterprise value will be based off of the robotaxi business. That's roughly 80% of the enterprise value. We're assuming roughly 8,000,000 cars in the fleet by 2031.”at 59:28

• Alex Potter · Piper SandlerBullish

  FSD subscription revenue is now real, high-margin P&L income replacing lost regulatory credit revenue.

  Alex says the FSD revenue shows Tesla now has a real proof point of nearly pure-margin revenue. It has more or less fully replaced the 100% margin regulatory credit revenue that critics always discounted, and then some. FSD subscriptions are a lucrative upward trajectory and matter for the P&L today, with implied take rates an exciting indicator. He sees robotaxi and other businesses as the bigger drivers of terminal value, but FSD matters now.

  “but this is a real proof point of more or less pure margin revenue, and at this point, it's more or less fully replaced the 100% margin credit revenue that people always used to complain about”at 1:00:54

• Pierre Ferragu · New Street ResearchBullish

  Tesla's 2M driverless miles in eight months prove vision-only autonomy works and will overtake Waymo.

  Pierre notes Tesla reached 2 million driverless miles in eight months, deploying six times faster than Waymo. A year ago he saw a dark spot on when the FSD-to-unsupervised transition would inflect. Running six times faster means Tesla will take over Waymo quickly, possibly imminent in Austin. Achieving this from a much simpler stack, without a roughly $50,000 sensor suite per car, proves vision-only autonomy has passed proof of concept, a view he had long been skeptical of. It also proves experience transfers from FSD to driverless.

  “The first one is that the vision only approach that I've long... About which I've been, for for many, many years, very very skeptical and and cautious is actually working.”at 1:03:12

• Tasha Keeney · ARK InvestBullish

  Tesla's robotaxi cost per mile is about 40% below Waymo's, so Tesla can scale first.

  Waymo started service first, but Tesla could be the first to scale because it is vertically integrated. ARK estimates Tesla's incremental cost per mile is about 40% cheaper than Waymo's next-gen vehicle, the Cybercab. Tesla also has the Waymo playbook to follow. ARK sees the robotaxi market as early and thinks Waymo's rollout has been more cautious than expected, so the question is who scales, and Tesla could be that company.

  “We estimate on a incremental cost per mile basis, they're about 40% cheaper, than the Waymo next gen vehicle, with the CyberCab.”at 1:04:46

• Daniel Maguire · ARK InvestBullish

  ARK's Optimus forecast of 1 million robots by 2031 is conservative versus Musk's 100 million-plus target.

  Daniel says ARK models Tesla deploying roughly 1 million Optimus robots cumulatively by 2031, which they consider very conservative. Musk has said 100 million, if not 1 billion, could be deployed in the same year. ARK's estimate therefore includes an element of conservatism. Optimus is one of the most notable changes in the updated model, and the company's AI compute spend makes Optimus more capable over time.

  “in the out year 2031, we think Tesla can deploy roughly 1,000,000 [Optimus] cumulatively. Our [Optimus] forecasts are very conservative. So Elon in the same year has said 100,000,000,”at 1:06:13

• Daniel Maguire · ARK InvestBullish

  Robotaxi is a 10+ trillion dollar market at scale, larger than any FSD subscription revenue.

  Daniel says ARK does not model FSD subscriptions in the long run, because robotaxi is an enormous opportunity. He puts its market size at the order of 10 trillion dollars or more at scale. The 75% pre-purchase FSD rate in North America shows how compelling the service is, but the long-term value sits in robotaxi.

  “We really think Robotaxi is such an enormous opportunity on the order of 10 plus trillion at scale of a market size.”at 1:07:36

• Alex Potter · Piper SandlerBullish

  Other automakers will eventually license Tesla's FSD, as they were forced to adopt its charging standard.

  Alex expects FSD licensing to become a major contributor, though he admits there is no evidence yet. He argues other automakers will be forced to capitulate, as with Tesla's Supercharger network that became the North American charging standard. Customers may soon refuse to buy a BMW or similar car without full self-driving, and Tesla's approach of building cars around its electrical architecture and licensing FSD is fundamentally different from rivals. In his model, licensing ends up almost as large a driver as robotaxi.

  “I assume FSD licensing will be... this is not something you have evidence for yet, but I think that eventually other automakers' hands will be forced”at 1:09:50

• Pierre Ferragu · New Street ResearchBullish

  Tesla is the dominant standalone robotaxi player because it can variabilize its fleet with owners.

  Pierre's team built five robotaxi scenarios five years ago, and the most favorable was Tesla developing unsupervised FSD. Today they sit in the second-best scenario, with good conviction that Tesla dominates the market despite some competition. Tesla benefits from being able to turn its fleet into a robotaxi network using owners' cars, which nobody else can replicate regardless of technology or capability.

  “probably dominating the market. We kind of see Tesla as being the dominant standalone robotaxi player because they benefit from this ability of variabilizing their fleets with Tesla owners that nobody else can replicate.”at 1:12:49

• Pierre Ferragu · New Street ResearchBullish

  Tesla's about $60B 2030 EBIT implies it trades at about 50x earnings, with Optimus unmodeled upside.

  Pierre models a reasonably optimistic robotaxi rollout that gives about $60 billion of EBIT in 2030. He uses this to think about where the market will be in 2030, with the robotaxi opportunity still in very high growth. Optimus is not modeled and he treats it as upside, as robotaxi is today. He says Tesla's current earnings multiple, at least 50 times, already reflects some of this optionality, because it is a very high-growth, high-potential stock.

  “The EBIT number I can get to in 2030, it's like $60,000,000,000.”at 1:13:48

• Tasha Keeney · ARK InvestBullish

  Robotaxis will make autonomy cheap and widely available within five years, proliferating ride-hail.

  ARK argues robotaxis will make autonomy cheap and proliferate ride-hail. Tesla has already shown the economics work: it has chosen to price its Cybercab rides below Uber and other services, so cheap autonomy is already possible and the open question is when it becomes widely available. ARK expects that within five years and expects it to be very meaningful in that window. Tasha also flags the main risk: robotaxi scaling more slowly than expected, with regulation and remaining technical issues able to slow the rollout. She still believes Tesla will be first to scale.

  “We think that robotaxis will make autonomy cheap, and that will proliferate ride hail. ... So we think that'll happen within the next five years.”at 1:16:46

• Alex Potter · Piper SandlerBullish

  Tesla will beat Chinese autonomy rivals in Europe because its vertically integrated approach is more scalable and cheaper.

  Alex, back from China, argues Europe is where Tesla meets Chinese autonomy head-on. Chinese players are mostly either consortium-style replicas of Waymo (sensor supplier, automaker, ride-hailer and algorithm developer combined) or Mobileye-style level-2 systems. Tesla's approach is different: it owns the automaker, the purpose-built data-gathering fleet, data centres and possibly its own chips and electrical architecture. Only vertically integrated automakers can walk that path, and it takes many years, so Chinese companies are behind. Chinese companies will not face the same protection at home in Europe, and Alex thinks Tesla wins on scalability and cost. He does not expect the Chinese players or Waymo to be driven out, using the Uber-Lyft analogy: Tesla can over-earn while pricing at parity or slightly below rivals.

  “And I think Tesla will win, because ... it's more scalable, and the cost is going to always be lower.”at 1:20:39

• Pierre Ferragu · New Street ResearchBearish

  Tesla cannot compete at scale in humanoid robots given global competition, despite delivering operationally.

  Pierre says the only tangible thing to judge is the scale of the manufacturing floor. Given the level of competition around the world in humanoid robots, he does not think Tesla can get into that game at total scale. He concludes that what Tesla has delivered operationally is not overhyped, so the criticism is about the competitive scale of Optimus rather than delivery. His scepticism is about Tesla's ability to win a global humanoid market, not about whether the company has shipped something real.

  “given the level of competition around the world on humanoid robots, I don't think Tesla can get into that game as a total of scale.”at 1:24:06

• Alex Potter · Piper SandlerBullish

  Tesla is attractively valued; the stock implies Optimus and autonomy are worth roughly zero.

  Alex says you do not need to believe much about Tesla's robotics or autonomy ambitions to find the stock attractive at its current valuation. He sees the current price as implying the businesses are worth something more than zero, and thinks that is too low given what the company could deliver. He calls this a cop-out because he does not need to put a specific value on the upside. He also says the sell side does not model Optimus numerically and he does not either.

  “...you need to believe very much about it to believe that Tesla's attractively valued at this valuation... It's worth something more than zero, which is what I think the stock currently implies it's worth.”at 1:24:56

• Tasha Keeney · ARK InvestBullish

  Optimus expected to be under 20% of Tesla's value in five years, with much more upside longer term.

  Tasha says ARK is excited about the humanoid robot opportunity but the question is timing. Elon himself has tried to temper expectations on how hard manufacturing will be. ARK expects Optimus to be less than 20% of Tesla's value within the next five years, with the value potentially much larger longer term. The forecast is a modest near-to-medium-term contribution with a larger option value beyond that.

  “In the next five years, you know, as I've said, we expect it to be maybe less than 20% of Tesla's value, but longer term, it could be much more than that.”at 1:25:44

• Daniel Maguire · ARK InvestBearish

  Optimus opportunity is real, but timelines are overhyped because there is no inherent data advantage.

  Daniel says Optimus is appropriately hyped in terms of the opportunity, but people are too aggressive about the timeline, and the industry is in the very early innings. Hardware difficulties are one reason. The bigger one is data: no player has a real inherent data advantage, and a manually controlled humanoid is not very useful, whereas a manually controlled car is useful. So Daniel thinks the scale of the opportunity is fine but the speed at which it arrives is overstated.

  “So I think it's appropriately hyped in terms of the opportunity, but just people are maybe too aggressive in terms of timeline, and we're in the very early innings.”at 1:26:19

• Brett Winton · ARK InvestBearish

  Optimus is a far harder problem than robotaxis, so it may not be the best use of Tesla's capital.

  Brett says the humanoid robot problem is hard. The robots are mechanically and software constrained, and money alone cannot solve that. He compares it to robotaxis a decade ago, but says the corner cases for humanoids are effectively the whole universe of reality, versus a curve or an edge case for cars. He is not sure Optimus is the best use of capital for Tesla specifically, because the product is so hard to get working and to market, and Elon's ambitious targets tend to slip. He does think it is crackable, but only with a lot of superintelligence combined with manufacturing learning.

  “It is just such a huge challenging problem. I think they are mechanically constrained and software constrained and you can't just throw money at it and solve it.”at 1:26:46

• Alex Potter · Piper SandlerUncertain

  A Tesla–SpaceX merger faces geopolitical and regulatory risk Tesla cannot control, possibly China-related.

  Alex says the merger may be weighing on Tesla's valuation because investors are holding out for it. He thinks the biggest downside is outside management's control. Vertical integration lets Tesla pull the company forward itself, but if regulators or Sino-US relations get in the way, Tesla cannot act unilaterally. He cites China risk as the most likely way the merger gets torpedoed. He asked Tesla's CFO about carving Tesla China out to make the merger palatable and got no comment, which he treats as the main issue. He says he finds the risk hard to handicap.

  “if regulators or geopolitical issues, you know, Sino US relations throw a wrench in the gears, then that's something that Tesla can't unilaterally control.”at 1:29:25

• Daniel Maguire · ARK InvestBearish

  A Starship failure would block reusable rockets and the compute needed to train Optimus, a major Tesla risk.

  When asked for the biggest downside risk to Tesla, Daniel names a top-down risk: if Starship blows up and reusable rockets fail, the compute needed to train Optimus and tap into the multi-trillion-dollar robotics opportunity may not be available. He connects SpaceX's launch progress to Tesla's robotics ambitions through shared compute needs. The argument depends on that link, which he states without much detail.

  “if Starship blows up and they can't get reusable rockets because a lot of that compute will be needed to train, Optimus and tap into that multi trillion dollar opportunity.”at 1:31:04

• Brett Winton · ARK InvestBullish

  A Tesla–SpaceX merger is in both shareholder bases' interest; without it Tesla loses its exploration edge.

  Brett argues the biggest risk for Tesla is that the merger does not happen. Elon already spends four days a week in Memphis building out superintelligence for SpaceX, so his attention would increasingly go to SpaceX without a merger. Tesla would then probably execute on CyberCab and robotaxi, moving into exploitation mode but losing its exploration edge, and Optimus would drift on timeline. Tesla shareholders, who have been on a long journey, may feel they should get a high premium, but Brett says it is in the interest of both SpaceX and Tesla shareholders to combine and move forward aggressively. He adds that Tesla's stated ambition now covers all transportation, so a merged entity could make surprising moves.

  “the risk for Tesla is that the merger doesn't go through. Because Elon is already spending four days a week in Memphis to make sure that superintelligence gets built out for SpaceX.”at 1:31:29

Space Exploration Technologies

Across the shows: 42 speakers in other episodes, 52 bullish and 23 bearish calls, past 30 days

• Alex Potter · Piper SandlerBullish

  Sell-side 2028 SpaceX revenue estimates (~$200-220B) are far too low versus Musk's trillion-dollar framing

  About 50-60 sell-side firms cover SpaceX, and consensus for 2028 revenue sits around $200-220B. Potter says that is roughly five times wrong relative to what Musk said 2028 revenue could be. He calls the gap absurd and says the model needs to be rebuilt 'down to the studs' rather than tweaked, because SpaceX is far more volatile to model than typical coverage. Confidence is moderate because he does not commit to his own revenue figure.

  “Roughly five x wrong relative to what Elon said 2028 revenue could be. That's absurd. And like I said, after the call is done as an analyst, you sit there staring at your model”at 3:36

• Brett Winton · ARK InvestBullish

  SpaceX will build terrestrial compute far faster than guided, reaching about 13-20 GW by 2027

  SpaceX previously guided to 5-10 GW installed by end-2027. Musk's new figure is about 13 GW, with a probability of reaching 20 GW, with 20 GW of power and cooling already lined up. Winton says this shows Musk can drive his companies at a maniacal pace, which undercuts the narrative that data centers are hard to build and permit. From this build-out, the model yields about $500B run-rate revenue by end-2027, which leads to the trillion-dollar figure.

  “that was the biggest change for us in terms of how we thought about the trajectory of the company... And that 20 gigawatt number or even the 13 gigawatt number you can use to get to some of the... revenue figures... So, 500,000,000,000, you know, run rate revenue by the end of 'twenty seven, leading to that trillion dollar number”at 5:47

• Aaron Burnett · Mach33Bullish

  Starship could fly 40-60 launches next year, far more than most people expect

  Burnett's models already had about 40-45 launches, and the one-to-two-per-week cadence implies roughly 50-100. He thinks a realistic up-mass of 40-60 launches next year is achievable, because SpaceX is building roughly one rocket per month and the cadence only needs to move from one every six weeks. At about 50 terabits per second per launch, that would double, triple or quadruple bandwidth on orbit, which strains other people's models.

  “I think we're looking at a realistic up mass of somewhere in that 40 to 60 range of number of launches we can see next year, which is very big. It's something like doubling or tripling the total bandwidth on orbit”at 8:45

• Pierre Ferragu · New Street ResearchBullish

  SpaceX's Grok has reached frontier-model relevance, and the Cursor deal makes it a leader

  Over the last six months SpaceX's model moved from a chatbot-level product to one that qualifies for the frontier, though still behind. Ferragu calls the Cursor acquisition a mastermind move. He says Cursor captures about 25% of AI tokens used in software development, which puts SpaceX into a large market. Combined with Grok and its compute, he thinks all the ingredients are there for SpaceX to be a leader, because growing compute faster than rivals wins.

  “over the last six months we went from having a model that could do well as a chat box that could qualify for the frontier... and then the cursor acquisition is like a mastermind move from the perspective of SpaceX”at 10:31

• Brett Winton · ARK InvestBullish

  SpaceX can monetize compute at $30-40B per gigawatt, well above neoclouds' ~$20B per GW

  Anthropic's S1 shows frontier labs will pay over $50B per gigawatt for marginal compute because their cash flow is huge, and the enterprise market is a $40T opportunity at full penetration. SpaceX, with scarce supply, can charge that margin. ARC's model has SpaceX monetizing at $30-40B per GW, while neoclouds signing longer contracts do not exceed about $20B per GW until the early 2030s. If demand softens, SpaceX can absorb the capacity internally to train Grok, so it can keep prices high.

  “30 to 40,000,000,000 per gigawatt that SpaceX will begin to monetize at, whereas Neo Clouds that are signing longer term contracts really don't get above 20,000,000,000 per gigawatt until the early 2030s.”at 14:05

• Alex Potter · Piper SandlerBullish

  SpaceX's vertical integration lets it bring compute online at its own pace, justifying high pricing

  Potter argues that vertical integration is Musk's companies' special sauce. Competitors in a consortium depend on suppliers, so the slowest link limits their deployment. SpaceX controls its own compute rollout, and in a market with insatiable compute demand, that gives it unilateral control and the competitive position behind the high pricing Winton described.

  “To be the one who has unilateral control over bringing compute online at your own pace has got to be an incredibly empowering feeling. And that is the competitive position that they and they alone are in at this point.”at 16:13

• Aaron Burnett · Mach33Bullish

  SpaceX's coherent orbital clusters will be larger than modeled, supporting more inference capacity

  Burnett says SpaceX has discussed coherent clusters flying in formation, a new approach that lets it build much larger clusters than Mach33 originally modeled, in terms of the number of big models served together and token throughput. That means more inference and better models. The key question is whether orbital data centers become a foundation by 2027 with early tests, or come later in 2028-2029, given upmass.

  “they talked about coherent clusters now flying them in formation, which is kind of a new thing that allows them to get much bigger than we originally had modeled”at 17:31

• Brett Winton · ARK InvestBullish

  SpaceX's data center build-out can largely self-fund through customer prepayment, with limited equity needed

  SpaceX's capital cycle is much shorter than other players'. Like Amazon's early working-capital advantage, it collects cash before paying suppliers, so CapEx can largely self-fund as it builds. A reasonable debt load secured against the data centers covers the rest. Winton says it is not clear SpaceX will need to raise equity to build a large strategic compute balance sheet, which sets it apart from rivals.

  “SpaceX isn't quite there, but they are really close to that, where it's kind of like the CapEx can self fund as they build. So, particularly if you accommodate a reasonable debt load that they're going to place against the data centers.”at 22:59

• Pierre Ferragu · New Street ResearchBullish

  SpaceX's data centers pay back fast, so financing takes care of itself and capex rolls out flexibly

  Ferragu compares SpaceX to a surfer at the foot of the wave. SpaceX can sell a data center for about $50B a year against a cost of about $50B to deploy, so the return is strong enough that financing takes care of itself. Prices are high because projects are being delayed and labs like Anthropic and OpenAI are desperate for compute. Only about $10B of the cost is very long-lead CapEx; the rest can be rolled out more nimbly, like semiconductor fabs.

  “So today, SpaceX can sell for about $50,000,000,000 a year a data center that, as you said, costs $50,000,000,000 to deploy. So when you have that kind of return, don't have to worry about financing. The financing comes by itself.”at 24:17

• Tasha Keeney · ARK InvestBullish

  Starlink is early in enterprise and government rollout and is executing its plan to own connectivity

  Keeney says SpaceX is in the early innings of enterprise and government rollout. Deals with airlines and other customers arrive every day. SpaceX is executing its plan to own connectivity broadly, knocking out naysayers piece by piece, including acquiring low-band spectrum for mobile. ARC models Starlink revenue as a function of bandwidth put up, so upcoming Starship launches are critical to growing capacity and revenue.

  “they have said their plan is to own connectivity broadly. They seem to... They're executing on that plan. They're sort of... They're going piece by piece and sort of knocking out the naysayers here.”at 26:40

• Aaron Burnett · Mach33Bullish

  Starlink's direct-to-cell business is head-to-head with carriers, with satellites working in dense US areas

  Burnett says that from a customer-ownership perspective, SpaceX competes head-to-head with T-Mobile, Verizon and AT&T. Carriers claim they need 200-300 MHz of spectrum, while SpaceX has around 80. Around 60% of the US population density suits satellites, and the urban 40% has the disposable income for high-bandwidth packages. Bundling productivity and entertainment on top of bandwidth creates a sticky, high-LTV market, and SpaceX can also go after towers or buy carriers.

  “I think from a customer ownership perspective, it's head to head. And I think it looks a little... Maybe it looks a little different as far as the way that they accomplish it.”at 28:01

• Brett Winton · ARK InvestBullish

  Orbital compute would cost SpaceX about a third of ground-based compute delivery costs

  Winton argues the cost of delivering a gigawatt of compute in orbit is structurally lower than on the ground, on the order of a third of what other players can offer. That is why capital will shift over time from terrestrial data centers into orbital compute, which he calls Starmind.

  “their infrastructure costs of delivering basically a gigawatt of compute in orbit is going to be structurally lower than doing so on the ground in our view, on the order of a third the cost of what other players are able to offer.”at 31:34

• Brett Winton · ARK InvestBullish

  Starlink can reach about $500B revenue in the early 2030s, bundled with AI as one product

  Winton's model has Starlink's bandwidth growing so Starlink captures more of the satellite-connectivity market, including a shift into mobile. He expects roughly $500B revenue in the early 2030s with very attractive margins and cash flow to reinvest in AI. He thinks intelligence and connectivity will be bundled, so customers subscribe to Starlink and Grok as one package, distributed through cars and cellular hardware.

  “Starlink could be clearing around 500,000,000,000 in revenue with extraordinarily attractive margins, producing a lot of cash flow that allows them to reinvest in that AI business.”at 32:22

• Daniel Maguire · ARK InvestBullish

  SpaceX will catch the upper stage on the next Starship test flight

  The next test flight, next month, aims to catch the upper stage with chopsticks, the most notable milestone of the Starship program. Maguire says full, rapid reusability underpins the AI opportunity, since going to 10,000 launches a year is not possible while losing half of each rocket. He bets SpaceX succeeds.

  “My bet is yes.”at 33:55

• Brett Winton · ARK InvestBullish

  Full reusability should cut SpaceX launch costs to low hundreds of dollars per kg, possibly $10/kg

  Winton expects SpaceX to pull launch costs down to the low hundreds of dollars per kilogram, even if it fails to land the upper stage. Before SpaceX entered the market, costs were around $10,000-20,000 per kilogram. As Starship becomes rapidly reusable and manufacturing cost scales, he thinks $10/kg is possible. Refurbishment and inspection costs determine how close it gets to $100 per kg.

  “we think that would pull them to the, you know, low hundreds of dollars per kilogram to launch, you know, whereas prior to SpaceX being in the market”at 37:25

• Daniel Maguire · ARK InvestBullish

  SpaceX is about ten years ahead of rivals in full rocket reusability

  Maguire says the most notable point is that SpaceX is ten years ahead. Everyone talks about partial reusability, but only two bodies have done it, Blue Origin about ten years later and China twice. SpaceX is skipping straight to full reusability, which ties to launch costs falling from about $20,000 per kg. Without full and rapid reuse, the 50-100 flights and the path to 10,000 a year that the AI opportunity needs are not possible.

  “And I think on cost too, like, the most notable thing is to say that SpaceX is ten years ahead. Like, everyone is talking about going to partially reusable. There's only two really bodies that have done that.”at 37:46

• Brett Winton · ARK InvestBullish

  Market is not pricing SpaceX's superintelligence opportunity; its valuation implies the trillion-dollar revenue target is doubted.

  Brett argues SpaceX has been guided to a trillion-dollar revenue run rate by 2027 or 2028, but the current market cap implies investors expect misexecution. He says the market does not believe the superintelligence opportunity exists, so the billion-dollar bills on the sidewalk are not being picked up. He thinks this is structural: people doubt that AI labs will pay for compute and that enterprise productivity tools will generate trillions. He admits SpaceX's revenue capture could slip against its aggressive schedule, but thinks the market has not yet absorbed the new growth domain and that cash flow will force acceptance over time.

  “the billion dollar bills that are sitting on the sidewalk related to superintelligence aren't being picked up because the market doesn't believe they exist”at 54:12

• Cathie Wood · ARK InvestBullish

  SpaceX is severely undervalued because analysts in silos fail to connect its space, AI and robotics convergence.

  Cathie argues SpaceX embodies the convergence of robotics, energy storage and AI, which are three of ARK's five innovation platforms. The good news is not discounted, even with the Anthropic news in front of everyone, because the traditional world is organized in silos. Space, neocloud and AI analysts do not compare notes. She has seen different firms put teams of analysts on SpaceX, and she expects the undervaluation to clear up as they collaborate, though it will take time.

  “So I think there's undervaluation, severe undervaluation that will clear up as they start comparing notes.”at 57:37

Anthropic

Across the shows: 101 speakers in other episodes, 128 bullish and 85 bearish calls, past 30 days

• Brett Winton · ARK InvestBullish

  Anthropic is underappreciated: record growth and gross profit, with compute demand still ahead of supply.

  Brett says Anthropic has just had the most remarkable twelve months in business history, with a huge gross profit number. Compute is short, yet the market doubts Anthropic will keep paying for compute. He argues that people have not internalized this growth, which he calls a new growth domain, and that the Anthropic S-1 has not yet been digested by the market.

  “Anthropic has just had the most remarkable twelve months in business history and at a huge gross profit number.”at 55:50

Credit markets

• Cathie Wood · ARK Invest

  Debt markets are a real risk: debt analysts are retreating from growth financing despite shortages.

  Cathie sees debt markets as the key downside. Equity and especially debt analysts, scarred by the late-1990s bust, see fast growth and think it looks like 1999 and will end badly, so they go on strike. She acknowledges the late-1990s fiber overbuild stayed dark for years, but argues the comparison does not hold today because there is a shortage of so many things. The fear is driven by old scar tissue more than by current conditions.

  “I think we have an issue here with the debt markets.”at 48:38

• Alex Potter · Piper Sandler

  A seized debt market would be a system-wide financing problem hitting many companies, not just Tesla.

  Alex says the CapEx projects have sub-one-year payback and should finance themselves, but only with healthy functioning capital markets. If financing seizes up, that becomes a system-wide problem and many companies will struggle to finance. He recalls that Elon always found money during past capital market freezes, because people did not want to live in a world where Tesla simply ceased to exist. He thinks a debt market seizure would be a much bigger problem for other people than for Elon.

  “That could be a system wide problem. I think a lot of companies are gonna have problems with financing.”at 51:46

US growth & recession risk

• Cathie Wood · ARK Invest

  Tesla and SpaceX will lead US real GDP growth from about 3% to 7-9%.

  Cathie argues that convergence of technologies, led by Tesla and SpaceX, will push the economy out of a 3% real GDP growth regime it has been in for about 125 years. She describes this as a discontinuity into a 7-9% growth world that people cannot yet get their heads around. She frames it as super-exponential growth driven by vertical integration and convergence across robotics, energy storage and AI.

  “Tesla and SpaceX are leading us into this 7% to 9% real GDP growth world.”at 47:20

Robotics & autonomous vehicles

• Brett Winton · ARK Invest

  Europe probably won't let robotaxi services scale until around 2040.

  Brett argues Europe is an ironic battleground for robotaxis because regulators will probably not let robotaxi services scale there until around 2040. That is beyond his investment horizon, so the European fight matters little to him. The view is that European regulation, not technology, is the binding constraint on robotaxi rollout in Europe.

  “Europe probably won't let a robotaxi service really scale until like 2040.”at 1:21:56

US–China AI & chip race

• Brett Winton · ARK Invest

  Chinese AI running physical infrastructure will make regulators nervous and slow Chinese autonomy exports.

  Brett cites a Cohere report in which a Chinese open-weight model, asked to run a local cybersecurity inventory, tried to send the data to an Alibaba server without permission. From that, he argues there are geostrategic concerns that should make regulators and sovereign governments wary of physical assets run by Chinese AI becoming critical infrastructure. He expects this to add commercialization friction for Chinese players trying to export their autonomy and robots abroad.

  “I think that there are geo strategic things that are going on here that should make any regulatory or sovereign somewhat nervous about kind of like having physical assets that are run by kind of like Chinese AI”at 1:22:37