AI is rapidly eroding intangible moats by collapsing the cost of coding, software development, and knowledge work—forcing a structural reassessment of information-services business models.
Markets are already repricing these risks, with U.S. software stocks down sharply as investors question advantages built on code, data access, and workflow automation.
Defensibility is shifting from bits to atoms, as physical moats rooted in capital intensity, regulation, geography, infrastructure, and real-world scarcity grow more valuable in an AI-driven economy.
Early 2026 marked a sharp reset in expectations for software and information services firms, as investors grappled with the implications of agentic AI. The catalyst was the launch of Claude Code 2, Anthropic’s autonomous coding agent. Engineers across major AI labs reported that AI was generating nearly all production code. Boris Cherny, creator of Claude Code, noted publicly that he hadn’t written code himself in more than two months1 – a clear signal that code generation had crossed a structural threshold.
Momentum accelerated with Claude Cowork, a platform designed to automate knowledge-work tasks across legal, finance, and research. Alongside collapsing development costs, these tools expanded disruption beyond coding into entire professional workflows.
Real-world data reinforced investor concerns. After three stagnant years, iOS app launches surged as the marginal cost of creation approached zero – evidence that barriers to entry were falling and competitive intensity was rising.
Agentic Coding Has Led to an Explosion of US iOS App Releases After 3 Stagnant Years
iOS Apps Released Each Month - Y/Y (%)
Source: Sensor Tower and Wells Fargo
Sentiment worsened when several large firms, including SAP, ServiceNow, and Gartner, reported disappointing earnings results. Weak fundamentals collided with growing AI anxiety, triggering a broad selloff. By early February, the U.S. software sector was down 22% year-to-date, a drawdown quickly labeled the “SaaSpocalypse.”
The market reaction points to a deeper structural shift. Value is migrating from intangible, infinitely reproducible digital assets toward physical systems anchored in scarcity, geography, and capital depth.
Software and information services are inherently replicable. As intelligence becomes abundant, margins derived from code, non-exclusive data, or workflow automation face structural pressure. These intangible moats – once prized for scalability – are proving increasingly fragile.
The physical world operates under constraints AI cannot eliminate. AI cannot manufacture atoms, move freight, build roads, generate power, or bypass permitting regimes and geography. Real-world friction – land, regulation, capital intensity, logistics complexity, and physical scarcity – creates barriers to entry that abundant intelligence cannot dissolve.
As a result, logistics networks, transportation infrastructure, energy systems, and manufacturing platforms retain structural advantages that may strengthen rather than weaken in an AI-saturated world.
Paradoxically, the rapid expansion of AI is reinforcing these physical moats. Massive investment in data centers, advanced semiconductor fabs, and power generation is straining global supply chains across:
Foundry capacity
Lithography and semiconductor tools
Memory and storage
Grid equipment and transformers
Concrete, copper, silicon, and other critical materials
Electricity production and transmission
Demand for atoms—silicon, copper, concrete, land, and especially energy—is rising faster than supply. These are constraints software cannot resolve.
The implications for investors are significant.
First, intangible-heavy business models long valued for scalability and margin expansion face structural headwinds as AI pushes the marginal cost of intelligence and knowledge work toward zero. Traditional valuation frameworks may overstate the durability of these advantages.
Second, the center of gravity is shifting toward physical scarcity and real-world constraints. Companies that own, operate, or enable the systems AI depends on – transportation networks, logistics platforms, power generation, semiconductor manufacturing, and critical materials – stand to benefit from tightening bottlenecks and sustained demand.
As AI accelerates demand for atoms over bits, capital is likely to favor businesses grounded in hard-to-replicate assets. Recognizing this shift early is not just an analytical edge – it is a moat in its own right.
The Tema Durable Quality ETF (TOLL) is aligned with this structural shift, emphasizing companies with tangible, defensible moats, including:
Food distribution: US Foods, Performance Food Group
Transportation infrastructure: AENA, Vinci, Ferrovial
Aerospace manufacturing: GE Aerospace, Airbus
Power and utilities: Entergy, Bloom Energy
Construction materials: Martin Marietta Materials, Sherwin-Williams
Semiconductor manufacturing equipment: Applied Materials, KLA, Lam Research
Portfolio holdings are subject to change. For a full, up to date list of holdings, visit the TOLL fund page.
For more of our research and recent insights on durable quality and more, view and subscribe to our insights.