From a $2 billion acquisition to the final narrative before a trillion-dollar IPO.

In the late hours of April 23, 2026, Tesla filed a quarterly report with the SEC. Buried in a footnote on the last page, in less than two lines of text, a transaction was disclosed: Tesla agreed to acquire an AI hardware company for up to $2 billion, paid entirely in common stock and equity incentives, with approximately $1.8 billion contingent upon achieving technical milestones.

Image source: Internet

This filing came just hours after Tesla released its Q1 earnings report. The report showed a year-over-year decline in revenue and profits falling short of expectations, with after-hours stock price dropping over 3%.

But market attention was entirely captured by those few lines hidden at the end of the footnotes.

This wasn't Tesla's first quiet integration of Musk's AI assets. In January 2026, Tesla had already invested approximately $2 billion to acquire shares in xAI; in February, SpaceX announced its acquisition of xAI with a merged valuation of $1.25 trillion; on April 22, SpaceX secured an option to acquire AI programming unicorn Cursor for $60 billion.

Image source: Internet

Combined with the March 21 announcement of the Terafab chip factory plan, multiple Musk tweets on X, and the impending mass production timeline for Optimus Gen3—this capital layout (capital deployment) spanning from Earth to space, from electric vehicles to humanoid robots, can no longer be explained merely as "expansion." It resembles more a person preparing for a ultimate exam.

01

Who Exactly Is the Acquired AI Hardware Company?

As of now, Tesla has not disclosed the name of the acquired company. Major media outlets have used terms like "unnamed" or "unclear" in their reports. This is highly unusual—typically, for mergers and acquisitions exceeding $1 billion, both parties disclose names in announcements, or at least provide sufficient information in filings for market speculation, even during confidentiality periods.

Tesla's choice to remain silent suggests only two possibilities: either the target's shareholder structure is overly complex, or the transaction itself carries high uncertainty.

However, considering the context, several directional speculations warrant tracking:

Direction 1: Inference chip companies. Tesla's self-developed Dojo supercomputer is primarily used for training, while inference still heavily relies on NVIDIA GPUs. Providing local inference capabilities for Optimus and FSD requires customized inference chips. Current reasonably valued targets with technology in AI inference chips include Groq (partially integrated by NVIDIA), Cerebras, and some smaller startups.

Image source: Internet

Direction 2: Network interconnect hardware companies. Communication bandwidth between chips is a critical bottleneck for AI cluster performance. If Tesla is to build large-scale computing clusters (such as Terafab), high-speed network switching chips and optical modules are indispensable components.

Direction 3: Sensor/actuator-related hardware. Mass production of Optimus requires large quantities of high-precision sensors and joint actuators, which are also the most scarce links in the current supply chain.

Image source: Internet

Regardless of the target's identity, the $2 billion valuation implies the company already possesses certain scale and productization capabilities, rather than being a pure early-stage startup.

02

What Did Musk Say on X?

Around the time of Tesla's quarterly filing, Musk's posts on platform X provided the most direct annotations for the acquisition's intent.

March 11: Musk confirmed that Tesla and xAI's joint project "Macrohard" (translated as "Jù Yìng" by Chinese media) officially entered full-scale advancement. Led by Tesla AI software vice president Ashok Elluswamy, the project aims to simulate a complete software company's functions using AI—Grok serves as the high-level decision-maker, while Tesla's AI agents handle real-time screen, keyboard, and mouse operations to execute specific software development tasks.

Image source: Internet

March 21: Musk announced the Terafab plan—a joint venture by Tesla, SpaceX, and xAI to build the world's largest chip manufacturing facility in Austin, Texas, targeting annual production of 1 terawatt (1TW) of AI computing power, roughly half of current global total capacity.

Image source: Internet

Mid-April: Musk posted on X that "Tesla will be one of the companies to achieve AGI, and quite possibly the first to implement AGI in a 'humanoid + atomic-level shaping' form." This tweet was widely cited by the industry as a clear signal of Musk directly binding humanoid robots with artificial general intelligence (AGI).

April 23: After earnings, Musk did not directly mention the $2 billion acquisition on X but retweeted multiple promotional contents about Optimus Gen3's impending mass production, adding the comment "Tesla is building the most important product on Earth."

Image source: Internet

These posts share a common narrative thread: Tesla is not just an electric vehicle company but an AI+hardware company, with humanoid robot Optimus serving as the ultimate carrier of its AI capabilities. The $2 billion acquisition represents the latest piece in this narrative puzzle—providing more autonomous computing power support for Optimus's scaled production.

03

Optimus Gen3 Mass Production Imminent: Real Progress on Humanoid Robots

On April 23, Tesla announced via its official Weibo account that the third-generation Optimus (Gen3) is expected to officially debut by mid-year, with plans to initiate large-scale production at its Fremont, California factory between July and August 2026, targeting an annual output of 1 million units.

Image source: Internet

This marks Tesla's first explicit timeline for humanoid robot mass production. Previously, the planned production of 5,000 Optimus units in 2025 had been postponed multiple times. The Fremont factory is undergoing production line modifications, with Model S and Model X lines halted in the first half of this year to make capacity for robots.

Key Gen3 upgrades: A newly designed robotic hand with significantly improved precision and degrees of freedom; comprehensive upgrades to joint motors throughout the body, with maximum torque output exceeding Gen2 by approximately 40%; equipped with an inference version of Tesla's self-developed FSD chip for real-time local environmental perception.

Image source: Internet

Meanwhile, the advancement of the Macrohard project opens a new usage scenario for Optimus—robots in factories not only perform manual labor but also execute knowledge-based tasks like "software testing" and "data entry." This means Optimus's target market expands from "manufacturing labor replacement" to "white-collar work assistance and replacement."

04

SpaceX Merges with xAI and Acquires Cursor: Narrative Sprint Before IPO

On February 2, 2026, SpaceX published a statement signed by Musk on its official website, formally confirming its merger with xAI. The statement said the merged company integrates artificial intelligence, rocket technology, satellite internet, and communication services to create a "comprehensive innovation engine." The new company is valued at approximately $1.25 trillion, becoming the world's highest-valued unlisted company.

On April 22, SpaceX struck again, announcing an agreement with AI programming unicorn Cursor: SpaceX secured an option to acquire Cursor for $60 billion or choose to pay $10 billion to advance deep collaboration. Cursor is one of the world's fastest-growing AI programming tools, seen as Microsoft GitHub Copilot's strongest competitor.

Image source: Internet

The logic behind these two transactions is highly consistent: paving the way for SpaceX's IPO.

SpaceX plans to go public later in 2026, with its valuation estimated by multiple investment banks to exceed $1.5 trillion. During the critical pre-IPO window, Musk needs to make one thing clear to secondary markets: SpaceX is not just a rocket company but a next-generation infrastructure platform with AI as its foundation and space as its scenario.

xAI's Grok large model, Cursor's programming tools, Terafab's computing power—injecting these assets before the IPO transforms SpaceX from a "space transportation company" into an "AI+space infrastructure company," fundamentally elevating its valuation logic.

Image source: Internet

05

The Finale of Empire Integration: Earth AI Closed Loop or Space Data Empire?

When placing all these moves on a single diagram, a complete puzzle begins to emerge.

Musk currently has four main business lines: Tesla (electric vehicles + robots), SpaceX (rockets + satellite internet), xAI (AI large models), and X (social media). Deep technological and capital integration is occurring among them:

Tesla invests in xAI, gaining exclusive rights to use Grok models in vehicle infotainment systems and humanoid robots

SpaceX acquires xAI, integrating AI model capabilities into satellite communications and space data center businesses

SpaceX acquires Cursor, injecting AI programming capabilities into its software ecosystem

Terafab simultaneously serves the chip demands of Tesla, SpaceX, and xAI, achieving computing autonomy

Tesla acquires an AI hardware company, addressing the inference computing power shortage for Optimus and FSD

These four business lines are forming a complete closed loop from energy collection (solar) to power conversion (electric vehicles) to computing power manufacturing (chips) to AI models (large models) to application terminals (robots, cars, satellites).

This raises the question: What is the ultimate endpoint of all this—AI hegemony on Earth or preparation for space expansion?

From a commercial logic perspective, the two are not contradictory and likely both correct.

The Earth AI market is valued in the trillions of dollars. Tesla, SpaceX, and xAI first establish competitive advantages on Earth, gaining sufficient revenue and valuation support; the narrative of space data centers provides even greater imagination space for long-term valuation—Starlink's satellite network can offer low-latency data return channels for space AI computing; if humanity needs to establish communication infrastructure between Earth and Moon in the future, SpaceX's competitive edge will be unparalleled.

Image source: Internet

The pre-IPO "narrative sprint" essentially sells a future story to secondary markets. The core of this story is not "rockets are cheap" but "Musk's AI empire will cover the entire physical world from Earth to Moon in the next decade."

06

A Question Worth Pursuing

Musk's capital integration is accelerating at an increasingly rapid pace, with ever-larger moves. However, behind each acquisition and merger lies a common underlying tension: Can these companies truly generate a 1+1>2 synergistic effect when operating independently, or are they merely propping up their valuations with capital narratives?

SpaceX's acquisition of xAI highlights a stark cultural contrast—SpaceX is ultra-pragmatic and engineering-driven, while xAI emphasizes rapid iteration and product-centricity. Post-merger, cultural clashes and resource allocation battles will pose long-term risks.

Image Source: Internet

Terafab's $20 billion chip factory plan represents an ambitious goal, but chip manufacturing success requires more than just capital—process expertise, yield control, and supply chain management demand over a decade of refinement. TSMC's decades-built competitive moat won't be replicated overnight in Austin.

Image Source: Internet

Optimus Gen3's target of 1 million annual production capacity marks an order-of-magnitude leap from actual 2025 shipments. The manufacturing challenge lies not in production capability but in achieving reasonable costs and stable yields.

Yet these remain "rational doubts."

Musk's core strategy never hinges on "success at every step." He bets on all possible directions simultaneously, letting gains from successful ventures offset losses from failures. SpaceX lost three rockets in its early days; Tesla's Model 3 production ramp nearly collapsed the company; xAI's Grok faced initial criticism for being "too conservative." But each failure was followed by faster iteration and bolder bets.

This $2 billion AI hardware acquisition continues that logic—trading capital for time and scale for certainty.

Whether this empire becomes the "Microsoft of the AI era" or collapses like early-stage SpaceX in its next explosion—the answer may only become half-clear when SpaceX goes public.

Core sources: Tesla Q1 2026 10-Q filing (April 23, 2026, submitted to SEC); SpaceX official statements (February 2, 2026); coverage by Wall Street See, Caijing 11, Jiemian News, Sina Finance, and other media on Musk's capital integration efforts.

- END -