Trajectory Daily Brief: 12 February 2026

Europe | Hypersonic | Startups and defence primes race to close hypersonic gap with Russia and China but face a decade-long structural deficit

Situation

German-British startup Hypersonica completed a Mach 6 test flight covering 300 km on February 3, then closed a €23.3 million Series A. Other ventures — Destinus and Polaris Spaceplanes — are pursuing cruise missiles, interceptors, and reusable test vehicles. The European Defence Fund’s 2026 programme allocates €168 million for hypersonic defence and €68 million for a classified glide-vehicle demonstrator.

Russia operates three hypersonic systems (Avangard, Kinzhal, Zircon). China fields the DF-ZF on DF-17 missiles and is testing the DF-27. Europe has no operational hypersonic weapon and no indigenous interceptor designed to engage one. NATO’s own assessment confirms its current systems were never built to track or intercept hypersonic threats.

Context

The startup model faces a proportion problem. Hypersonica’s entire raise is roughly what MBDA spends in a quarter. Hypersonic weapons demand ceramic matrix composites requiring years to qualify, guidance systems that must function inside plasma sheaths, and scramjet engines no country has made reliably sustainable. The US ARRW programme was cancelled after repeated failures. Missile development typically takes a decade — for organisations with unlimited budgets and deep institutional knowledge.

Europe’s incumbent primes offer no shortcut. France’s ASN4G nuclear-capable hypersonic missile targets 2035. MBDA’s HYDIS2 interceptor aims for the same date. Russia deployed Avangard in 2019; China fielded the DF-17 in 2020. By the time Europe arrives, adversaries will have fifteen additional years of production refinement and operational doctrine. The IISS found that Europe’s exo-atmospheric interception capability is provided solely by the United States — not a gap but a structural dependency.

The EU defence-industrial base remains fragmented across 27 member states with overlapping agencies and sovereign procurement politics that block the consolidation needed for scale.

Trajectory

Europe may be building toward a capability whose strategic premium is already eroding. Directed-energy weapons — the US Navy’s HELIOS laser, advancing solid-state systems — could invert hypersonic economics: defenders spending kilowatt-hours against attackers spending tens of millions per missile. AI-driven sensor networks threaten to collapse the decision-time compression that gives hypersonic speed its battlefield value.

The likeliest outcome is not that Europe closes the gap but that it arrives in the mid-2030s with a first-generation capability entering a battlefield already shaped by countermeasures designed to neutralise it. The strategic question shifts from “can Europe build hypersonic weapons” to whether the billions required would be better spent leapfrogging to the defensive technologies that make them obsolete.

China | Indo-Pacific | AUKUS submarine dispersal to Western Australia opens targeting gap but creates isolated, undefended logistics node

Situation

Under AUKUS Pillar I, up to four U.S. and one British nuclear-powered submarine will rotate through HMAS Stirling near Perth beginning as early as 2027. The base sits more than 5,500 kilometres from China’s nearest confirmed conventional missile reach, placing it beyond the DF-26’s range.

Australia is pouring up to A$8 billion into Stirling’s expansion, with A$738 million in priority infrastructure works already scoped. The surrounding suburbs of Rockingham, Kwinana, and Henderson are experiencing a construction boom driven by an estimated A$70 billion in defence-related spending.

The base currently lacks any integrated air and missile defence — no Patriot batteries, no THAAD, no layered architecture against the bomber and hypersonic threats China is expected to field by the early 2030s.

Context

The strategic logic is sound in principle. Dispersing submarines beyond first-strike range forces Beijing to divide its targeting across more nodes. Four Virginia-class boats departing Stirling could be transiting the Lombok Strait, patrolling the Malacca approaches, or operating in the eastern Indian Ocean — China must plan for all possibilities. Its ISR constellation now exceeds 510 satellites, but a dived submarine remains effectively invisible.

The logic frays at the pier. Submarines in port are fixed, observable targets with predictable maintenance cycles. Stirling is connected to industrial Australia by a single rail corridor through some of the continent’s most sparsely populated terrain, making it far less logistically resilient than Guam’s Pacific network. The U.S. Navy’s existing maintenance backlog already strains four dedicated submarine shipyards; adding a fifth node at the end of an 8,000-nautical-mile supply chain risks diluting capacity rather than multiplying it.

China’s H-20 stealth bomber, anticipated to exceed 10,000-kilometre range, could close Stirling’s distance advantage by the early 2030s — converting a sanctuary into a target on a timeline set by Beijing’s procurement schedule.

Trajectory

Stirling’s value is real but time-limited and conditional. It complicates China’s calculus today because distance exceeds missile range and submarines at sea are hard to track. That window narrows as Chinese long-range strike and ISR capabilities mature.

The deeper risk is strategic lock-in. Billions in sunk infrastructure creates political gravity that resists adaptation. If Stirling becomes the southern anchor of allied submarine operations without layered defences, logistic redundancy, or hardened dispersal options across northern Australia, it replicates the concentration problem it was designed to solve — just further from the fight and harder to sustain.

The measure of success is not whether Stirling complicates targeting today. It is whether the investment remains strategically sound when Beijing’s kill chains catch up with the geography.

Middle East | Analysis | US Navy expended 220 missiles defending Red Sea shipping against $20,000 drones, exposing unsustainable cost-exchange ratio in carrier strike group operations

Situation

Between November 2023 and early 2025, twenty-six US Navy vessels rotated through the Red Sea to counter Houthi drone and missile attacks on commercial shipping. The fleet conducted 380 separate engagements, expending 220 missiles and 160 five-inch shells. The USS Carney alone fought for ten continuous hours in October 2023, the Navy’s most intense surface combat since World War II.

Every engagement was a tactical success — no carrier was hit. But interceptors costing $1–5 million each were destroying Iranian-design Shahed drones worth roughly $20,000 apiece. The Navy cannot reload vertical launch cells at sea; each ship must withdraw to port for weeks to rearm, creating a fight-withdraw-reload rotation that concedes continuous presence.

Context

The carrier strike group’s layered air-defence architecture — Aegis radar, Standard Missiles, Evolved SeaSparrows, Phalanx CIWS — was engineered to defeat expensive Soviet-era anti-ship cruise missiles launched from targetable platforms like bombers. The Shahed-136 doesn’t defeat this system. It exhausts it. A slow, low-flying drone with a tiny radar cross-section built from commercial automotive components forces the same defensive response as a sophisticated threat, at a hundredth of the cost.

The Navy’s own Navigation Plan 2024 identifies terminal defence and contested logistics as top capability priorities, effectively acknowledging the fleet was designed for a different threat environment. SM-6 production runs at roughly 500 per year. Iran and proxies manufacture drones in days using globally sourced parts. An Arleigh Burke destroyer’s 96 vertical launch cells represent a fixed, non-replenishable magazine — a closed system fighting an open one.

This is not a novel dynamic. Iran’s drone programme was never designed to sink a carrier. It was designed to make carrier operations so costly and logistically burdensome that sustained presence becomes a strategic liability.

Trajectory

China is the audience that matters. Chinese military researchers have studied the Red Sea campaign closely and already launched a prototype drone carrier. The lesson is straightforward: force US ships into a defensive posture and their magazines become the binding constraint. In the Taiwan Strait — 130 km wide, with no permissive rear area for rearming — the geometry is far worse than the Gulf of Aden.

The Navy is pursuing directed-energy weapons and cheaper interceptors, but none are fielded at scale. Until the cost-exchange ratio inverts, every dollar spent on carrier strike group operations buys less deterrence than it did a decade ago. The era of disposable weapons is compressing the strategic utility of history’s most expensive military asset.

Global | Technology | AI-enhanced cyber espionage operates inside detection blind spots, with dwell times exceeding 200 days against sensors built for human-speed threats

Situation

Current cyber espionage detection infrastructure—endpoint sensors, network monitors, behavioural analytics, SIEM platforms—was architected to catch human operators making mistakes at human speed. AI-enhanced reconnaissance makes fewer mistakes and operates at machine tempo. Mandiant’s 2024 data puts espionage-specific dwell time past 200 days, while red teams achieve objectives in five to seven.

Fifty-seven state-linked threat actors from China, Iran, North Korea, and Russia used AI operationally in 2024 for malware development, phishing, reconnaissance, and automated attack sequencing. A tool called Villager, built on DeepSeek AI, was downloaded over 17,000 times.

SIEM platforms typically retain detailed logs for 30 to 90 days. Espionage campaigns routinely unfold over months or years, operating entirely within the gap between what happened and what defenders can still retrieve.

Context

The failure is structural, not incremental. Each detection layer assumes something AI-enhanced attackers have learned to deny it. EDR assumes malicious code will execute—living-off-the-land techniques use only signed, trusted system tools. NDR assumes malicious traffic will cross monitored segments. UEBA assumes adversary behaviour will deviate from baselines. An AI agent using native tools can randomise timing, vary command syntax, and distribute activity across sessions until it dissolves into legitimate administrative noise.

The Salt Typhoon penetration of major US telecom providers in 2024 exposed a compounding problem: cloud and telecom control planes generate logs that are often incomplete, proprietary, or inaccessible to the organisations they serve. Defenders cannot detect what infrastructure was never designed to make visible. AI-driven offensive operations now exploit four compounding advantages simultaneously—reconnaissance at scale, social engineering without detectable patterns, adaptive evasion of deployed detection rules, and autonomous lateral movement that eliminates the command-and-control cadence defenders rely on to spot intrusions.

Trajectory

The asymmetry is mathematical, not budgetary. Offensive AI needs to find one viable path; defensive AI must secure all of them. Every detection rule deployed reveals its boundary to an adversary capable of probing it. The kill chain has become a kill graph—branching, adaptive, and resistant to linear defensive models.

Organisations still measuring security posture by alert response times are optimising for a contest that has already shifted beneath them. The capability gap will widen until detection architectures are rebuilt around the assumption that adversaries operate with machine-speed decision-making, leave no behavioural signature, and persist beyond log retention windows. That rebuild has not begun at scale.

Until tomorrow.