Version 3

SpaceX's S-1 IPO filing, published May 20, 2026, confirmed a $1.25 billion-per-month Cloud Services Agreement with Anthropic for COLOSSUS and COLOSSUS II compute capacity through May 2029 — roughly $45 billion total [2][3]. The filing simultaneously disclosed SpaceX's absorption of xAI into a new SpaceXAI division and its ambition to build orbital data centers, with SpaceNews reporting Anthropic has agreed to 'consider using' SpaceX's planned satellite-based compute facilities [15] — potentially explaining the deal's previously opaque 'space development component' [1]. On the buyer side, Anthropic's financial picture has sharpened considerably: annualized revenue is reportedly $30 billion as of April 2026 [18][19], its valuation has reached $900 billion [20], and Reuters reported it is nearing its first quarterly profit [17]. On the seller side, Morningstar has expanded its institutional critique of xAI's finances — which burned $6.4 billion last year [26] — with multiple outlets now amplifying the 'financials look reckless' verdict [22][21][23].

The story has evolved from a single large deal into a window on two simultaneous high-stakes bets: Anthropic is growing fast enough that $90 billion in contracted compute across SpaceX, Amazon, and Google may be sustainable rather than reckless; and SpaceX is betting that orbital data centers — which its own S-1 warns may not pay off [16] — can extend its infrastructure dominance from ground to orbit, with Anthropic as a potential anchor customer for both. The financial paradox at the center remains: SpaceX's model business (xAI/Grok) is losing billions while its infrastructure arm collects billions from the AI labs it competes against, and investors are being asked to price that contradiction at up to $2 trillion [10].

The Anthropic–SpaceX compute deal entered the public record in two phases. A CNBC report from May 7, 2026 described a joint announcement covering a compute partnership and a space development component [1]. The full financial terms became visible on May 20 when SpaceX published its S-1 IPO filing: Anthropic agreed to pay $1.25 billion per month for capacity on the COLOSSUS and COLOSSUS II data centers in Memphis, Tennessee, running from May 2026 through May 2029 and totaling approximately $45 billion [2][3]. Capacity is ramping at reduced fees through mid-2026 before reaching full monthly payments, and either party can exit with 90 days' notice [3]. COLOSSUS I's full capacity is dedicated to Anthropic [4][5], while COLOSSUS II is equipped with NVIDIA GB200 NVL72 systems also used to train Grok 5 [6].

The S-1 simultaneously disclosed that SpaceX has formally acquired xAI and integrated it into a new SpaceXAI division overseeing the Grok model family [7]. The filing's strategy section is titled 'Our AI Compute Infrastructure Advantage and Growth Strategy' and positions SpaceX's launch and Starlink businesses as supporting infrastructure for its AI ambitions [8][7]. The company claims a $26.5 trillion total addressable market and cites infrastructure cost research to argue its build-cost advantage is a durable moat [9]. Valuation targets referenced in financial media range from $1.75 trillion [4] to $2 trillion [10]. The SpaceX deal is one piece of a broader Anthropic compute strategy that also includes a reported $33 billion commitment with Amazon [11] and a large Google Cloud arrangement [12] — a total contracted compute footprint exceeding $90 billion.

The 'space development component' of the deal [1] — previously unexplained — has now gained a plausible interpretation. SpaceX has filed regulatory plans for a million-satellite orbital data center constellation [13], and Elon Musk has publicly stated that 'SpaceX will be doing this' on orbital compute [14]. SpaceNews reported that Anthropic has agreed to consider using SpaceX's orbital data center satellites [15], which would make it a potential anchor customer for both ground and orbital infrastructure. The ambition, however, comes with a caveat from SpaceX itself: the company's S-1 warns investors that orbital AI data centers 'may not pay off' [16] — an unusual disclosure that simultaneously sells the vision to customers and hedges it for regulators and investors.

On the financial sustainability of the overall picture, Anthropic's position has sharpened. Reuters reported the company is nearing its first quarterly profit [17], and multiple sources — including a Reddit discussion and Binance financial coverage — cite annualized April 2026 revenue of $30 billion [18][19], with valuation reportedly reaching $900 billion following Google and Amazon investment commitments [20]. If accurate, these figures suggest Anthropic's revenue growth may be outpacing even its extraordinary infrastructure costs. Against this, Morningstar has substantially expanded its institutional skepticism about the SpaceX side of the equation: beyond its initial 'financials look reckless' verdict on xAI [21], Morningstar has published that 'xAI is draining SpaceX's cash, with little to show for it' [22] and that 'the Grok maker's losses are accelerating' [23][24] — a critique now amplified by PitchBook [21] and InvestmentNews [25], which flagged rising concentration risk from AI bets embedded in seemingly non-AI assets like SpaceX. TechCrunch further detailed the $6.4 billion xAI burn rate and the trajectory suggesting that spend is not declining [26].