Strategic Dependence – Strength or Weakness

Understanding C4ISR-C6ISR System Procurement and Implementation Decisions

tractor sketch 1

Critical defense systems become strategically dangerous when private vendors become so deeply embedded in the operational architecture that replacing them threatens the integrity of the entire defense ecosystem.

Anduril and Palantir appear to be approaching the same integrated defense offering from opposite directions. Anduril began with autonomous hardware, sensors, and field systems, then expanded upward into orchestration, software integration, deployable compute, and cross-domain mission systems. Palantir began with data integration, analysis, targeting, and decision support, then expanded downward toward operational platforms, defense applications, and execution environments.

The danger is not that either company is malicious or legally monopolistic. The danger is that procurement success can quietly become architectural dependence. Once software binds sensors, satellites, drones, tactical edge compute, targeting systems, and command workflows into one proprietary operating environment, replacing a vendor no longer means replacing a product. It can mean reconstructing the operational nervous system.

Disclosure: I am an Anduril fan. I admire both the company’s pace of innovation and its speed of execution. My concern in this instance is not about corporate intent, but the strategic dependence that can emerge when any vendor becomes deeply embedded in critical operational architecture.

  • Modern defense capability is increasingly produced by integrated systems rather than stand-alone platforms.
  • The most visible value sits in decision support and operational advantage, but the most durable power sits in the software and control layer that shapes those decisions, assigns authority, routes data, and records mission evidence.
  • Classified deployments make the real depth of dependence difficult for the public, policymakers, and even some oversight bodies to assess.
  • Interoperability claims are not the same as practical substitutability. A technically open interface can still sit inside an ecosystem that is economically, operationally, or institutionally hard to leave.
  • The correct governance question is not whether one vendor is currently dominant. It is whether the system remains governable, auditable, and replaceable if that vendor becomes indispensable.

Centralized hyperscale infrastructure and deployable tactical compute are not competing bets, they are complementary layers. Hyperscale systems support training and global workloads, regional infrastructure supports distribution and continuity, and tactical edge systems support low-latency, autonomous operation when communications are degraded, denied, jammed, intermittent, or intentionally unavailable. That layered architecture increases resilience, but it also increases the number of integration points where proprietary control can accumulate.

Strategic dependence exists when replacing a vendor becomes more costly, disruptive, or operationally risky than remaining with that vendor, regardless of whether competitors exist. The dependency arises because the vendor has become embedded in the operational architecture rather than because it controls the market. A legal monopoly concerns market power, and limits who is able to sell you a capability. Strategic dependence limits your practical ability to replace the capability after it has been integrated into your mission. One is primarily an economic condition, the other is a functional condition. Markets can remain competitive while customers become operationally captive.

Although they began from different starting points, Anduril and Palantir appear to be increasingly converging on the same category of integrated defense capability. Anduril has expanded upward from autonomous hardware, sensors, and tactical systems into AI-enabled command, deployable compute, and mission orchestration. Palantir has expanded downward from data integration, analytics, and intelligence platforms into operational command, targeting, and mission execution. Their products remain distinct, but both appear to be evolving toward integrated operational environments that combine sensing, data fusion, decision support, command, and execution within a single architectural ecosystem.

Anduril moving upward into orchestration

  • U.S. Army Next Generation Command and Control (NGC2): Anduril selected to prototype the Army’s software-defined command-and-control architecture, demonstrating expansion beyond autonomous platforms into operational orchestration. (Army)8
  • Menace-I: extends Lattice, mesh networking, mission planning, and battle management to tactical edge compute.3

Palantir moving downward into operations

  • U.S. Army TITAN description: processes multi-domain sensor data and shortens the sensor-to-shooter timeline.7
  • Maven Smart System: positioned as an enterprise mission-command and CJADC2 platform.6

Modern defense capability no longer resides primarily in individual platforms. It resides in the software layer that binds those platforms together. Drones, satellites, sensors, edge compute, communications, and cloud infrastructure all produce value, but increasingly only when integrated into a common operational environment. The software layer authenticates users, ingests data, fuses information, coordinates autonomous systems, routes decisions, records evidence, and presents a coherent operational picture. It therefore becomes the system through which every other capability is exercised. Replacing a drone changes one component. Replacing the orchestration layer changes every component connected to it. The greater the number of integrated systems, the greater the leverage exercised by the software layer that coordinates them. Much of that leverage is not visible on product pages. It emerges as deployments accumulate, integrations deepen, and operational dependence increases.

The F-35 provides a clear example of strategic dependence. Its capability depends on a tightly integrated ecosystem of mission data, program managed software updates, sustainment, logistics, engineering support, and contractor expertise. The aircraft may remain sovereign property, but its operational effectiveness depends on an external architecture that extends far beyond the airframe. Recent European criticism5 has described this arrangement as a “sustainment monopoly,” reflecting concern that control over upgrades, software, and operational data can constrain national autonomy.

A civilian analogy is the modern software-dependent tractor. The farmer may own the machine, but diagnostics, software access, repair authority, and proprietary service tools can still limit practical control over that investment. Defense systems operate under far stricter security constraints, but the architectural lesson is similar: ownership of the platform does not guarantee independence from the ecosystem required to keep it functional.

Procurement and deployment may occur through classified programs, special-operations channels, intelligence authorities, or fragmented funding streams that reveal spending without revealing operational scope. For example, U.S. Special Operations Command awarded Palantir a multiyear contract worth hundreds of millions of dollars. Public records may show the money while still obscuring deployment volume, geographic spread, mission integration, and operational dependence.4

Interoperability Is Not Replaceability: The NHS Lesson

Understanding why interoperability can still produce strategic dependence requires understanding how modern operational platforms are actually constructed. Open standards do not guarantee practical exit simply because they are “open.” Large operational platforms do more than store data. They transform it into ontologies, schemas, workflows, permissions, operational logic, decision-support systems, and applications. This transformation begins at the start of the project. Technical dependence does not require the vendor to own the customer’s data. The customer may retain legal ownership while losing practical control over its use. Once data has been cleaned, normalized, enriched, joined, and reorganized into vendor-shaped operational structures, departing is no longer a simple export. It becomes an institutional reconstruction problem.

NHS England provides a useful example. As with most customers, the NHS retains ownership of its data, yet the operational architecture becomes increasingly platform-dependent. NHS England describes its Federated Data Platform ontology as the collection of curated and transformed datasets that supplies national products and manages data exchange, cleaning, standardization, deduplication, enrichment, reference joins, metrics, and application-specific services. The strategic dependence is therefore not created by ownership of the data itself, but by the architecture built around it.1 The UK Parliament has warned that vendor lock-in and “debilitating dependencies” can create an “unacceptable point of weakness” in critical public infrastructure.²

Control Domains as the Governance Test

AI Control Domains, developed by Borealis Traders of New England, is a freely available framework intended for use by governments, industry, researchers, and practitioners to identify where authority, evidence, policy, technical enforcement, and human control must exist within AI-enabled systems. The framework includes identity and authentication, authorization, governance, policy enforcement, tool access, runtime monitoring, model integrity, data provenance, memory custody, auditability, shutdown authority, and human oversight. When applied to AI and advanced technology systems, the framework gives governance practitioners a repeatable method for identifying applicable control points, testing whether controls actually exist, and exposing gaps between declared governance and operational reality.

The question is not whether Anduril or Palantir are good companies. The question is whether we are building critical systems that remain governable when either company becomes too deeply embedded to replace. Strategic dependence rarely announces itself. It accumulates quietly through successful integrations until the cost of leaving exceeds the cost of staying.

Acknowledgment: This paper was prompted by a LinkedIn post from Krishang Sheth discussing Anduril’s Menace edge-compute platform. While that post sparked the initial line of thought, the strategic dependence analysis, governance conclusions, and AI Control Domains mapping presented here are entirely my own.

*AI Control Domains is a freely available governance framework developed by Borealis Traders of New England. Organizations are encouraged to use and reference the framework with attribution.

Contact Us Now
Browse Essays

References

Scroll to Top