SpaceX Q2: $7.8B Revenue Clashes With $15.8B AI Spend

Woofun AI reports that SpaceX’s first quarterly earnings reveal a stark financial dichotomy, where soaring revenues collide with unprecedented capital expenditure, leaving Wall Street Insights and analysts like Yanwai Yizhi questioning the sustainability of Elon Musk’s triple-threat strategy. The market’s reaction to this contradictory picture was immediate and negative, driven by concerns over liquidity and the sheer scale of investment required to maintain momentum across three distinct business verticals. CFO Bret Johnsen and Musk spent the earnings call attempting to reconcile these divergent narratives, but the disconnect between cash burn and realized profit remains the central tension for investors.

The second quarter financials present a complex landscape of growth and expenditure. Revenue reached $7.8 billion, marking a 92% year-on-year increase from $4.1 billion in the same period last year. Adjusted EBITDA climbed to $3.5 billion, significantly outpacing the $1.2 billion recorded in the prior year.

However, capital expenditures surged to $18.4 billion, which is 2.4 times the total revenue. Of this massive outlay, $15.8 billion was allocated to AI-related initiatives, translating to an average daily spend exceeding $170 million. This aggressive spending pattern caused the stock price to drop more than 7% in after-hours trading, as investors struggled to value a company where capex dwarfs income.

Market anxiety is further compounded by an impending liquidity event that management addressed only perfunctorily. On August 6, 910 million shares of internal stock held by SpaceX will become available for sale, a volume that accounts for 1.4 times the current circulating supply. The failure to proactively discuss this dilution risk during the main presentation, responding only briefly during the Q&A session, has heightened investor uncertainty. This structural overhang casts a shadow over the reported profitability, as the potential selling pressure could significantly impact valuation metrics regardless of operational performance.

Starlink continues to demonstrate robust profitability, generating $4.29 billion in revenue, a 66% year-on-year increase. Operating profit reached $1.66 billion, up 79% year-on-year, resulting in a healthy profit margin of 38.6%. The user base expanded by 1.7 million, bringing the total to 12 million users across 167 countries. Enterprise and government clients contributed $1.8 billion in revenue, more than doubling from the previous year. The U.S. government awarded contracts worth over $6 billion in the second quarter, while American Airlines signed an agreement for in-flight Wi-Fi services. CEO Gwynne Shotwell stated that Starlink aims to become the fourth-largest telecom operator in the U.S., trailing AT&T, Verizon, and T-Mobile, whose combined annual revenue is approximately $600 billion.

Operational metrics for Starlink show stability in average revenue per user (ARPU), which remained at $66, down from $85 in the second quarter of last year but consistent with the first quarter of this year. The fastest-growing user bases are in Latin America, Africa, and Southeast Asia, where lower pricing drags down the average compared to North America.

However, high-value contracts from aviation, maritime, and government sectors offset this dilution. Currently, over 10,200 satellites are in orbit, with each V3 broadband satellite offering a downstream capacity of 1 Tbps, which is 10 times that of V2 satellites. Shotwell noted that V3 satellites will not reach full deployment until mid-next year. If annualized, Starlink’s second-quarter figures suggest $17.2 billion in revenue and $6.6 billion in operating profit, a scale still far below T-Mobile’s $81 billion revenue and $13 billion profit, but growing at a much faster rate.

The AI business segment experienced a dramatic surge in activity, generating $2.56 billion in revenue, a 247% year-on-year increase and a 213% quarter-on-quarter jump. Cloud service agreements contributed around $1.6 billion, while revenue from AI solutions skyrocketed from $475 million in the first quarter to $2.194 billion. Adjusted EBITDA turned positive for the first time, rising from -$609 million in the first quarter to $1.146 billion. Despite this improvement, operating losses still amounted to $1.257 billion when depreciation and amortization were excluded. This mixed profitability picture highlights the transitional nature of the AI division, which is rapidly scaling revenue but has not yet fully converted that growth into bottom-line operating income.

Woofun AI data shows that the capital intensity of the AI division is the primary driver of market concern, with $15.8 billion spent on infrastructure, surpassing NVIDIA’s entire annual R&D budget of $12.9 billion last year. This quarterly spend is nearly 60% of Meta Platforms’ total annual capital expenditures of $27.2 billion. Most of this investment was directed toward the Colossus data center in Memphis, Tennessee. The nominal hash rate capacity at the end of the second quarter was 1.4 gigawatts, with a target of reaching 2 gigawatts by the end of the year. Musk stated that the internal goal is to achieve 20 gigawatts in power and cooling by the end of 2027, though he personally expects around 15 gigawatts. This represents a roughly 10-fold increase from the current 1.4 gigawatts, signaling an aggressive expansion in computational infrastructure.

Regarding the return on investment, CFO Johnsen claimed the payback period for new hash rate capacity is less than one year, suggesting that capital expenditures can be treated as operating expenses given current contract scales. In the first three weeks of Q3, the company signed additional cloud contracts worth $6.7 billion, with a service period of about half a year starting from October.

However, analysts noted that $7.7 billion was spent last quarter, bringing the two-quarter total to $23.5 billion, and questioned when corresponding returns would appear on the income statement. Musk projected that by December, annualized monthly revenue would reach $100 billion, and moved the internal target of $1 trillion in revenue from 2031 to 2030, with a "non-zero probability" of achieving it by 2029. The market remains skeptical of these long-term projections in the face of immediate cash burn.

The launch services segment generated $962 million in revenue, a 29% year-on-year increase, but incurred operating losses of $542 million. Thirty-eight launches were completed in the second quarter, with Starship being the primary source of losses. In the past 90 days, Starship V3 completed two successful flights: the 13th flight tested orbital mission capabilities and tower capture, while the 14th flight successfully deployed a Starlink V3 satellite into operational orbit. Musk stated that the heat shield issue, previously described as the biggest technical obstacle, has been resolved.

The next test flight, scheduled for the end of this month, will attempt to capture and recover both the first-stage booster and the spacecraft. R&D expenses increased by $389 million year-on-year, resulting in an adjusted EBITDA of -$200 million for the aerospace segment. Musk’s goal is to achieve at least one Starship launch per day within a year and reduce orbital launch costs to less than 1% of traditional methods, claiming that competitors’ launch capacity would be just one pixel on a global chart.

The conclusion drawn from this quarterly report is that Musk is attempting to balance three vastly different business pressures: rapid profitability from Starlink, aggressive expansion in AI, and long-term R&D for Starship. With control over 82% of voting rights, he is driving a strategy that includes ongoing compute rental agreements with Anthropic worth $1.25 billion per month through 2029 and the release of Grok 4.5, which tripled token consumption. While Starlink offers a clear path to becoming a major telecom player, the AI division’s $15.8 billion quarterly spend and Starship’s persistent losses create a volatile financial profile. As Johnsen noted, the gap between accounting treatment and investor perception remains wide, and the market awaits a definitive answer on whether a single entity can successfully execute all three strategies simultaneously.