Musk Accelerates "Superintelligence" Ambitions: SpaceX (SPCX.US) and Tesla, Inc. (TSLA.US) Discuss $3.5 Trillion Merger, ASML Holding NV ADR Expedites Delivery for Terafab

After the U.S. market close on Tuesday, during SpaceX's (SPCX.US) Q3 earnings call, Musk tied together two threads that had each been circulating separately. Both point to the same thing: across the AI compute chain, Musk is simultaneously solving the two problems of "can't build it" and "can't combine." One is chips: he said ASML Holding NV ADR (ASML.US) and other key equipment suppliers are already accelerating equipment deliveries for Tesla, Inc.'s (TSLA.US) and SpaceX's Terafab project. "These dates are very aggressive, and we are very grateful to ASML Holding NV ADR and all our other suppliers for accelerating equipment deliveries to achieve this result." The other is capital markets activity: SpaceX and Tesla, Inc. have each established special committees composed of independent directors to explore the possibility of a mergerthe most complete public statement to date on this potential transaction. Terafab's timeline: test wafer in May 2027, mass production in Q3 2028 This time Musk gave the first set of specific internal milestones for Terafab, and proactively added a qualifier"with a wide margin of error." According to him, the research fab in Austin aims to get its first test chip running in May or June of next year; this chip is "like an SRAM chip," used to validate the production line; after that, the goal is mass production in Q3 2028 at the main campus, Gibbons Creek. Only by putting these two milestones back into the project itself can their significance be seen. Terafab was officially announced in Austin on March 21, 2026, jointly advanced by SpaceX, Tesla, Inc., and xAI, with the goal of producing 1 terawatt (1 TW) of AI compute per yearthe reason SpaceX gave in the project announcement was that "the combined chip demand of SpaceX and Tesla, Inc. is expected to exceed 1 TW of compute, significantly greater than current global supply." The campus is located at the former Gibbons Creek coal-fired power plant site in Grimes County, Texas, covering more than 6,000 acres, about 90 miles northwest of Houston; the groundbreaking ceremony was held on October 6, land clearing began in August, formal construction is set for December 1, and full completion will not come until the end of 2028. The investment figures are likewise layered: Phase 1 investment is $16.8 billion, confirmed by SpaceX's August 6 announcement and Texas Governor Abbott; according to the incentive application filed with Texas, the total project value after full completion could reach up to $119 billion. The Texas Enterprise Fund provides a $30 million grant for Phase 1, and the county commissioners court passed a 100% property tax abatement in June by a 4-to-1 vote. The real manufacturing capability is not in Musk's hands. Intel Corporation (INTC.US) joined the project on April 7, committing to cover the full vertical chain of design, manufacturing, and packaging; Terafab's mass production will use Intel Corporation's 14A process; on October 2, market rumors emerged that Taiwan Semiconductor Manufacturing Co., Ltd. Sponsored ADR (TSM.US) might participate, and on October 7, Musk drew a clear line on X"We will build and operate this fab," and Taiwan Semiconductor Manufacturing Co., Ltd. Sponsored ADR could at most sublease a portion. Yole Group analyst Adrien Snchez put it more bluntly: Terafab "is more like an Intel Corporation fab expansion, with Tesla, Inc., SpaceX, and xAI as anchor customers." ASML Holding NV ADR's role: the only bottleneck, and also the most direct beneficiary Musk's named thanks landed on a company whose capacity is already booked out to 2030. ASML Holding NV ADR reported record Q3 results on October 14: net sales of 11.662 billion, up 25.0% quarter over quarter and 55.2% year over year, but slightly below the Wall Street average estimate of 11.735 billion; full-year net sales guidance was raised to 44 billion45 billion, and gross margin guidance was raised to about 56%. CEO Christophe Fouquet had already confirmed direct communication with Musk about Terafab at the time, describing him as "very serious," while adding a preconditionthe project can only be called an opportunity if ASML Holding NV ADR itself does not become a supply bottleneck, and the company is not yet doing that even for its existing customers. Equipment may arrive earlier than the building. According to chip industry observer Tim Culpan, citing people familiar with the matter, Terafab's first ASML EUV lithography machine may be delivered in the middle of next year; while SemiAnalysis estimates show that the main structure of the Austin research fab will not be completed until the end of 2027, with equipment move-in even later, meaning this high-priced single machine will sit idle for quite a while. For a traditional chipmaker, that is a heavy financial burden; for Terafab, it is not. Why insist on building it yourself? An earnings report that writes AI into the core business The urgency of Terafab can be quantified in the earnings report from the same evening. SpaceX's Q3 revenue was $14.575 billion, up 161% year over year ($5.576 billion in the year-ago period); operating profit was $2.791 billion and net profit was $4.404 billion, both turning profitable; adjusted EBITDA was $8.809 billion, up 306% year over year; GAAP diluted earnings per share were $0.35. All three exceeded analyst expectations compiled by S&P Global Market Intelligence (revenue of about $12.8 billion, net profit of about $2 billion, adjusted EBITDA of about $6.8 billion). The structural shift matters more than the total: the AI business, which the company calls "Super Intelligence," had Q3 revenue of $7.56 billion, up 830% year over year, surpassing Starlink for the first time to become the largest revenue source, accounting for about 52% of total revenue, and posting $431 million in operating profit. DRIVE is cloud compute leasesAnthropic agreed in May to pay nearly $45 billion over three years to SpaceX's Colossus data centers, and Alphabet Inc. Class C signed in June to pay $920 million per month from October through June 2029 (AFP). The cost is written in capital expenditure: Q3 capex was $19.593 billion, of which $16.4 billion went to the Super Intelligence business, for a nine-month total of about $48 billion. Compute capacity stood at 1.7 GW as of September 30, versus 0.6 GW a year earlier; Musk said it will reach 2.53 GW by year-end, with confidence of adding another 10 GW in 2027, and even "a slight possibility" of 20 GW, and he expects SpaceX's annualized recurring revenue (ARR) to reach $500 billion to $1 trillion by the end of 2027. How aggressive that range is can be calculated: Q3 revenue of $14.575 billion annualizes to about $58.4 billion; to reach $500 billion in 15 months would mean ARR growing at least 8.5 times. Put alongside Terafab, the logic is self-consistentwhen the slope of compute demand is this steep, the supply chain itself is the biggest constraint, and Musk's solution is to build the fab himself. Merger signal: from "can't discuss on an earnings call" to special committees What really made the market hold its breath was the second item. In July, when asked by a Wells Fargo & Company analyst about a merger of the two companies, Musk's answer was that "this kind of thing can't be discussed on an earnings call; it needs to go through the proper process." On October 20, he gave the process itself: "SpaceX and Tesla, Inc. have both established special committees at shareholders' request to discuss this idea, and these special committees are composed of independent directors. Since I am an interested party, I actually cannot participate in the discussion. It's a bit like a bizarre kabuki dance, with two groups talking and I can't be part of it." He also stressed three boundaries: the special committees have not agreed on any transaction terms, and "obviously there can be no guarantee" they will; the deal needs shareholder support, and "on the Tesla, Inc. side, 50% of all shares must vote in favor, not just of the shares that vote"which he called "a fairly high threshold"; "some people outside think the merger is a done deal, and I think that's somewhat exaggerated." But he also stated his own view without reservation: "Do I think it's a good idea? Yes." The reason is the synergy listSuper Intelligence software, chip design, Terafab, Siasun Robot&Automation technology, orbital launch capacity for the StarMind constellation, and "connectivity," which President Shotwell added on the spot. If it happens, the scale would be unprecedented: the two companies are worth about $2 trillion and $1.5 trillion, respectively. The governance tension is more noteworthy. According to SpaceX's IPO filing, Musk holds more than 82% of the voting power at SpaceX, versus about 20% at Tesla, Inc.; Barclays analyst Dan Levy estimates that if SpaceX acquired Tesla, Inc. in an all-stock deal, Musk would hold about 37% of the economic interest and about 56% of the voting power in the combined companya "minority equity, majority control" transaction. The market is already pricing it, but disagreement is also clear The sell side has laid out a fairly specific path. Barclays' Dan Levy judges the deal could be announced before Tesla, Inc.'s annual shareholder meeting (the company has announced it for December 3), with SpaceX acquiring Tesla, Inc. at a premium in an all-stock deal; based on the SpaceX $2.1 trillion and Tesla, Inc. $1.5 trillion valuations he cites, Tesla, Inc. shareholders would receive about 0.7 SpaceX shares per share; if that window is missed, an announcement could still come within 6 to 12 months. Wedbush's Dan Ives puts the probability of completing the merger by the end of 2027 at "over 80%." RBC Capital has already incorporated a 25%30% SpaceX acquisition premium into its Tesla, Inc. target price calculation. Skeptics are equally direct. The Future Fund's Gary Black warned bulls to "stop hyping the TSLA-SPCX deal," arguing that a stock-for-stock merger would be priced at Tesla, Inc.'s lower valuation multiple and that arbitrage would erode the premium paid for SpaceX shares before closing; Morningstar earlier this year flagged that valuation disagreements could be large, especially on the Tesla, Inc. shareholder side. On target prices, the sell side is not far apart but broadly conservative: the 12-month average from 38 analysts is $397.83, with JPMorgan at $415, Morgan Stanley at $400, Goldman Sachs Group, Inc. at $360, and UBS Group AG at $391, mostly neutral or hold ratings. Prediction markets give another set of numbers: on Polymarket, the contract for "formal merger announcement before the end of 2026" is priced at about 23.5% (up 6 points within 24 hours), and "before the end of 2027" at about 64.5%that is traders' pricing, not the existence of an agreement. The price tag on ambition: $48 billion in capex meets a tightening financing environment Taken together, what Musk has offered is a plan to "trade balance sheet for time." The hard constraints on progress remain clear: SemiAnalysis estimates the research fab needs 15 months to build, plus about 6 months to install equipment before it can start running wafers; Intel Corporation's 14A process transition itself requires extensive equipment qualification; ASML's High-NA EUV capacity will still need to prioritize existing customers through 2028. Musk's own qualifier for these timelines is "with a wide margin of error." The funding constraint is equally real. SpaceX's capex in the first nine months was about $48 billion, with Q4 capex guidance of $35 billion$40 billion. Meanwhile, financing conditions for AI infrastructure are weakening: CoreWeave and Nebius shares both fell about 28% in October, IREN fell 35%, and Australian data center operator Firmus withdrew its IPO plan citing market volatility. Procedural uncertainty is the most easily underestimated. As of now, the existence of the special committees comes only from Musk's verbal statement on the earnings callit is not mentioned in SpaceX's earnings filings or in any Tesla, Inc. document, and neither company has announced committee members or any transaction proposal (Drive Tesla Canada). In addition, the deal would need to pass two shareholder votes, antitrust review, and the national security review that could be triggered by a defense contractor absorbing an automaker with deep China exposure. In other words, the signal Musk released that evening is clear and coherentequipment has been expedited, committees are in place, and compute targets are written into guidance; but from "signal" to "fact," what stands in between is exactly the threshold he repeatedly emphasized: the shareholder vote.