2 Space Stocks to Buy in September

The article frames the global space sector as potentially reaching $1.8T by 2035 and highlights two investment angles: MDA Space and SpaceX. MDA reported adjusted net income of CA$102.5M for the period ending June 30, 2026 (up from CA$84.4M prior year) and the stock is up 48.6% YTD versus the S&P 500’s 13.2%. For SpaceX, the key bull case is AI-driven revenue capacity: projected $26B annual compute rentals from Anthropic and Alphabet (vs. 2025 sales of $18.7B) and additional potential from Reflection AI, while analysts (75% buy) cite a median $217 one-year target (~51% above the Aug. 31 close of $143.69), despite high stated price volatility.

Analysis

The more investable angle is not the headline TAM, but who captures scarce, high-margin bottlenecks as the market matures. MDA looks like the cleaner way to own the buildout: mission-critical hardware with lower execution beta, while most of the upside in SpaceX remains dependent on proving that AI-linked infrastructure can scale without destroying returns on capital. The second-order winner set is broader than named here: launch-adjacent suppliers, radiation-hard components, thermal management, and ground-station software vendors should benefit if orbital compute moves from narrative to procurement.

The market is likely over-discounting near-term revenue optionality from space-based AI. The binding constraint over the next 12-24 months is not demand for compute; it is capex intensity, launch cadence, power density, and regulatory/orbital debris risk. If those constraints bite, the multiple should compress first in the highest-expectation name (SPCX) before any of the smaller, asset-backed beneficiaries like MDA rerate.

Contrarian read: the consensus is treating this like a straight-line AI monetization story, but the more likely path is lumpy contract announcements followed by long gaps before cash flow shows up. Alphabet is a modestly positive read-through only if it is using SpaceX as a tactical supply source, not if it is forced to subsidize an uneconomic platform. The thesis breaks if SpaceX can show sequential improvement in utilization and unit economics over the next 2-3 quarters; otherwise this remains a narrative trade rather than a fundamental one.

Key Decisions for Investors

• Long MDA / short SPCX as a 6-12 month relative-value pair: own the tangible, lower-duration cash flows and fade the most crowded AI-space expectation premium; stop if SPCX proves recurring AI backlog conversion or if MDA guidance stalls.
• Buy SPCX only on 15-20% pullbacks, not strength, and size it like a venture-style position: the risk/reward is asymmetric only if you can tolerate 12-24 month drawdowns while waiting for proof of margin durability.
• Add GOOGL as a small optionality hedge on AI compute sourcing diversification, but treat it as de minimis to earnings; this is a watch item, not a primary catalyst trade.
• Set an alert on MDA earnings/backlog and margin progression over the next 1-2 quarters; if adjusted earnings growth decelerates or order conversion weakens, the stock’s rerating thesis is likely done.
• If a public pure-play launch or satellite-services proxy emerges, favor a basket of picks-and-shovels over SPCX; that basket should have better risk/reward if orbital AI remains capital intensive and slow to monetize.