By Nancy Kondelidou, an investment analyst in the Liontrust fixed Income team
SpaceX hosted one of the most interesting roadshows I have attended in a long time. Unlike a typical credit presentation, which tends to focus primarily on leverage, liquidity and financial metrics, the discussion centred on the company’s longer-term strategic vision across launch, connectivity and AI.
While the presentation was certainly more thematic than most, beneath the ambition sits a very real business generating meaningful cash flows today, which ultimately explains why the transaction attracted such strong demand.
The key takeaway from the roadshow is that SpaceX should no longer be thought of primarily as a rocket company. Management’s vision can be summarised as Rockets → Satellites → Connectivity → Compute → AI Applications, with each layer supporting the next.
See also: Knacke’s money maps: SpaceX and AI’s search for cheapenergy
While the headlines focus on Starship, AI and Mars, the economic engine of the business today is unquestionably Starlink. The connectivity segment generated approximately $11.4bn of revenue and $7.2bn of EBITDA (earnings before interest, tax, depreciation and amortisation) in 2025, with over 10 million subscribers across 164 countries.
In contrast, launch economics are relatively modest and the AI segment remains loss-making. Put simply, Starlink is funding much of the broader ecosystem.
Investment grade ratings despite negative free cashflow
From a bondholder’s perspective, this distinction is critical. Much of the debate surrounding the transaction has focused on whether the rating agencies have been unusually generous in assigning investment-grade ratings.
SpaceX currently exhibits characteristics that are atypical for a Baa1/BBB credit. Its cash generation is offset by enormous capital expenditure (capex) plans, meaning that free cashflow – a frequently used financial metric which subtracts capex from operating cash flow – is negative.
There are also expectations of rising leverage over the next several years; S&P has suggested debt could increase materially as the company funds AI infrastructure and other growth initiatives.
However, rating agencies appear willing to look through today’s cash burn because of three assets that are exceptionally difficult to replicate: Starlink, launch dominance and the emerging AI infrastructure platform.
What makes the story particularly interesting is that SpaceX is not simply attempting to build another AI model. Through Colossus and Colossus II, management is building large-scale compute infrastructure while simultaneously developing Grok and broader AI applications. In effect, the company wants to monetise AI through both selling compute and selling intelligence.
The recently announced compute agreements with Anthropic and Google suggest infrastructure monetisation is already beginning, reducing reliance on Grok becoming an immediate commercial success. In many respects, the company appears to be positioning itself somewhere between a hyperscaler, a telecommunications operator and an AI developer.
Difficulty in valuing the risks
That said, the bond market’s initial reaction highlights investors are still struggling to determine how to value the risk. Despite massive demand during the offering process, Bloomberg reports that the bonds have widened significantly in secondary trading, with the 30-year bonds moving around 28bps wider than issue levels and approximately $305m of mark-to-market losses accruing to investors shortly after pricing.
Several market participants attributed this weakness to fast-money accounts participating in the deal with the intention of flipping bonds rather than holding them long term.
The longer-dated maturities appear to have suffered the greatest pressure, reflecting investor scepticism around underwriting such a long-duration story where much of the value depends on future execution rather than current cashflows.
See also: WisdomTree Space Economy ETF adds SpaceX ahead of the index
Importantly, I do not think this secondary market weakness necessarily changes the long-term credit case. In fact, much of the widening appears technical rather than fundamental.
The deal was upsized significantly, a large amount of new investment-grade supply has come to market as technology companies race to finance AI investments, and SpaceX remains difficult to compare with traditional corporate issuers.
Investors are effectively trying to fit a company with characteristics of a telecom operator, launch provider, hyperscaler and AI developer into a conventional investment-grade framework.
Why we passed on the deal
The more speculative elements of the story – orbital AI data centres, lunar manufacturing, asteroid mining and the broader space economy – undoubtedly made for compelling roadshow content.
Management’s thesis that power, cooling and land constraints could eventually make space-based compute attractive is intellectually intriguing, particularly given SpaceX’s unique combination of launch, satellite and connectivity assets.
However, as bondholders, these initiatives should be viewed primarily as upside optionality rather than contributors to the current investment case. Today’s value remains overwhelmingly driven by Starlink’s subscriber growth and cash generation.
From a valuation perspective, I can understand why the deal attracted such strong interest, particularly given the quality of the Starlink franchise, the company’s dominant position in launch and the optionality embedded in its AI ambitions. However, we ultimately chose not to participate in the transaction.
See also: SpaceX: Should investors believe the hype?
While we have a high degree of conviction in the underlying quality of the business and the strategic moat SpaceX has established across launch, satellites and connectivity, a meaningful portion of the investment case remains dependent on businesses whose economics are yet to be fully proven.
The AI infrastructure build-out, Grok monetisation strategy and longer-term concepts such as orbital compute undoubtedly have significant potential, but at this stage require investors to underwrite a considerable amount of future execution.
As a result, we felt the risk-adjusted opportunity was not sufficiently compelling to participate at issuance.
The recent widening in the bonds perhaps reflects this challenge, with investors continuing to grapple with how to value a company that sits somewhere between a telecommunications operator, launch provider, hyperscaler and AI developer.
While the secondary market weakness may prove largely technical, it reinforces our view that there is still a meaningful gap between today’s cash-generating business and some of the longer-dated opportunities embedded in the valuation.
Nevertheless, SpaceX remains a highly interesting credit story and one we expect to revisit as the various elements of the investment case continue to develop.
