The Case for Data Centers in Space Is Getting Stronger

The article argues that space-based AI data centers are becoming more feasible as launch costs fall, citing SpaceX launch economics dropping from about $5,400 per kilogram on Saturn V to $1,500 per kilogram on Falcon Heavy, with Elon Musk targeting well under $1,000 per kilogram. It also highlights Rocket Lab’s Neutron, which aims to lift 13,000 kilograms to orbit, enough for a space-based data center. The piece is largely speculative and long-dated, but constructive for SpaceX, Rocket Lab, and the broader space infrastructure theme.

Analysis

The first-order read is that orbital compute is a long-duration infrastructure theme, but the second-order winner is not necessarily the eventual data-center operator. The market is likely to price the enabling stack first: launch cadence, reusable rockets, satellite-grade communications, thermal management, and high-reliability components. That makes ASTS the highest-beta beneficiary on the comms side and RKLB the cleanest leverage point on launch-cost compression; TSLA and AMZN matter more as capital allocators and potential ecosystem catalysts than as direct earnings beneficiaries.

The key economic inflection is that launch cost has to fall into a regime where depreciation and replacement losses stop overwhelming the power/cooling savings. Until that threshold is reached, orbital AI remains a demonstrator, not a margin-accretive alternative to terrestrial hyperscale. That creates a multi-year window where the trade is really about narrative optionality and contract wins, not current fundamentals. In other words, the stock reaction can outpace revenue recognition by 12-24 months.

Contrarianly, the consensus is probably underestimating the bottlenecks outside launch: radiation hardening, on-orbit servicing, latency for inference-heavy workloads, and insurance underwriteability. Those frictions favor a phased adoption model—edge inference and specialized workloads first, general-purpose training last—so the biggest upside may accrue to infrastructure vendors selling picks-and-shovels rather than the end-user platform. If launch economics continue to improve, the market could re-rate a basket of space enablers before it assigns any meaningful terminal value to orbital data centers themselves.

Near term, the main risk is that the story is too early and becomes a multiple-expansion trap if milestones slip. A credible reversal would be any launch failure, delayed reusable medium-lift rollout, or evidence that uplink/downlink economics dominate the cooling savings. On a 6-18 month horizon, the right lens is optionality: pay for exposure to the enabling chain, but size it as a venture-style barbell rather than a core fundamental value trade.

Key Decisions for Investors

• Long ASTS on a 6-12 month horizon as the highest beta comms beneficiary; use a tight trailing stop because the upside is narrative-driven and vulnerable to execution slippage.
• Long RKLB into Neutron progress milestones over the next 6-18 months; this is the cleanest direct exposure to falling launch costs and should rerate ahead of revenue inflection if cadence improves.
• Pair trade long RKLB / short a basket of mature terrestrial data-center infra names on a 12-month view to express the migration-to-space optionality while hedging generic AI capex exposure.
• Buy TSLA or AMZN only as a small catalyst-option position, not a core thesis, since the economic benefit is indirect and likely to show up first in ecosystem influence rather than earnings.
• If ASTS or RKLB rallies sharply on headlines without confirmed contracts, fade part of the move with short-dated call overwrites; the setup is prone to overextension before tangible cash flow arrives.