SpaceX bets big on Nvidia, burn rate worries grow

SpaceX is executing an unprecedented capital expenditure program, with 2025 spending reaching $20.7 billion and Q1 2026 alone hitting $10.1 billion, primarily funneled into AI infrastructure, Starship development, and satellite manufacturing. The lion's share—over $15 billion in Q2 2026—is dedicated to building out the Colossus AI supercomputer, the largest coherent AI training facility on Earth, with plans to deploy orbital versions powered by solar energy. This aggressive CapEx strategy underpins SpaceX’s ambition to scale compute capacity from 0.4 gigawatts a year ago to an expected 2 gigawatts by the end of 2026 and potentially 10 gigawatts by late 2027, positioning the company as a formidable player among hyperscalers.

Despite the rapid expansion and soaring AI-related revenues—up 247% year-over-year to $2.56 billion in Q2 2026—SpaceX’s AI compute cloud remains unprofitable, with operating losses of $1.3 billion in the quarter driven largely by fast depreciation of GPUs. This capital intensity has pushed the company’s debt and finance leases from $22 billion to nearly $37 billion within a quarter, while cash reserves have sharply declined, reflecting a high cash burn rate that challenges free cash flow sustainability. However, management anticipates margin improvements in the AI cloud rental business, targeting low to mid-teens margins, albeit still below traditional hyperscaler levels.

SpaceX’s AI infrastructure buildout is tightly integrated with an exclusive partnership with Nvidia, leveraging the Vera Rubin architecture for superior compute performance and collaborating on AI satellite payloads such as the Starmind AI1 satellite. This strategic alignment not only ensures cutting-edge hardware for their massive compute clusters but also likely facilitates vendor financing to support projected 2027 capital expenditures between $300-500 billion. Elon Musk’s commitment to Nvidia chips underscores SpaceX’s ambition to become a sixth major hyperscaler, joining industry giants like Alphabet and Microsoft in the AI infrastructure race.

While SpaceX’s core businesses like Starlink and space launches remain relatively small compared to its AI ambitions, the company is targeting a $100 billion annual recurring revenue run rate by the end of 2026 and a staggering $1 trillion revenue milestone by 2030, with AI infrastructure and cloud services expected to be the primary growth engines. This aggressive expansion also includes ventures into mobile markets via Starlink Mobile, aiming to capture a slice of the $600 billion annual revenue generated by top US carriers, highlighting a diversified approach to monetizing its vast compute and satellite assets.

SpaceX's exclusive partnership with Nvidia centers on the adoption of Nvidia's Vera Rubin GPU architecture for all its AI workloads, a decision driven by the chips' superior performance and efficiency. Elon Musk emphasized this choice, stating, 'We've decided to build exclusively on Nvidia because we think the Vera Rubin architecture is the best architecture,' highlighting the strategic importance of this hardware in powering SpaceX's AI ambitions. The Vera Rubin superchip delivers roughly ten times more performance per watt compared to its predecessors, combining Vera CPUs and Rubin GPUs with NVLink interconnects, which cements Nvidia's technological edge in AI compute infrastructure.

Underpinned by this exclusive Nvidia partnership, SpaceX is aggressively scaling its AI compute capacity, aiming to triple its footprint by 2027 with over two gigawatts of compute expected online by the end of 2026 and scaling to ten gigawatts by the end of 2027. This rapid expansion positions SpaceX as a potential sixth hyperscaler in the AI infrastructure space, with Morgan Stanley estimating that hyperscaler AI investments could reach $110 billion by 2028. Beyond hardware supply, the collaboration extends to co-developing compute payloads for projects like the Starmind AI1 satellite, showcasing a deep integration of Nvidia's technology into SpaceX's AI ecosystem.

SpaceX's strategy leverages its Nvidia-powered AI compute infrastructure not only for internal model development, such as the Grok AI system, but also by leasing excess GPU capacity to external customers, effectively transforming itself into a 'neocloud' provider. This approach monetizes otherwise idle silicon, turning expensive capital expenditures into revenue streams while maintaining flexibility in AI model development. However, this compute leasing model carries inherent risks typical of capital-intensive infrastructure businesses, including hardware depreciation and potential downward pressure on rental prices as more data centers come online and AI models become more efficient.

The exclusive Nvidia partnership has had a pronounced positive impact on Nvidia's market valuation and earnings outlook, with Nvidia's stock rising following the announcement and analysts projecting a 45% annual earnings increase over the next three years. Nvidia's current PEG ratio of 0.75, the lowest in five years, alongside a median target price implying a 33% upside, reflects Wall Street's confidence in underestimated hyperscaler AI infrastructure spending, much of which is fueled by SpaceX's ambitious compute expansion plans. This symbiotic relationship underscores how SpaceX's AI ambitions are tightly coupled with Nvidia's growth trajectory in the AI hardware market.

By early 2026, SpaceX is riding a wave of record revenue growth fueled predominantly by its Starlink connectivity segment and burgeoning AI infrastructure business. Starlink alone accounts for 61% of total revenue, generating $11.4 billion with an impressive 50% year-over-year growth driven by both consumer broadband and enterprise/government contracts, while the AI segment contributes $3.2 billion, growing 22% annually through subscriptions and infrastructure sales despite advertising declines. This dual-engine growth propelled SpaceX’s total revenue run rate to approximately $19 billion, marking a 33% increase year-over-year and underscoring the company’s successful diversification beyond its traditional aerospace roots.

SpaceX’s profitability is anchored by Starlink’s robust adjusted EBITDA of $7.2 billion at 63% margins, which not only cements the connectivity business as the company’s financial backbone but also fuels ambitious future growth targets. With a substantial backlog of $28.4 billion in multi-year launch contracts and enterprise Starlink deals, alongside a $47.5 billion AI segment backlog driven by marquee clients like Anthropic and Google, SpaceX has a clear line of sight on accelerating top-line momentum and margin expansion in the near term. This financial foundation supports the company’s bold projections to reach a $100 billion annual recurring revenue (ARR) by the end of 2026, reflecting confidence in sustained demand for its satellite and AI compute services.

Looking further ahead, Elon Musk’s vision for SpaceX is nothing short of transformative, with internal forecasts now targeting $1 trillion in total revenue by 2030—an ambitious goal that has been accelerated from earlier estimates of 2031 and carries a 'non-zero chance' of being achieved as early as 2029. This aggressive outlook is underpinned by rapid expansion in AI compute capacity, which has surged from 0.4 to 1.4 gigawatts within a year, and the deployment of next-generation V3 satellites that Musk describes as 'an order of magnitude more capable.' Meanwhile, Starlink aims to dominate global internet connectivity, with plans to challenge major U.S. mobile carriers in the $600 billion market through its Starlink Mobile initiative, signaling a strategic push to capture significant new revenue streams.

The evolving revenue mix reveals a strategic pivot as SpaceX’s AI compute segment now represents 77% of its run rate revenue, effectively recasting the company as a leading AI infrastructure powerhouse with a rapidly growing customer base that includes Microsoft and Anthropic. This shift from a primarily aerospace-focused firm to a compute-centric enterprise is critical to meeting its staggering $1 trillion ARR target by 2030, with intermediate milestones such as $500 billion ARR expected in the nearer term. Despite investor concerns over heavy capital expenditures, these aggressive investments in AI infrastructure and satellite technology underpin SpaceX’s confidence in sustaining exponential revenue growth and reshaping its market identity.

Investors are increasingly uneasy about SpaceX's aggressive capital expenditures, which have led to significant negative free cash flow despite robust revenue growth. For instance, the company’s FY 2025 CAPEX of $20.7 billion dwarfed its operating cash flow of $6.8 billion, creating a $14 billion funding gap primarily covered by debt and preferred equity raises. The AI infrastructure segment, while fueling growth, remains a drag on profitability, with losses of $1.2 billion in adjusted EBITDA in 2025 and operational losses of $1.9 billion in Q1 2026 due to rapid GPU depreciation, highlighting the challenges of balancing rapid expansion with financial sustainability.

SpaceX’s stock price has experienced pronounced volatility, tumbling over 56% since its IPO and dropping more than 19% after recent earnings, reflecting investor skepticism about the company’s spending spree and uncertain returns. This volatility is exacerbated by a surge in the tradable float post-IPO, with approximately $100 billion in shares becoming transferable and raising concerns about potential sell-offs. Although the anticipated wave of insider selling following the lock-up expiry did not materialize, investors remain wary, especially given the company’s net loss of $541 million and ongoing GAAP losses despite a 92% year-over-year revenue increase.

Skepticism persists among analysts and investors regarding CEO Elon Musk’s ambitious growth targets and the sustainability of SpaceX’s financial model. While Musk aims for a $1 trillion revenue run rate by 2030 and daily Starship launches by next year, experts like Tim Farer deem these goals 'not credible' in the short term, and some label the stock as overvalued by a factor of three. Additionally, the company’s AI cloud rental business, expected to yield low to mid-teens margins, pales in comparison to the mid-30% margins of established hyperscalers, fueling doubts about profitability timelines and the company’s ability to finance its expansion without repeatedly tapping capital markets.

Investor uncertainty is further compounded by SpaceX’s modest financial disclosures and lack of executive commentary during earnings releases, which limits transparency around key segments like Starlink Mobile and AI revenue projections. CFO Bret Johnson’s brief remarks have done little to assuage concerns, especially as the market eagerly awaits clearer guidance on AI revenue streams that are pivotal to future valuation. This opacity, combined with the enormous scale of capital deployment—such as the $18.4 billion spent in Q2 2026, far exceeding expectations—continues to fuel tensions between the company’s operational successes and investor confidence in its long-term financial health.