Technology thesis · Computing Infrastructure
high conviction growthDigital twins
Digital twins are the operating end of a physics-AI surrogate stack that now starts at the design bench; value is moving from the validated solver to the proprietary engineering data it trains on.
Position maintained continuously · last reviewed Jun 24, 2026
The thesis
State of the art (2026)
In 2026 the centre of gravity has moved from visualisation to physics-AI surrogates that collapse the line between design and operations. PhysicsX raised a $300M Series C at roughly $2.4B (8 June 2026, Temasek-led), running one model from early design to live operational twins. NVIDIA Omniverse plus Cosmos anchors the AI-native 3D stack; Siemens is training a 150-petabyte Industrial Foundation Model after its Altair buy, and Synopsys closed its $35B Ansys deal in July 2025. The contested layer is no longer the solver but the high-fidelity simulation and fleet-telemetry data the surrogate learns from, which OEMs, not vendors, increasingly own.
Design and operations are converging on one model
The same physics-AI surrogate that explores a design space in seconds becomes the live digital twin of the equipment in service. PhysicsX runs that arc from early-stage design through to real-time operational twins; in power, AI-driven electricity demand is pulling turbine and plant twins (GE Vernova's APM SmartSignal, AI-assisted tuning) into the same loop. The design tool and the operational twin stop being separate products.
The moat is the training data, not the solver
Once a neural surrogate reproduces a solver's output in seconds, the marginal cost of an extra design evaluation falls to near zero and the per-seat solver licence decouples from value. What remains scarce is the high-fidelity simulation and operational data the surrogate learns from. The solver vendors generate simulation data; the OEMs own fleet telemetry and physical test results the vendors do not. For the highest-value physics the data owner is the OEM, which is why Siemens is buying its way to 150 petabytes.
Everything below is live inside CanaryIQ
The full analysis behind the verdict — the structure is real; the content unlocks when you log in.
Signal stack
Evidence stacked leading → lagging
Technology-native KPIs
Metrics that predict trajectory, tracked over time
Landscape map
Who builds what — and who depends on whom
Catalyst calendar
Dated events that will move the position
Technology roadmap
Milestones on the path to maturity
Watchlists
Companies, people and papers — each with a remove-by condition
Decision frameworks
The same call, framed for your desk
Thesis changelog
When our view changed, and why
Change our mind
4 disconfirming conditions
The rest is inside
You've read the verdict. The file is much deeper.
The full signal stack, technology-native KPIs tracked over time, the landscape of who depends on whom, the dated catalyst calendar, decision frameworks for every desk, live watchlists and the changelog of every time our call on Digital twins has changed — all live inside CanaryIQ.