Before the Golden Dome: Canada's Case for Sovereign Offshore Defence Infrastructure
In February 2026, Prime Minister Carney stood before a parliamentary committee and confirmed what defence analysts had been quietly documenting for years: 75 cents of every dollar Canada spends on defence capital procurement leaves the country. The F-35 program returns 6.1 cents per dollar in domestic industrial benefit. Canada is, in a very literal sense, paying for its own strategic dependency.
Then came the Golden Dome proposal — a US-designed, US-hosted, US-operated missile defence architecture that Canada is being invited to join. The pitch is compelling on its face: continental defence, shared burden, integrated capability. The question that doesn't get asked loudly enough is simpler: what happens when Washington decides to adjust the terms?
We've seen how fast that can happen.
The Coverage Gap Nobody Talks About
Canada administers the world's largest Exclusive Economic Zone — 5.6 million square kilometres of ocean. Current persistent sensor coverage sits at approximately 34%, stitched together from RADARSAT passes, coastal radar stations, and data feeds from US-controlled NORAD infrastructure. The remaining 66% is monitored intermittently at best.
Russia has deployed more than 40 military installations north of the Arctic Circle since 2014. China declared itself a "near-Arctic state" in 2018 and has been systematically mapping Arctic seabed topology ever since. Both nations understand that the Arctic is not a frozen desert — it is the shortest route between the Pacific and Atlantic, and the most significant untapped resource corridor on the planet.
Canada's response has largely been to point at NORAD and call it sovereignty.
A Norwegian Startup Changes the Equation
In March 2026, San Francisco-based Aikido Technologies announced the AO60DC — a semisubmersible floating offshore wind platform that integrates, on a single football-field-size structure:
- A 15–18 MW floating wind turbine
- A 10–12 MW liquid-cooled AI compute module, cooled passively by seawater
- Subsea fibre communications backbone
- Onboard battery storage for grid stress events
The platform is self-powered, self-cooled, and anchored to the seabed. A 100-kilowatt prototype deploys in Norway's North Sea by end of 2026. A commercial 10–12 MW version is targeted for UK waters by 2028.
The observation that follows is straightforward: if you bolt a sensor suite onto that platform — marine radar, passive sonar array, AIS with spoofing detection, optical and infrared cameras — you have a persistent, self-powered, sovereign domain awareness node in Canadian territorial waters. One that nobody in Washington can switch off.
What the Model Says
The CanuckDUCK Ducklings simulation platform ran the cascade for Proposal E-101 (Canadian Sovereign Maritime Sensor Network Act) across a 6-epoch cycle. The cross-proposal wiring tells an interesting story:
Platform deployment (+100%) cascades through defence spending into NATO eligibility ratios, employment, GDP across all provinces, and provincial surpluses within the first three epochs. The spending stays in Canada.
Canadian UBO share (+50%) — the requirement that contracts flow to Canadian-owned entities — flips the capital outflow equation within one epoch, improves trade balance by epoch 2, and reaches Arctic transit revenues and critical mineral processing by epoch 3. A defence procurement decision connects to a mining variable in 18 months of model time. That's not a coincidence — it's the supply chain.
US defence data dependency — currently modelled at 0.71 on a 0–1 scale — drops to approximately 0.21 by the end of Phase 1 deployment (3 platforms, 5 years). Not zero. NORAD doesn't disappear, nor should it. But the relationship shifts: Canada arrives at the table with demonstrated, sovereign, independent capability rather than as a dependent asking for access to its own approaches.
The sovereignty index gain follows one epoch after the capability gain. The model captures something real there — having the sensors and using them diplomatically are two different decisions.
The Case For
The argument for this approach over Golden Dome participation rests on three pillars:
Strategic independence. Platforms in Canadian sovereign waters, built by Canadian firms, operated by Canadian crews, running on Canadian wind, cannot be conditioned or withdrawn by a foreign government. The relationship with Washington matters — but it matters more when Canada isn't entirely dependent on it.
NATO compliance without capital flight. Every dollar spent building and deploying these platforms counts toward NATO's defence spending targets. It is simultaneously NATO-eligible and domestically retained. The 75% capital leakage problem Carney identified is structurally solved — you can't offshore-manufacture a platform that has to be built in Halifax.
Dual-use economic return. A defence platform that also generates 15–18 MW of clean power and hosts sovereign AI compute infrastructure is not a sunk cost. It is infrastructure. The grid gets renewable capacity. Canadian researchers and government agencies get compute that isn't sitting on AWS or Azure. The Arctic gets a permanent Canadian presence that isn't a flagpole and a ranger station.
The Case Against
The counterarguments deserve equal weight.
Cost and complexity. Three Arctic-hardened platforms at approximately $1.08B each — a 35% premium over North Sea specifications — represents $3.2B in Phase 1 capital. For a government managing a structural deficit, the opportunity cost question is legitimate: what else could $3.2B buy? The Aikido model is proven at small scale in the North Sea. Arctic hardening at this scale has no direct precedent. The engineering risk is real.
Integration, not independence. Canada's defence relationship with the United States is not simply a data dependency — it's the foundational architecture of continental security. NORAD, Five Eyes, joint Arctic command — these are not optional add-ons. A sovereign sensor network supplements that relationship; it cannot and probably should not replace it. Critics argue that the Golden Dome framework, whatever its political complications, represents a genuine upgrade to continental defence capability that a handful of offshore platforms cannot replicate.
Maintenance in the Arctic. Crew transfer vessels operating in the Beaufort Sea in February is not a North Sea problem. Ice-class vessel requirements, extreme cold electronics hardening, and the logistics of keeping platforms operational in conditions that regularly destroy equipment add operational cost that the capital estimate doesn't fully capture. Beaufort Sea oil platforms provide precedent, but offshore oil infrastructure has also produced some of the most expensive maintenance failures in Canadian history.
Technology readiness. The Aikido AO60DC is a prototype. The combined wind-compute-sensor integration at Arctic scale sits at roughly Technology Readiness Level 6–7. Canada would not be buying proven infrastructure — it would be co-developing it. That is a different risk profile than the proposal's confident tone suggests.
The Question the Model Can't Answer
The simulation can trace capital flows, sovereignty indices, and employment cascades with reasonable confidence. What it cannot model is political will.
Canada has been announcing Arctic sovereignty infrastructure for thirty years. The Diefenbaker-era NORAD debates, the Trudeau-era Arctic waters protection, the Harper-era icebreaker programs, the Trudeau II-era defence reviews — each cycle produces compelling analysis and underwhelming follow-through. The platforms that would change the dependency calculation require a government willing to sustain capital commitment across multiple electoral cycles, through cost overruns and engineering setbacks, against a defence establishment with deep institutional ties to US procurement pipelines.
Whether that political will exists is not a modelling problem.
For Discussion
- Is the Golden Dome framework fundamentally incompatible with Canadian strategic sovereignty, or is participation the pragmatic choice given the state of the relationship with Washington?
- At what point does defence spending that qualifies for NATO targets but stays entirely in Canada become the standard expectation rather than the exception?
- Should Canada be co-developing emerging offshore infrastructure technology with a 5-year capital commitment, or waiting for the technology to mature before procuring proven systems?
- The model shows D-05 UBO procurement requirements rippling into Arctic transit revenues and critical mineral processing within three epochs. Does that supply chain connection change how you think about defence procurement as industrial policy?
- If Canada had deployed three sovereign sensor platforms in the Beaufort Sea in 2020, would the current conversation about US Arctic encroachment look different — and does that counterfactual change the investment calculus?
This article references simulation outputs from the CanuckDUCK Ducklings platform (Proposal E-101, deployed March 2026). Model outputs represent causal cascade projections based on registered variable relationships — they are analytical tools for structured debate, not predictions.