Rob Spivey has spent years tracking every major move by Elon Musk. His track record of three consecutive correct predictions about Musk’s business pivots has drawn attention from institutional investors and individual traders alike. Today Spivey argues that the real story behind SpaceX’s public listing is not the rockets themselves but a rapidly expanding “AI landlord” business that could reshape how the world’s largest technology companies access computing power. In his analysis, the companies best positioned to benefit are not SpaceX itself but a select group of specialized suppliers feeding Elon’s data-center and orbital-compute build-out. This article lays out the full picture Spivey presents, the logic behind his recommendations, and why he believes access to his firm’s research service, Hidden Alpha, gives ordinary investors a practical way to position themselves for the opportunity.
Rob Spivey’s Track Record of Elon Predictions
Spivey co-founded a financial research firm that later became known as Altimetry. Over more than fifteen years the firm’s work has been followed by professionals at major money managers including Goldman Sachs, JPMorgan Chase, BlackRock, and Fidelity. Spivey’s own focus in recent years has been mapping the flow of capital that follows Musk’s strategic decisions.
In January 2025, before the official launch of the Department of Government Efficiency, Spivey predicted that Musk’s involvement would center less on pure budget cutting and more on opening federal data vaults for AI training. Subsequent reporting from Wired, Politico, and the Washington Post described DOGE’s emphasis on data access. Spivey’s firm recommended two stocks tied to the theme; both rose more than 130 percent in the following months.
In April 2025 Spivey correctly anticipated Musk’s step-back from day-to-day DOGE operations. In September 2025 he forecast that Tesla’s November shareholder meeting would mark a decisive pivot toward Optimus humanoid robots, with Musk stating that robots would eventually represent the majority of Tesla’s future value. One robotics-related name the firm highlighted advanced nearly 100 percent.
Spivey frames these calls not as luck but as pattern recognition: when Musk commits capital and engineering resources at scale, secondary industries and suppliers experience rapid demand spikes. He applies the same framework to the SpaceX story.
What Wall Street Missed in the SpaceX Filing
SpaceX went public at a valuation near $1.8 trillion. Most coverage focused on Starship, Starlink, and launch cadence. Spivey contends that the more consequential disclosure lay in the commercial arrangements for computing capacity at the Memphis-area Colossus complex.
Colossus began life as an abandoned Electrolux factory. In roughly four months the site was converted into an operational AI training cluster. Colossus 1 housed approximately 200,000 high-end chips. Colossus 2 expanded the total to roughly 555,000. A third facility across the Mississippi border is expected to add another 220,000 to 300,000 chips, bringing the campus toward one million interconnected processors. Power comes from on-site natural-gas turbines, more than 160 Tesla Megapack batteries, a relocated overseas power plant, and an $80 million water-recycling system handling 13 million gallons per day.
Spivey estimates that across Colossus, Tesla’s Austin and Buffalo facilities, and X’s Atlanta data center, Musk-controlled entities operate between 750,000 and 800,000 of the most powerful AI chips currently deployed. That concentration, he argues, places a single decision-maker in a position to rent capacity to the largest AI developers at a time when traditional hyperscalers face multi-year grid interconnection queues, permitting delays, and chip shortages.
The AI Capacity Crisis
Three structural bottlenecks constrain conventional data-center expansion. First, nearly half of planned AI data centers scheduled for the current year face delay or cancellation; in the first quarter of 2026 alone, 75 projects totaling $130 billion were blocked or postponed. Goldman Sachs has estimated that only 50 to 60 percent of announced capacity will arrive on schedule. Second, utility interconnection queues now average nearly five years and can stretch to twelve years when major transmission upgrades are required; roughly 80 percent of sponsors ultimately withdraw. Third, high-end Nvidia H100 rental prices have risen even as newer chips appear, reflecting persistent scarcity that is not expected to ease before late 2027.
Companies that need training capacity today therefore face a binary choice: wait years for their own infrastructure or rent from whoever already possesses operational clusters. Spivey identifies Musk’s constellation of facilities as the largest readily available pool.
Major Capacity Deals Disclosed
According to Spivey’s reading of the SpaceX S-1 and subsequent announcements, four significant customers illustrate the emerging landlord model.
Anthropic, creator of the Claude models, is described as renting the full capacity of Colossus 1—approximately 300 megawatts and 220,000 chips—for $1.25 billion per month, or $15 billion annually through May 2029, totaling $45 billion. Cursor, an AI coding-tool company, signed a capacity agreement and was later acquired, converting external rental revenue into internal use. Reflection AI, an open-source model developer valued at $25 billion, agreed to pay $150 million per month for Colossus 2 access, equating to roughly $6.3 billion through 2029. Google is reported to be paying $920 million per month, or $11 billion per year, for a total of $30 billion through June 2029. Google has also been a SpaceX shareholder since 2015 and holds a board seat, creating a dual shareholder-customer relationship that Spivey views as mutually reinforcing.
Taken together, these contracts represent tens of billions in contracted revenue flowing into the same infrastructure footprint. Spivey characterizes the arrangement as the beginning of a near-monopoly on rapid-deployment AI compute.
Orbital AI and the Next Expansion Layer
Earth-based clusters remain constrained by power, water, and local opposition. Spivey highlights a parallel initiative: orbital data centers. Google’s Project Suncatcher, announced in 2024, aims for prototype launches around 2027. Because Google does not operate a launch fleet, SpaceX becomes the logical partner. Musk has publicly discussed the advantages of space-based compute: continuous solar power, passive radiative cooling in the vacuum of space, and freedom from terrestrial permitting.
To manufacture the large solar arrays, electronics, and satellite structures required, SpaceX is constructing Gigasat in Bastrop, Texas—an eleven-million-square-foot facility more than ten times larger than existing spacecraft plants at Starbase. The plant is designed for vertical integration of solar panels, circuit boards, terminals, and finished spacecraft. Spivey notes that first orbital AI prototypes are targeted for launch in the near term, with Gigasat ultimately sized for thousands of satellites annually.
Complementing the satellite factory is Terafab, a joint chip-manufacturing effort involving Tesla, SpaceX, xAI, and Intel. A pilot line in Austin is scheduled for late-2026 production, followed by a full-scale plant in Grimes County, Texas, intended to reach one terawatt of annual AI compute capacity. Spivey presents these projects as Musk’s attempt to internalize the entire supply chain from silicon to orbit.
Despite the scale of the opportunity, Spivey does not recommend purchasing SpaceX equity at current valuations. He cites several factors: the multi-trillion-dollar market capitalization already prices in substantial success; ongoing AI-related capital expenditures weigh on near-term profitability even while Starlink is profitable; a 180-day post-IPO lock-up expiration in December could introduce selling pressure from insiders and early employees; and Musk’s historical timelines for complex hardware programs frequently slip. If orbital data centers encounter delays measured in years rather than months, the share-price reaction may lag the underlying infrastructure story.
Instead Spivey points investors toward the specialized suppliers that must deliver hardware, cooling, power equipment, and related systems into the build-out. Historical precedent, he argues, favors this approach.
Historical Supply-Chain Multipliers
When Apple shifted to high-end OLED displays, Universal Display Corporation, holder of critical patents, rose from roughly $50 to more than $200. When Face ID required miniaturized laser arrays, Lumentum Holdings climbed from the $30 range past $100. Luxshare Precision Industry, once a cable supplier, captured AirPods manufacturing and advanced from $10 to more than $60. In the first wave of generative AI, Vertiv Holdings, a data-center cooling specialist, moved from a 2023 low near $12 to over $189. SK Hynix, dominant in high-bandwidth memory for Nvidia GPUs, gained more than 350 percent over three years. Super Micro Computer, provider of modular liquid-cooled servers, rose from under $8 to an adjusted peak above $114.
Spivey contends that the same dynamic is forming around Musk’s AI infrastructure program. Capital expenditures by Amazon, Microsoft, Alphabet, Meta, and Oracle are projected to reach $700 billion in 2026 alone. Global data-center infrastructure spending is expected to approach $1 trillion by 2030. The suppliers that solve the bottlenecks—servers, cooling, power, specialized semiconductors, and satellite components—stand to capture outsized returns if the build-out proceeds at the pace Musk has demonstrated with Colossus.
The Altimeter System and Institutional-Grade Research
Altimetry’s core analytical tool is the Altimeter, a proprietary ranking system developed by Spivey and his partner Joel Litman.
Back-tests over nearly three decades show that the top thirty companies identified by the system substantially outperformed the broader market.
The same framework flagged Meta Platforms in 2013 before its multi-thousand-percent advance, Moderna before its pandemic-era surge, and Advanced Micro Devices in 2015 when the firm was featured in Barron’s; AMD subsequently rose more than 4,000 percent. A semiconductor name still held in the firm’s open portfolio has delivered a gain exceeding 1,000 percent.
Spivey emphasizes that the system is designed to surface companies whose economic reality differs from reported accounting metrics. In the case of Dell Technologies, for example, as-reported return on assets appears modest at approximately 6 percent, yet the Altimeter calculation yields roughly 41 percent—nearly seven times higher—because it adjusts for items that obscure underlying profitability. Spivey expects those returns to expand further as demand for AI server infrastructure continues.
Free Research Bundle and the Hidden Alpha Subscription
To help investors act on the SpaceX AI landlord thesis, Spivey offers a package of research materials.
The first report, titled The SpaceX AI Landlord Supply Chain: The Stocks Positioned to Capture Elon’s $50 Billion Buildout, provides names, ticker symbols, buy parameters, and detailed analysis of the suppliers Spivey considers most leveraged to the Memphis expansion, Gigasat, and related projects.
The second report, Dark Energy: Stocks That Could Soar as AI Goes Off the Grid, examines companies positioned for alternative power solutions that bypass traditional utility constraints.
A third report lists ten widely held stocks Spivey believes investors should avoid because capital is flowing away from their business models.
Subscribers also receive twelve months of access to the Altimeter itself, allowing real-time evaluation of hundreds of stocks, and the monthly Hidden Alpha research service. Hidden Alpha delivers new investment ideas, ongoing updates on existing recommendations, and a model portfolio. Over the past two years the average gain on recommendations has been reported at approximately 50 percent, more than double the market’s return over the same period. The service normally carries an annual fee of $499 or higher; the current offer is a discounted trial with a full 30-day money-back guarantee.
Spivey stresses that all investing involves risk and that past performance does not guarantee future results. Testimonials from subscribers describe significant gains, yet the firm notes that such outcomes are not typical and that capital can be lost.
Dell Technologies as the Featured Free Recommendation
Within the broader supply-chain thesis, Spivey singles out Dell Technologies (NYSE: DELL) as the top free recommendation. Dell has supplied server infrastructure for the original Colossus installation and continues to work with xAI and SpaceX on AI Factory solutions. Spivey argues that the company’s true economic returns are substantially higher than as-reported figures suggest and that rising demand for AI servers should further expand those returns. He presents Dell as one concrete way for investors to gain exposure while awaiting the full set of supplier recommendations inside the research package.
The Broader Customer Pipeline
Beyond Anthropic, Cursor, Reflection AI, and Google, Spivey anticipates additional demand from Meta’s Llama development, OpenAI’s next-generation models, U.S. defense agencies seeking secure compute for classified workloads, and sovereign wealth funds in the Middle East racing to build national AI capacity.
High-frequency trading firms may eventually pay premiums for orbital latency advantages, while intelligence agencies could process imagery and signals in space before data reaches the ground. In each case, Spivey argues, the party that already controls large operational clusters and the means to expand them into orbit holds pricing power.
Positioning Before the Next Capital Wave
Spivey’s central claim is that Musk has decoupled a meaningful portion of AI compute from the terrestrial grid and is now building the manufacturing capacity to extend that advantage into orbit. The capital already committed by the largest technology companies, combined with government and international interest, creates a multi-year spending cycle. Investors who buy the over-hyped parent company at elevated valuations may face lock-up selling and timeline slippage. Investors who identify the bottleneck suppliers early, Spivey contends, stand to capture the same asymmetric returns seen in prior technology transitions.
Hidden Alpha and the accompanying research reports are presented as the practical mechanism for obtaining the specific names, entry criteria, and ongoing updates. The 30-day guarantee is intended to remove friction for readers who wish to evaluate the material themselves. Spivey closes with a straightforward exhortation: the infrastructure cycle is already underway, the customer pipeline is forming, and the window for positioning before the next wave of announcements is finite.
Readers interested in the full set of supplier recommendations, the Dark Energy power-theme analysis, the list of stocks to avoid, twelve months of Altimeter access, and the Hidden Alpha monthly service can review the current subscription offer and begin a risk-free trial. The research is delivered electronically shortly after enrollment, allowing immediate review of the SpaceX AI landlord supply-chain thesis and the concrete stock ideas Spivey believes are best positioned to benefit.
The story Spivey tells is ultimately one of capital following engineering. Musk has repeatedly demonstrated the ability to compress timelines that conventional operators treat as fixed. Colossus rose in months rather than years. Gigasat and Terafab aim to repeat the pattern at larger scale.
If the pattern continues, the specialized companies that supply the racks, cooling systems, power equipment, chips, and satellite components will experience demand that few other sectors can match. Rob Spivey’s research is designed to map that demand onto specific, actionable investment ideas and to keep subscribers informed as the picture evolves. For investors seeking a structured way to participate in the AI landlord supply chain without relying solely on the parent company, the Hidden Alpha package offers a direct path to the analysis.
