On May 7th Eastern Time, Cathie Wood, founder and CIO of ARK Invest, expressed strong optimism in an interview regarding SpaceX's upcoming historic IPO and Tesla's competitive position in the autonomous taxi market. The star investor, known for her steadfast bullishness on disruptive innovation, described SpaceX's orbital data center business as a massive opportunity that could "dwarf" Starlink. She also systematically argued the underlying logic that Tesla's vertical integration will give it a crushing cost-structure advantage over all competitors. Wood's comments come as Tesla's stock has been mired in a slump since the start of the year and as SpaceX's approaching IPO sparks valuation debates. Her sharp stance and pointed arguments have quickly ignited Wall Street discussions about the true value of disruptive innovation.
Regarding the SpaceX IPO, Wood characterized current market demand for its shares as "ravenous." She noted that while SpaceX aims to raise approximately $75 billion through the IPO, boosting its valuation to $1.75 trillion—a figure more than double Saudi Aramco's nearly $30 billion record fundraising in 2019—this amount may still fall far short of meeting the demand from institutional and high-net-worth investors. She predicted "supply-demand imbalances" in the initial IPO phase, leading to significant stock price volatility, but believes long-term investor enthusiasm for SpaceX will withstand short-term fluctuations. Wood emphasized, "Think about how SpaceX has reawakened dreams of space exploration. This has captured the imagination of investors, and everyone."
ARK Invest's venture fund, ARKVX, currently holds assets exceeding $850 million, with SpaceX being its largest holding. As of the end of Q1 2026, SpaceX comprised 17.02% of ARKVX's net assets. Wood explained that because ARKVX is a "public-private hybrid" fund, it is not forced to sell SpaceX shares even after the IPO, but hinted at a potential partial sell-down post-IPO to meet the fund's target allocation of 80% to private companies.
A central, impactful argument from Wood in the interview was positioning orbital data centers as SpaceX's most significant growth catalyst. ARK Invest research suggests the Starlink satellite internet business alone holds revenue potential of up to $160 billion. However, Wood believes the commercial opportunity for orbital data centers will "dwarf Starlink" and potentially increase company revenue by "orders of magnitude—ten or twenty times." She provided a specific figure: if SpaceX successfully deploys orbital data centers, its revenue could leap from the estimated $160 billion for Starlink to a range of $3 to $5 trillion. ARK Invest has published a valuation model for SpaceX on its website. Wood specifically noted that this model does not yet incorporate the revenue potential of orbital data centers, meaning the current valuation framework does not fully reflect the company's most disruptive growth engine. Back in January 2026, Wood stated, "We would not be surprised if SpaceX becomes the first company to reach a $1 trillion market cap, especially with this new opportunity—space data centers." Elon Musk responded to Wood's analysis, leaving a brief, intriguing comment "Interesting" under the published interview video.
While Wall Street eagerly pursues SpaceX, its valuation faces unprecedented scrutiny. Jay Ritter, a University of Florida scholar known as "Mr. IPO," publicly warned that market optimism about SpaceX is "overly naive," questioning whether Starlink can maintain high profit margins while reducing prices, even as launch costs fall. He cited data showing that new issues with price-to-sales ratios exceeding 40x often significantly underperform the market within three years of listing. Ritter stated bluntly, "Even great companies aren't necessarily good investments. If the post-IPO valuation truly reaches $2 trillion, I would consider shorting it." More notably, the American Federation of Teachers has formally written to the SEC requesting a "special review" of the SpaceX IPO. The union, with 1.8 million members and managing large pension funds, pointed out that under current rules, large companies can be added to major indices like the S&P or Nasdaq via a "fast entry" process within just 15 trading days. This forces passive index-tracking funds—including countless American retirement accounts—to automatically buy the highly-valued stock soon after its listing. Union President Randi Weingarten stated in a public letter that retail investors, including teachers, nurses, and public servants, are being directed towards a company with ambiguous financial disclosures, and rules might "force Americans into an overvalued, extremely risky gamble just days after its listing."
According to listing documents disclosed on May 6th, SpaceX will implement a dual-class share structure (A/B shares) granting Musk near-absolute control. By holding Class B shares with 10 votes per share and owning just 42.5% of the equity, Musk will control 83.8% of the voting rights. More startling for governance experts, the charter stipulates that Musk's removal "can only be effected by the vote of Class B shareholders"—meaning the only person who can fire Elon Musk is Musk himself. Harvard Law School corporate governance authority Professor Lucian Bebchuk commented that this clause is "highly unusual" and sets a precedent that other founder-controlled tech companies might emulate. Furthermore, SpaceX is incorporated in Texas and mandates that all shareholder disputes be resolved through arbitration, eliminating investors' rights to a jury trial or class-action lawsuits.
In analyzing the autonomous taxi market competition, Wood systematically argued, using a precise cost-structure framework, that Tesla will ultimately prevail. She established a core analytical framework: while current ride-sharing services like Uber cost about $3 per mile, ARK's analysis suggests Robotaxi costs could fall to 25 cents per mile with the scaling of autonomous technology, compressing transportation costs by over 90% and unleashing new demand. In this cost competition, the degree of vertical integration determines the final profit structure. In Wood's calculations, by 2030, Waymo's cost structure will be at least 50% higher than Tesla's. "Because they rely on other automakers and parts of the supply chain that Tesla does not," she explained. This conclusion is based on key differences: Tesla's Robotaxi solution is built on its own platform, like the Cybercab and existing models like the Model Y. Its integrated, full-stack in-house model for hardware and software allows it to scale its fleet at near-zero marginal cost. In contrast, while Waymo's technology is mature, each operational vehicle must be purchased from third-party automakers and retrofitted with sensors and computing systems, making its cost structure inherently dependent on the supply chain.
Recently, ARK's model significantly raised Tesla's 2026 price target to $4,600 per share, primarily driven by Robotaxi expectations—estimating that by 2026, the autonomous taxi business could contribute about 60% of Tesla's expected value and over half of its EBITDA. If Tesla successfully transitions to a high-margin tech platform driven by software and autonomous services, Wood believes its gross margins could reach 70% to 80%, levels unimaginable for any traditional automaker. Wood's latest comments coincide with Tesla's full push towards physical AI. On April 24th, Musk announced the start of mass production for the Cybercab, a vehicle seen as a key piece of Tesla's "Master Plan Part 3" for physical AI, moving from concept to production ramp-up. In late April, Tesla expanded its Robotaxi service from initial test areas in Austin and the San Francisco Bay Area to Dallas and Houston, directly competing in markets where Waymo already operates. However, initial production volumes remain cautious—Musk clearly indicated that output for new supply chain products like the Cybercab and Semi will be "extremely slow" initially, and the goal of covering about a dozen states by year-end still needs to overcome hurdles like regulatory approvals and safety data validation.
On another front, while SpaceX's orbital data center is not a direct functional module for Tesla's autonomy, it serves as a critical computing foundation and data engine supporting the latter's scaled commercial闭环. This hidden but vital connection is becoming a core link in the logic chain supporting the "trillion-dollar valuation" thesis for Musk's business empire.
Autonomous driving's demand for computing power is growing exponentially, while ground-based computing infrastructure approaches physical limits. Training Tesla's FSD neural network requires massive GPU clusters—currently, Tesla's AI4 chips are fabbed by Samsung, and xAI is a major customer of NVIDIA. However, ground-based data centers face three rigid constraints: GPU shortages, expensive electricity, and soaring cooling costs. The disruptive potential of orbital data centers lies in fundamentally overcoming these three constraints: Unlimited Energy: Sun-synchronous orbits in space provide 24/7 solar power, free from terrestrial grid and land limitations, with a Power Usage Effectiveness nearing 1.0 (far better than the 1.4–1.8 typical for ground data centers); Zero-Cost Cooling: The vacuum environment enables passive cooling, allowing chips to operate safely at temperatures far exceeding ground limits; Layered Computing: SpaceX's planned "orbital data center system," comprising up to 1 million satellites, would form a three-tier computing network for edge computing (500km orbit), large model training (1000km orbit), and global scheduling (2000km orbit).
In April 2026, Tesla unveiled the D3 chip at its Terafab announcement—a chip customized for the space vacuum environment with strong radiation resistance, not reliant on the fragile ground power grid, intended for integration into hundred-kilowatt orbital server racks. Musk's judgment at the announcement was clear: "Within just a few years, the cost of launching chips into space will be lower than building traditional ground-based data centers." This means orbital data centers could not only provide computing power for model training but also directly support real-time inference needs for future autonomous vehicle fleets—satellites in orbit could perform data cleaning and intelligence extraction, avoiding the discard of 90% of raw data due to bandwidth limitations.
The key lever turning this blueprint into reality is the "Terafab" chip manufacturing joint venture officially announced by Musk in March 2026. According to the latest documents SpaceX submitted to the Texas government on May 6th, the project's initial investment is approximately $55 billion, potentially rising to $119 billion if all phases are completed. The facility will use Intel's 14A process technology to produce chips specifically for Tesla's autonomous driving system, the Optimus humanoid robot, and SpaceX's space data centers. Terafab's long-term goal is an annual computing power output of 1 terawatt, equivalent to roughly 50 times the current total annual global output of AI chips. Musk stated directly that 80% of this capacity will be allocated to space-based orbital AI infrastructure, with only 20% for terrestrial applications—a distribution ratio that strongly underscores the thesis that the "center of gravity for computing power is shifting to space."
In summary, for SpaceX, Wood's investment logic rests on three progressively layered pillars: Starlink provides the commercial foundation with $160 billion revenue potential; Orbital data centers open the long-term option for a trillion-dollar "space-based AI compute" opportunity; and the inaugural flight of Starship V3 is the crucial technical test for validating the entire strategy. However, the $75 billion fundraising size, the $1.75 trillion valuation, and the governance structure offering near-zero constraints on the founder are forcing investors to choose between an "unprecedented opportunity" and an "unprecedented ceding of control"—a conflict pushed onto the SEC's review table by challenges from unions and economists.
For Tesla, Wood's arguments are equally fundamental: Robotaxis will fundamentally alter the economics of transportation—the gap between 25 cents per mile and $3 per mile is enough to create entirely new markets within highly price-elastic transportation demand. In this market, the vertically integrated player will enjoy a structural cost advantage. However, Cybercab production is just beginning, the goal of covering about a dozen states with Robotaxi service by year-end still faces regulatory and safety validations, and Waymo's intense expansion, with hundreds of thousands of weekly paid rides, presents real competitive pressure.
History will ultimately judge these bets. But for now, this star investor, who called for a $4,600 Tesla price target back in early 2026, is declaring with renewed clarity that she sees not a bubble, but two super-sized opportunities to "reshape human civilization" that remain undervalued by the market.
Comments