The expansion of AI computing power opens up room for growth! Bernstein raised SpaceX (SPCX.US) price target to $268, and Damo maintains “gain” rating

The Zhitong Finance App learned that as SpaceX (SPCX.US) accelerates the deployment of terrestrial AI computing power infrastructure and promotes Starship (Starship) reusable technology and orbital data center plans, Wall Street Investment Bank remains optimistic about the company's long-term growth prospects. In a research report released on October 26, Bernstein raised SpaceX's target price from $248 to $268, maintaining the “outperforming market” rating. Morgan Stanley, on the other hand, maintained SpaceX's “gain” rating and a target price of $300 in the October 23 report, believing that the market has yet to fully reflect the potential value of the company's transformation from an aerospace company to a comprehensive AI platform.

Bernstein pointed out that the target price increase is mainly based on the speeding up deployment of SpaceX's terrestrial AI computing power and the improvement in pricing prospects for computing power services. Despite the recent decline in the valuation of AI-related stocks, the agency believes that SpaceX's layout in the fields of terrestrial computing power, satellite communications, and future orbital data centers is still expected to drive a significant increase in the company's revenue.

According to Bernstein Research, SpaceX's third-quarter revenue reached 14.58 billion US dollars, higher than the market's consensus forecast of 12.88 billion US dollars; after dilution, earnings per share were 0.35 US dollars, which significantly exceeded the market forecast of 0.15 US dollars. The total operating profit of the company's various business divisions was US$2.79 billion, which was also significantly higher than market expectations of US$1.21 billion.

Among them, the AI business became a highlight of growth. SpaceX's AI business revenue in the third quarter reached 7.56 billion US dollars, and operating profit was 431 million US dollars, all exceeding market expectations. Bernstein pointed out that this is mainly due to the faster deployment of terrestrial computing power facilities than expected and the higher price of computing power services.

By the end of the third quarter, SpaceX had about 1.7 gigawatts of nominal computing power capacity. The company expects this scale to grow to 2.5 to 3 gigawatts by the end of 2026, and plans to reach at least 13 gigawatts by the end of 2027.

Based on this, Bernstein raised his forecast for SpaceX's terrestrial AI computing power construction scale. The installed capacity is expected to reach 2.8 gigawatts, 12.8 gigawatts, and 20.1 gigawatts at the end of 2026, 2027, and 2028, respectively.

At the same time, the price of computing power services has also become an important factor supporting SpaceX's growth prospects. According to the report, SpaceX previously agreed on computing power deals with Google and other customers at about 30 to 50 US dollars per watt, and the company's management said that recent offers obtained during negotiations with potential customers are already higher than this level.

Although Bernstein raised his own computing power pricing assumptions, the forecast was still below the level described by management. The agency predicts that SpaceX's average computing power pricing will reach $26 per watt in 2027, then decline as supply and demand relationships gradually change.

However, aggressive AI infrastructure expansion also means huge capital expenses. SpaceX expects capital expenditure to reach 35 billion to 40 billion US dollars in the fourth quarter, far higher than the market's previous estimate of about 22 billion US dollars.

Bernstein hypothesizes that SpaceX will need to invest about 35 billion US dollars for every additional gigawatt of terrestrial computing capacity. The agency believes that despite high construction costs, if demand for computing power continues to be strong and contract pricing remains high, related investments may still generate considerable returns.

SpaceX management has proposed more aggressive targets for future revenue growth. The company expects the annualized revenue operating rate to reach at least 100 billion US dollars by the end of 2026, and believes that by the end of 2027, this figure may rise to 500 billion to 1 trillion US dollars, mainly driven by the terrestrial AI computing power business.

In contrast, Bernstein's predictions are more conservative. The agency predicts that SpaceX's annual revenue for 2027 and 2028 will reach approximately US$241.9 billion and US$423.7 billion, respectively, and raised its revenue forecast for 2031 to US$784 billion from US$655 billion previously. Among them, the AI business is expected to be the biggest source of revenue. Bernstein predicts that by 2031, SpaceX's AI business revenue is expected to reach about 571.2 billion US dollars, and the adjusted EBITDA is about 476.8 billion US dollars.

Despite raising profit forecasts, Bernstein also lowered the AI business valuation ratio from 21 times to 17 times to reflect the impact of the recent decline in AI-related stock valuations. This also partially offset the positive impact of the growth in the computing power business on the target price.

In addition to the ground-based AI business, Bernstein believes that orbital data centers are still an important factor supporting SpaceX's long-term valuation. SpaceX plans to launch its first orbital data center at the end of 2027 and aims to start operations in mid-2028. Bernstein pointed out that whether orbital data centers can be commercialized depends largely on whether starships can be completely reused, thereby significantly reducing space transportation costs.

According to the company's plan, the 15th Starship test flight will try to recycle both the booster and the main body of the spacecraft. If the test flight is successful, SpaceX is expected to reuse the spacecraft for the first time in the 16th flight.

Bernstein believes that a breakthrough in Starship reuse capabilities will provide an important foundation for orbital data center construction and may further improve the long-term economy of SpaceX's space business.

Meanwhile, SpaceX's Starlink (Starlink) business continues to show strong profitability. In the third quarter, SpaceX connectivity business revenue reached US$5.25 billion, higher than market expectations of US$4.86 billion, and operating profit margin reached 45%, clearly exceeding market expectations of 38%. Among them, the number of broadband subscribers of Starlink consumers increased by 1.6 million month-on-month to 13.6 million, and the average revenue per user rose to 69 US dollars. Although the rate of user growth has slowed, Bernstein believes that Starlink's profitability is improving faster than the number of users expanding.

The corporate and government connectivity business performed particularly well, with related revenue rising 91% year over year in the third quarter. SpaceX management anticipates that in the long run, the revenue contribution of this business may reach or even exceed that of the consumer broadband business.

The aviation internet market is also an important source of growth. Bernstein pointed out that SpaceX has signed cooperation agreements with at least 46 airlines around the world, involving about 12,000 aircraft, which is close to a quarter of the serviceable fleet.

In addition, SpaceX is also actively promoting the direct satellite mobile phone business and exploring the possibility of building terrestrial mobile communication networks. However, Bernstein believes that related businesses still face high capital investment and execution risks, so management's aggressive growth goals have not been fully included in the forecast.

Morgan Stanley, on the other hand, is optimistic about SpaceX from the perspective of longer-term industrial development. The agency believes that as AI investment gradually extends from data centers and software to physical industries such as manufacturing, robotics, energy, and aerospace, the market will usher in a stage of real economic development driven by artificial intelligence.

Morgan Stanley refers to this trend as the “real smart economy,” and believes SpaceX and Tesla (TSLA.US) are expected to play an important role in this process.

The report points out that the market still tends to view SpaceX's AI business as a traditional cloud computing infrastructure business, without fully considering its vertical integration advantages in rocket launches, satellite communications, AI computing power, and other technical fields.

According to Morgan Stanley's valuation model, the total value of SpaceX's aerospace and connectivity business is approximately $123 per share. Based on the approximately $145 share price used in the report, the implied value given to its AI business by the market is only about $22 per share. The agency believes this means SpaceX's AI business is still undervalued.

Morgan Stanley expects SpaceX's net profit to reach around $133 billion in 2028. Based on the market capitalization of about US$1.91 trillion used in the report, its expected price-earnings ratio in 2028 is about 14.3 times, which is lower than the valuation level of some large technology companies for the same period.

Morgan Stanley maintains SpaceX's $300 price target. Among them, the enterprise's AI business contributed a valuation of 166 US dollars per share, the connectivity business contributed 115 US dollars, X and Grok business contributed 11 US dollars, and the aerospace business contributed 8 US dollars.

Furthermore, Morgan Stanley also sees the 15th Starship test flight as an important catalyst for SpaceX's future. The agency believes that if SpaceX can successfully achieve dual recycling of starship boosters and spacecraft bodies, it will significantly reduce space transportation costs and improve the commercialization prospects of new businesses such as orbital data centers.

Notably, Morgan Stanley also mentioned the possibility of further strengthening cooperation between SpaceX and Tesla, believing that closer strategic collaboration between the two companies could help them integrate the large-scale capital and technical resources needed to develop the real smart economy. However, the relevant cooperation is still a potential scenario, and it is not a confirmed deal.

Both investment banks believe that SpaceX's future valuation logic is no longer limited to traditional rocket launches and satellite internet services, but will increasingly depend on the speed of expansion of AI computing power infrastructure, the ability to commercialize computing power services, and whether Starship reuse technology can achieve key breakthroughs.

However, related growth prospects are still accompanied by high uncertainty. Bernstein stressed that huge capital expenditure and AI industry valuation fluctuations may limit room for stock price increases; Morgan Stanley warned that the progress of Starship reuse, AI business profitability, infrastructure construction costs, and potential financing needs may all affect SpaceX's future valuation performance.