Drones are now expected to become a $250 billion market by 2035

Barclays projects the global drone market will reach $250 billion by 2035, with the market having doubled from ~$20B in 2020 to >$40B in 2025. The report frames a shift to 'Physical AI' that moves value from hardware to compute, data centers and software, front-loading costs into AI capex while promising lower long-term operating costs. Key risks for scaling autonomous swarms are energy availability and critical minerals bottlenecks, which investors should monitor as constraints on deployment and pace of innovation.

Analysis

The structural shift from discrete drone hardware to "Physical AI" reallocates economic value from low-margin airframes to high-margin compute, software and systems integration. That implies winners will be firms selling dense inference infrastructure, orchestration stacks and recurring SaaS for autonomy — not necessarily the legacy avionics OEMs that still trade as hardware plays. Expect profit pools to concentrate in firms that control data, models and edge/server co-design; this elevates suppliers of high-efficiency servers, power electronics, and real-time networking as implicit “defense primes.”

Second-order supply-chain stress will come from three chokepoints: high-density compute power (data-center and edge batteries), critical minerals for motors and batteries, and local-grid capacity for distributed launch/charging nodes. Those constraints create a multi-year capex curve — initial unit deployment is cheap, but scaled autonomous swarm economics require heavy up-front spending in power, cooling, and custom silicon; shortages or price shocks in any input can slow deployments by 12–36 months. Geopolitical moves (export controls, sanctions) are asymmetric catalysts: they can instantly reroute value to Western suppliers while starving adversary ecosystems, amplifying winners/losers beyond product specs.

Near-term reversal risks are concrete: a regulatory crackdown on autonomous lethal systems, a macro capex pause, or a sudden GPU/ASIC supply relief that rerates current compute suppliers downward. Conversely, a handful of defence contracts or a government-funded data-center program within 6–18 months would rapidly re-rate compute-centric vendors. Finally, market consensus underprices integration complexity and over-weights unit counts; the path to scale is operational and capital intensive, so public multiples should be tied to demonstrable recurring revenue and supply-chain control rather than unit shipment growth alone.

Key Decisions for Investors

• Long SMCI (overweight, 6–12 months): Buy stock or 9–12 month call spreads to express exposure to high-efficiency server demand for edge/defense AI. Risk/reward: target +40–80% if AI capex sustains, downside -25–35% if GPU/ASIC supply normalizes or enterprise capex stalls. Size: tactical overweight 1.5x normal sector weight with 20–25% stop-loss on equity basis.
• Pair trade — Long SMCI / Short AVAV (6–12 months): Go 1:1 notional long SMCI and short AVAV to capture secular shift from airframe manufacturing to compute/software. Target: SMCI outperformance of 30–50% over AVAV within 12 months. Tail risk: broad sector bid lifts both; cap position size so net market beta ~0.
• Long APP (6 months): Buy 6–12 month calls or 20–30% size equity exposure to play software/monetization upside from AI-driven edge data flows and ad-tech optimization that benefit platforms with strong LTV economics. Risk/reward: target +30–60% if monetization accelerates; downside -40% in ad slowdown. Use options to limit downside.
• Event hedge / watchlist — BCS (3–12 months): Keep a small long on BCS as a liquidity/underwriting proxy ahead of potential defense-tech M&A and financing waves; rotate into after M&A/contract announcements. Use this as smaller, tactical exposure rather than core position; Catalysts: announced government programs or large prime deals that require bank financing.