Terafab is not simply Elon Musk's attempt to build an enormous semiconductor factory. The planned Texas complex connects a much larger system stretching from AI models and data centers to rockets, satellites, autonomous vehicles, robots and the chips underneath them. The more important question it raises is what happens when AI competition moves beyond models toward control of the infrastructure required to run them. (Reuters, August 2026; SpaceX Prospectus, June 2026)

The easiest way to misunderstand Terafab is to look only at the factory. SpaceX and Tesla are initially committing $16.8 billion to a semiconductor complex planned for Grimes County, Texas, a proposed site encompassing approximately 100 million square feet intended to combine logic fabrication, memory fabrication, advanced packaging and testing within an unusually integrated manufacturing system. Earlier county materials contemplated a $55 billion project phase and as much as $119 billion across a potential full buildout, and those figures should not be blended: they describe different planning stages, not a single committed cost. (Reuters, August 2026; TechCrunch, August 2026) SpaceX's own offering materials describe the broader objective in terms that go beyond manufacturing capacity. Terafab is intended to extend vertical integration all the way to the foundational semiconductor layer, shortening the loop between chip design, manufacturing, testing and deployment, with SpaceX expecting the facility to eventually support silicon for Tesla vehicles, Cybercab and Optimus as well as chips for planned orbital AI infrastructure. (SpaceX Prospectus, June 2026) The project is therefore about compressing an AI industrial supply chain, and that supply chain is forming while the corporate structure surrounding it is becoming increasingly consolidated. SpaceX acquired xAI in February 2026. Tesla has a financial relationship with the combined SpaceX-xAI structure. SpaceX completed the largest IPO in history in June. Tesla and SpaceX are now cooperating on semiconductor manufacturing. Intel has entered the Terafab project as an announced manufacturing-technology partner. ASML says its future equipment-capacity planning already takes Terafab demand into account. (Reuters, August 2026; The Information, February 2026; SpaceX Form 8-K, June 2026; Reuters, July 2026) Tesla and SpaceX remain legally separate public companies, but whether they stay that way has become a serious Wall Street question, making Terafab more than a semiconductor story. It is a useful place to see the industrial architecture of Musk's companies before the corporate architecture around them is settled.

What Is Terafab?

Terafab is a planned semiconductor manufacturing complex in Grimes County, Texas, being developed through cooperation between SpaceX and Tesla, with Intel participating as an announced manufacturing-technology partner. (Reuters, August 2026; TechCrunch, August 2026) The current initial phase is valued at approximately $16.8 billion, a figure that should not be confused with earlier numbers associated with the project. A May 2026 Grimes County filing described an earlier proposed initial investment of approximately $55 billion and a potential full buildout reaching approximately $119 billion, which were figures from a different stage of planning. The current $16.8 billion represents the announced initial phase while $119 billion remains a potential long-term scenario rather than committed current spending.

SpaceX describes the intended architecture as a closed-loop manufacturing system. Instead of designing a chip, sending the design into an external manufacturing chain, coordinating with separate memory and packaging suppliers and then waiting to test the resulting hardware inside a vehicle, robot or satellite, the companies want to bring more of those stages into a tighter organizational and physical loop. Current plans span approximately 100 million square feet and would combine semiconductor functions that are frequently distributed among different companies and locations: lithography-mask design, logic fabrication, memory fabrication, advanced packaging and testing. (SpaceX Prospectus, June 2026) The potential advantage is time. Chip designs can be manufactured, packaged, tested in real systems, modified and returned to production through a shorter development cycle, and that is fundamentally different from simply owning a large factory.

Why Are Tesla and SpaceX Building Their Own Semiconductor Factory?

The project originates in a concern Musk has been expressing publicly since 2025: that the semiconductor capacity available from existing suppliers may eventually be insufficient for the scale of computing his companies intend to deploy. At Tesla's November 2025 annual meeting, Musk discussed the company's fifth-generation AI chip and raised the possibility that Tesla would ultimately need enormous internal fabrication capacity because even aggressive production assumptions from existing suppliers might not satisfy future requirements. Terafab became the industrial response.

SpaceX's prospectus states that the companies expect their compute requirements to become large enough to justify extending vertical integration into chip manufacturing itself, with the long-term target described as approximately one terawatt of annual compute production capacity. (SpaceX Prospectus, June 2026) That figure requires an important clarification: SpaceX is not saying the Terafab complex will continuously consume one terawatt of electrical power, but is instead describing a target for the amount of computing capability represented by the chips produced annually. The metric is unconventional enough that its precise technical definition remains important to watch as the project develops. The underlying demand sources are easier to identify: Tesla wants increasing quantities of specialized silicon for vehicles, autonomous driving, Cybercab and Optimus; SpaceX wants computing hardware for terrestrial infrastructure and eventually for AI computing systems operating in space; and xAI requires enormous computing resources for artificial-intelligence development and deployment. Terafab is being designed underneath all three requirements.

Is Terafab Really a Tesla, SpaceX and Intel Joint Venture?

Not in the conventional sense of a completed, binding three-company joint venture. Intel's involvement is real and potentially consequential, but the legal structure remains more conditional than a simple partnership description suggests. Musk has said Tesla intends to use Intel's forthcoming 14A semiconductor process at Terafab, and Reuters has described Tesla as Intel's first major outside customer for 14A, making the relationship an important potential validation point for Intel's attempt to rebuild its contract manufacturing business. (Reuters, April 2026)

Customer intent is not the same as demonstrated manufacturing success. Intel still has to prove that 14A can achieve the yields, reliability, volume and economics required by a project on Terafab's proposed scale. SpaceX's own disclosures impose another important boundary: the existing arrangement between Tesla and SpaceX provides a framework for collaboration, but specific Terafab projects remain subject to separate negotiations and definitive agreements, meaning neither Tesla nor Intel should be described as permanently bound to every future phase of the project simply because the companies are participating today. (SpaceX SEC offering materials, June 2026) The accurate description is narrower: Terafab is a SpaceX-Tesla initiative with Intel participating as an announced manufacturing-technology partner. The distinction matters because Terafab remains a developing industrial project, not a completed manufacturing alliance.

Why Does ASML Matter to the Terafab Story?

ASML provides one of the strongest pieces of independent evidence that Terafab has moved beyond Musk publicly describing an ambition. Modern leading-edge semiconductor manufacturing depends heavily on lithography systems supplied by ASML, and in July 2026 ASML Chief Financial Officer Roger Dassen said the company's capacity expansion planning for 2027 and 2028 already takes expected Terafab demand into account. (Reuters, July 2026)

That statement does not prove Terafab will be completed successfully. It does something different and considerably more significant. It establishes that a critical upstream semiconductor-equipment company is planning around the project. ASML is close to fully booking portions of its future EUV capacity years before delivery, meaning semiconductor fabrication cannot be scaled simply by deciding to purchase equipment after a building is finished. Tool availability, installation, qualification and process development have to be anticipated far in advance. Terafab appearing inside ASML's capacity planning therefore provides independent industrial corroboration of the project's seriousness. It is one thing for Musk to announce a semiconductor factory, and another for the company controlling the world's most critical advanced lithography technology to incorporate expected demand from that factory into its own future planning.

Is Terafab Trying to Replace TSMC, Samsung and NVIDIA?

No. SpaceX's own prospectus explicitly contradicts that interpretation. The company says it expects to continue sourcing a significant portion of its compute hardware from third-party suppliers even as Terafab develops, describing Terafab as complementary to those relationships rather than a complete substitute for them. (SpaceX Prospectus, June 2026)

Terafab does not need to make Musk's companies completely semiconductor-independent to be valuable. It needs to create optionality. Internal manufacturing capacity could give Tesla and SpaceX another source of specialized silicon, increase their ability to experiment with custom designs, reduce exposure to capacity constraints and potentially strengthen their negotiating position with external suppliers. This resembles vertical integration without requiring total autarky. TSMC, Samsung, NVIDIA and other outside suppliers can remain important while Terafab creates an internal manufacturing path for chips whose design or production requirements are particularly strategic. The actual objective is therefore better described as supply leverage and development control than complete semiconductor independence.

Why Would Terafab Produce Two Different Kinds of AI Chips?

Because the physical environments in which Tesla and SpaceX intend to deploy computing systems are fundamentally different. SpaceX's disclosures describe terrestrial silicon intended for systems including Tesla vehicles, Cybercab and Optimus, alongside a separate class of silicon designed for orbital computing infrastructure. (SpaceX Prospectus, June 2026) A processor operating inside a vehicle or humanoid robot does not face exactly the same constraints as hardware intended to operate in space: power budgets differ, thermal conditions differ, radiation exposure differs, reliability requirements differ and packaging constraints differ.

Terafab therefore should not be understood as one enormous production line producing one universal Musk AI processor. The project is intended to support at least two distinct hardware environments, and that makes the closed-loop development concept more consequential. Tesla can feed information from physical AI systems on Earth back into chip development. SpaceX can do the same from orbital systems. Manufacturing can theoretically evolve alongside the machines actually using the silicon. If that loop works, the competitive advantage would not simply be chip volume but iteration speed, which is a different form of vertical integration from merely reducing component costs.

How Does xAI Fit Into Terafab?

xAI is no longer merely another Musk company cooperating with SpaceX. It is inside SpaceX. SpaceX acquired xAI in February 2026 in an all-stock transaction reported at approximately $250 billion for xAI, against a SpaceX valuation of roughly $1 trillion at the time. The transaction brought xAI's models, computing infrastructure and the X platform inside SpaceX's corporate perimeter. (The Information, February 2026) That changed the context around Terafab substantially. Before the acquisition, Musk controlled several companies that could cooperate around AI infrastructure. After the acquisition, significant portions of that stack became one company, with SpaceX now containing rockets, Starlink communications infrastructure, xAI, X, terrestrial data-center ambitions and planned orbital AI infrastructure.

Tesla remains outside that corporate perimeter but is increasingly connected to it. Tesla agreed in January 2026 to invest approximately $2 billion in xAI Series E preferred shares. Following SpaceX's acquisition of xAI, that right converted into approximately $2 billion of SpaceX Class A shares, with Tesla completing the converted investment in March. Tesla and SpaceX are therefore not simply two unrelated Musk companies discussing a factory. One now owns the AI company. The other has invested financially into that combined structure. Both are participating in Terafab. That does not constitute a Tesla-SpaceX merger, but it does explain why the possibility of one is no longer being treated exclusively as internet speculation.

How Big Was SpaceX's 2026 IPO?

SpaceX's June 2026 IPO was the largest initial public offering on record, and the exact numbers require more precision than the shorthand figure normally attached to it. On June 11, SpaceX priced a base offering of 555,555,555 Class A shares at $135 per share, representing approximately $75 billion in base offering value, more than double Saudi Aramco's previous record. Trading began June 12 under the ticker SPCX. (SpaceX IPO pricing announcement, June 2026)

The transaction expanded when underwriters exercised their option in full. At closing on June 15, SpaceX had issued a total of 638,888,888 shares, including an additional 83,333,333 shares through the overallotment option. At $135 per share, that represents approximately $86.25 billion in aggregate offering value. SpaceX's offering materials identify approximately $85.7 billion, specifically about $85.675 billion, as net proceeds after underwriting discounts, commissions and offering expenses. (SpaceX Form 8-K, June 2026; SpaceX Form 10-Q and FWP, June 2026) Those three figures answer different questions: approximately $75 billion describes the original base offering; approximately $86.25 billion describes the aggregate value of all shares sold after full exercise of the underwriter option; and approximately $85.7 billion describes the resulting net proceeds to SpaceX after offering expenses. The company emerged from the transaction with extraordinary access to capital at precisely the moment it was moving deeper into AI infrastructure and semiconductor manufacturing, making Terafab a project sitting inside a corporate structure with financing capacity that would have been difficult to imagine for a private aerospace company only several years earlier.

Does Elon Musk Control SpaceX?

Yes, but economic ownership and voting control should not be confused. SpaceX's post-IPO share structure gives Musk approximately 82 percent of the company's voting power through super-voting Class B shares. That does not mean Musk economically owns 82 percent of SpaceX. Voting control determines how much influence a shareholder can exert over corporate decisions, while economic ownership determines the shareholder's financial interest in the company. SpaceX's structure gives Musk unusually strong governance control even after the company became publicly traded, and Tesla has a different ownership and governance structure. Any attempt to combine the two companies would therefore have to reconcile two public corporations, two shareholder bases, transaction valuation, governance rights, potential conflicts of interest and regulatory considerations, which is considerably more complicated than Musk deciding that his companies should become one company.

Are Tesla and SpaceX Actually Going to Merge?

No merger has been announced. The possibility has nevertheless moved far enough into mainstream financial analysis that dismissing it as fringe speculation would also be inaccurate. Wedbush analyst Dan Ives placed the probability of a Tesla-SpaceX combination at approximately 80 percent in June 2026. Deepwater Asset Management's Gene Munster separately placed the probability at approximately 90 percent over a two-year window. (TheStreet, Seeking Alpha, Barchart, June 2026; Deepwater Asset Management, 2026)

ARK Invest has offered another variation. ARK researchers said in August 2026 that they expect a merger to be announced before the end of 2026 while explicitly distinguishing an announcement from actual completion. Their reasoning acknowledges several substantial complications, including the premium SpaceX could have to offer Tesla shareholders and the geopolitical consequences of combining a major U.S. defense contractor with a company whose manufacturing economics remain meaningfully connected to China. Prediction markets have been more cautious: Fortune reported that Kalshi contracts in late May were pricing the probability of a Tesla-SpaceX merger before May 2027 at approximately 52 percent, a historical snapshot that should never be carried forward without a fresh timestamp since prediction-market prices change continuously. (Fortune, May 2026) The disagreement itself is informative. Some professional analysts see consolidation as highly likely while market participants have at times priced materially lower odds. No definitive transaction exists. The correct conclusion is not that Tesla and SpaceX are merging but that a merger has become a credible enough possibility to require serious analysis.

Why Would Tesla and SpaceX Merge?

The strongest argument is not simply that Musk controls both companies. It is that their industrial systems increasingly overlap. Tesla builds physical AI systems on Earth: vehicles, autonomy, robotics, energy and the infrastructure surrounding them. SpaceX increasingly controls another part of the stack: communications infrastructure, orbital systems, xAI, X and planned space-based computing. Terafab would sit beneath both companies as semiconductor infrastructure.

A combined organization could therefore contain a remarkably broad vertical stack: AI models from xAI; training and inference through terrestrial and eventually orbital infrastructure; connectivity through Starlink; custom silicon partly from Terafab; terrestrial deployment through Tesla vehicles, Cybercab and Optimus; orbital deployment through SpaceX systems; and energy systems through Tesla. The corporate logic is visible, but that does not mean the transaction logic is easy. Tesla shareholders would have to receive acceptable economics. SpaceX shareholders would have to believe the combination created more value than remaining independent. Governance conflicts would have to be addressed. Defense relationships would require scrutiny. Tesla's exposure to China would become relevant to a company deeply integrated into U.S. national-security infrastructure. The industrial architecture may increasingly look unified before the legal architecture does, and that tension is the real merger story.

Would a Tesla-SpaceX Merger Automatically Give Musk a $1 Trillion Pay Package?

No. That claim compresses several distinct provisions of Tesla's 2025 CEO Performance Award into a conclusion the documents do not support. The award covers approximately 423.7 million performance-based restricted shares and contains specific treatment for a change in control. Separately, Tesla defines a mechanism called Deemed Achievement that applies when qualifying Covered Events outside Tesla's control materially impair the company's ability to achieve certain product goals. (Tesla, 2025 CEO Performance Award disclosure) Those are different provisions, and the Deemed Achievement mechanism should not be described as Tesla's change-of-control clause.

A sufficiently valuable acquisition or merger could affect how portions of Musk's award are evaluated, depending on the transaction structure, Tesla's valuation and the precise operation of the award terms. But there is currently no Tesla-SpaceX transaction, no known transaction valuation and no known exchange ratio. There is therefore no basis for saying that a hypothetical merger automatically produces a $1 trillion payment to Musk. The compensation implications are real enough to analyze while the automatic trillion-dollar-payout formulation is not.

What Would a Tesla-SpaceX Merger Actually Create?

If a merger eventually occurred, describing the result as an automaker combined with a rocket company would badly understate what had been assembled. SpaceX already contains xAI, bringing artificial-intelligence development and the X information platform. SpaceX also brings launch systems, Starlink, satellite manufacturing and planned orbital computing. Tesla brings vehicles, autonomy, robotics, energy storage, manufacturing and enormous numbers of physical devices capable of becoming endpoints for AI. Terafab would potentially provide custom semiconductor infrastructure underneath both sides.

The result would resemble an integrated AI-industrial platform more than a conventional merger between companies operating in adjacent markets, and that distinction matters because the AI industry has spent much of the current cycle competing at the model layer: who has the best reasoning system, the largest context window, the strongest coding performance, the most capable multimodal architecture. Those questions remain important. Terafab introduces another one: who controls what the models run on.

Is AI Competition Moving From Models to Infrastructure?

Increasingly, both layers matter. Frontier models require enormous quantities of computation, and that computation requires accelerators, memory, packaging, networking, electricity, cooling, buildings and access to semiconductor manufacturing capacity. As model development scales, infrastructure stops being merely a support function and becomes a competitive constraint. SpaceX says this explicitly in its own prospectus, describing Terafab as a way to extend its vertically integrated compute platform into the foundational chip layer and arguing that controlling more of that chain can become a competitive advantage in scaling AI infrastructure. (SpaceX Prospectus, June 2026)

Musk's companies are not abandoning model competition: xAI remains part of that race. They are attempting to build underneath it. If successful, the system would not depend exclusively on purchasing every critical component from an external supplier and assembling those components inside third-party cloud infrastructure. It would control more stages of the chain, from models and data through deployment and connectivity to compute infrastructure, chip design and some chip manufacturing, and potentially, through SpaceX's orbital ambitions, the physical location of portions of the compute itself. That is what makes Terafab different from another semiconductor-capacity announcement.

Why Is the Closed-Loop Design Potentially More Important Than Terafab's Size?

Because semiconductor development is an iterative process. A chip is not valuable because a factory can physically manufacture it. It is valuable if the chip performs correctly, can be manufactured at acceptable yield, costs an economically viable amount and can be improved faster than competing designs. Traditional semiconductor supply chains distribute those activities across specialized companies: one designs the processor, another fabricates it, separate suppliers provide memory and packaging, and the finished system eventually reaches the company deploying it.

Terafab's proposed architecture compresses more of those steps into a single organizational loop, and for Tesla and SpaceX the deployment endpoints are also unusually accessible. (SpaceX Prospectus, June 2026) Tesla can test chips inside vehicles and robots. SpaceX can test chips inside communications and orbital systems. Performance information can theoretically return directly to the design and manufacturing organizations responsible for the next revision. If the system works, the important advantage may not be producing the greatest number of chips but learning faster, and that is a different form of vertical integration from merely reducing component costs.

What Could Make Terafab Fail?

Scale alone does not make a semiconductor factory successful. Terafab has to manufacture usable silicon at competitive yields. Yield measures how much of the silicon moving through an extraordinarily complicated fabrication process becomes functional, economically valuable product, and poor yields can destroy the economics of a fab regardless of how impressive the building appears. Equipment uptime, defect density, process maturity, memory integration, advanced packaging yield and workforce development all matter, and Intel's 14A process has to perform at the level required while the enormous facility itself has to ramp without turning into underutilized industrial capacity.

Capital intensity adds another layer of risk. The currently announced $16.8 billion initial phase is already substantial, and earlier project materials contemplated total investment potentially reaching $119 billion. A project of that magnitude has to produce enough valuable silicon for enough years to justify the capital deployed into it. Terafab should not be judged by whether Musk succeeds in constructing enormous buildings in Texas. The meaningful test begins when wafers start moving through them.

What About Water, Power and the Local Community?

The industrial scale that makes Terafab strategically interesting also creates real local consequences. The Grimes County site is adjacent to Gibbons Creek Reservoir, and the companies have stated that industrial water would be drawn from the reservoir rather than relying primarily on local groundwater. (Reuters, August 2026) That is a mitigation plan, not proof that every environmental or resource concern has been resolved. Large semiconductor facilities require significant water, energy, chemicals, transportation infrastructure and industrial support. Local opposition has emerged around environmental impact, resource use, tax incentives and disruption to the surrounding rural community, while state and local officials have simultaneously promoted the project for investment and employment. Both realities belong in the story. Terafab could strengthen domestic semiconductor capacity and create high-value employment while still imposing real environmental and infrastructure costs on the region hosting it, and those outcomes are not mutually exclusive.

Is Terafab a National-Security Project?

Terafab is a commercial project, but it sits directly inside a larger U.S. national-security concern: advanced semiconductor supply. Much of the world's most advanced semiconductor manufacturing remains concentrated in East Asia, and the United States has spent years attempting to expand domestic manufacturing because disruptions involving Taiwan, China, Korea or critical international shipping routes could affect industries extending far beyond consumer electronics. Terafab would add substantial domestic capacity if it reaches its intended scale.

Manufacturing chips in Texas does not make the entire supply chain American, however. Leading-edge fabs remain dependent on globally distributed equipment, materials, chemicals, intellectual property and specialized expertise: ASML itself is Dutch, and critical semiconductor materials and equipment originate across multiple countries. Terafab could therefore reduce one category of dependency without eliminating global semiconductor interdependence. The national-security benefit, if the project succeeds, is additional domestic capacity and optionality, not complete supply-chain sovereignty.

What Does Terafab Mean Two or Three Years From Now?

The most important consequences of Terafab may not be visible when the buildings first appear. If Intel 14A succeeds inside the program, Intel gains a major outside customer and potentially one of the most demanding validation environments available for its foundry technology. If Terafab achieves competitive yields, Tesla and SpaceX gain internal semiconductor capacity that can reduce their exposure to external shortages while accelerating custom-chip development. If SpaceX's orbital AI plans advance, Terafab could become part of a supply chain connecting semiconductor manufacturing in Texas directly to computing hardware deployed in orbit. If Tesla's robotics and autonomous-vehicle programs scale, the same manufacturing system could support enormous terrestrial demand. And if Tesla and SpaceX eventually merge, Terafab would cease to be an industrial collaboration between two Musk companies and become internal semiconductor infrastructure inside a single corporate system. None of those outcomes is guaranteed, but the architecture required for them is being assembled now, and that is why Terafab deserves to be watched before its final corporate consequences are obvious.

The Bigger Picture

The defining AI company of the next decade may not be the company with the best model on a particular benchmark. It may be the organization that can repeatedly turn intelligence into physical capability at enormous scale. That requires more than models: it requires capital, semiconductors, energy, manufacturing, networks, deployment platforms and, increasingly, control over the interfaces connecting those layers.

Musk's companies are assembling unusually large portions of that system. SpaceX has already absorbed xAI. Tesla remains separate but financially and industrially connected. Starlink supplies global connectivity. Tesla supplies physical AI endpoints through vehicles and robotics. SpaceX is pursuing orbital computing. Terafab is intended to bring semiconductor manufacturing deeper inside the system. The unresolved Tesla-SpaceX merger question sits above all of it. A merger would be consequential, but it is not necessary for the underlying industrial convergence to be real. Much of that convergence is already happening through ownership, investment, supply relationships and shared infrastructure. Corporate consolidation does not always begin with a merger agreement. Sometimes the infrastructure converges first, and Terafab is where that convergence becomes physical.

Fact Summary

What is Terafab? Terafab is a planned semiconductor manufacturing complex in Grimes County, Texas, being developed through a SpaceX-Tesla initiative with Intel participating as an announced manufacturing-technology partner. (Reuters, August 2026; TechCrunch, August 2026)

How much will Terafab cost? The currently announced initial phase is approximately $16.8 billion. An earlier May proposal described approximately $55 billion for a project phase and as much as $119 billion for a potential full buildout. Those are different planning figures from different stages and should not be combined. (Reuters, August 2026; TechCrunch, August 2026)

How large is Terafab? Current plans describe approximately 100 million square feet of manufacturing space. (Reuters, August 2026)

What will Terafab manufacture? SpaceX describes an integrated system spanning lithography-mask design, logic and memory fabrication, advanced packaging and testing, with chips intended for both terrestrial and orbital applications. (SpaceX Prospectus, June 2026)

Is Intel involved? Yes. Musk has said Tesla intends to use Intel's 14A process, but specific future project arrangements remain subject to negotiation and Intel's ability to deliver the necessary high-volume manufacturing performance remains unproven. (Reuters, April 2026; SpaceX SEC offering materials, June 2026)

Has ASML independently acknowledged Terafab? Yes. ASML CFO Roger Dassen said the company's 2027 and 2028 capacity planning takes expected Terafab demand into account. (Reuters, July 2026)

Will Terafab replace TSMC, Samsung and other suppliers? No. SpaceX explicitly says it expects to continue sourcing a significant portion of compute hardware from third parties. (SpaceX Prospectus, June 2026)

Does SpaceX own xAI? Yes. SpaceX acquired xAI in February 2026 in an all-stock transaction reported at approximately $250 billion for xAI. (The Information, February 2026)

How large was SpaceX's IPO? The June 2026 IPO began as a roughly $75 billion base offering. After full exercise of the underwriter option, approximately 638.9 million shares had been sold at $135 each, representing approximately $86.25 billion in aggregate offering value and approximately $85.7 billion in net proceeds after underwriting discounts and offering expenses. (SpaceX Form 8-K, June 2026; SpaceX Form 10-Q and FWP, June 2026)

Are Tesla and SpaceX merging? No merger has been announced. Analysts including Dan Ives and Gene Munster have assigned high probabilities to a future combination, while prediction-market pricing has at times been substantially more conservative. These are forecasts, not evidence that a transaction exists.

Would a Tesla-SpaceX merger automatically give Musk $1 trillion? No. Tesla's 2025 CEO Performance Award contains specific change-of-control treatment, while the separately defined Deemed Achievement mechanism applies to qualifying Covered Events. A hypothetical merger's effect on Musk's award would depend on the actual transaction structure, valuation and award terms. (Tesla, 2025 CEO Performance Award disclosure)

What is Terafab's long-term target? SpaceX describes a goal of approximately one terawatt of annual compute production capacity. This is a compute-output target, not a statement that the facility will continuously consume one terawatt of electricity. (SpaceX Prospectus, June 2026)

Why does Terafab matter to AI? It represents an attempt to extend vertical integration below AI models and computing infrastructure into the semiconductor layer itself, potentially connecting chip design, fabrication, packaging, testing and deployment through a shorter development loop. (SpaceX Prospectus, June 2026)

Sources

Reuters. "SpaceX, Tesla to Initially Spend $16.8 Billion on Terafab Chip Plant in Texas." August 6, 2026. Reuters

Space Exploration Technologies Corp. Prospectus. June 2026. Terafab strategy, one-terawatt annual compute-production target, third-party sourcing and vertically integrated semiconductor architecture. SpaceX

Space Exploration Technologies Corp. SEC registration materials. Terafab collaboration and vertical-integration disclosures. SEC

Reuters. Reporting on SpaceX, Tesla and Terafab semiconductor development. 2026. Reuters

Reuters. "Factbox: Elon Musk Lays Out Terafab AI Chip Project Plan." April 23, 2026. Intel 14A and terrestrial/orbital chip programs. Reuters

Space Exploration Technologies Corp. SEC registration and offering materials. Tesla-SpaceX collaboration framework and Terafab project disclosures. SEC

Reuters. "ASML Financial Guidance Includes Terafab Plans, CFO Says." July 15, 2026. Reuters

The Information. Reporting on SpaceX's acquisition of xAI. February 2, 2026.

SpaceX. "Space Exploration Technologies Corp. Announces Pricing of Initial Public Offering." June 11, 2026.

Space Exploration Technologies Corp. Form 8-K. June 15, 2026. 638,888,888 shares at $135 per share following full exercise of the underwriters' option. SEC

Space Exploration Technologies Corp. Form 10-Q and Free Writing Prospectus. June 2026. IPO proceeds disclosures. SEC

TheStreet, Seeking Alpha and Barchart. June 2026 reporting on Wedbush analyst Dan Ives' Tesla-SpaceX merger forecast.

Deepwater Asset Management. Gene Munster commentary on probability of Tesla-SpaceX combination. 2026.

Fortune. "SpaceX-Tesla Merger Would Create a $3.4 Trillion Company, With No Profits." May 31, 2026. Kalshi probability cited as a dated late-May market snapshot. Fortune

Tesla, Inc. 2025 CEO Performance Award disclosure. Change-in-control treatment and separately defined Deemed Achievement based on Covered Events.

TechCrunch. "Tesla and SpaceX will invest $16.8B to start building 'Terafab' chip factory in Texas." August 6, 2026. TechCrunch