An Independent Research Imprint · Victoria, British Columbia
Working Papers on
Pacific Sovereignty & Bilateral Architecture
Defensive options for middle powers in a period of dollar-system stress.
The North Pacific Strategy Initiative publishes reference-grade working papers on the financial, diplomatic, and infrastructure architecture available to middle powers seeking to diversify counterparty risk. The Initiative is editorially independent: it accepts no outside funding and no paid engagements, and no government, institution, or commercial entity directs the work. It is written and funded by one person; the editor’s outside interests are declared in full at /disclosure/.
Working papers are placed publicly under version control to enable transparent review by allied policy and academic communities. Named commentary is welcomed and may be incorporated into subsequent versions with attribution.
Previous Working Papers
No. 9 · v1.0 · 27 July 2026 · For Discussion
The Counterparty Problem
What a $15 billion Memorandum of Understanding is actually worth when the counterparty’s charter expires in five months.
“U.S. reliability” is not one question but two, and conflating them produces bad risk analysis. The United States has a real, documented record of formally exiting binding treaties — roughly fifteen significant withdrawals or repudiations since 2001 — but no honest failure-to-convert base rate exists for non-binding instruments like MoUs, because no registry of U.S. MoUs exists. The paper prices the 24 July 2026 PIF–EXIM $15 billion framework against the sharpest case study available: EXIM’s 2015–2019 board-quorum lapse, which cut the bank’s financing by almost 85 per cent, and a charter that expires again on 31 December 2026 with reauthorization still in committee. Released 27 July 2026.
No. 10 · Technical Series · v1.0 · 26 July 2026 · Superseded by No. 11
Fair Use for We, IP Theft for Thee
Nobody agrees what “open source” means in artificial intelligence. Washington is now writing policy inside that vacuum.
Between 16 and 24 July 2026 the United States government came closer to restricting foreign open-weight AI models than at any point in the technology’s existence — and not one of the statements driving that policy rests on an agreed definition of the thing being restricted. The paper reads the licences: of eleven major model families surveyed, three satisfy the OSI definition, and none operates at frontier scale. Every load-bearing claim carries an evidence grade. Released 26 July 2026; the PDF is the canonical v1.0 release. Superseded by Working Paper No. 11, which itemises four corrections; this paper remains available unaltered with the correction notice attached.
No. 7 · v1.0 · 12 July 2026 · For Discussion
Dazzle 2.0
Russia's “zebra” trucks and the contest between a can of paint and a machine that thinks it can see.
What the viral claim gets right, what it inflates, and what it means for how the West buys autonomy. Russian logistics trucks in Ukraine have worn high-contrast dazzle paint since late May 2026, and the anti-machine-vision intent is accepted by serious analysts — but no controlled test shows the tactic defeating any named detector. The durable finding is cost asymmetry: a paint can iterates in an afternoon; a fielded detector re-learns only as fast as its slowest update loop. The paper separates confirmed fact from inference and hype, and draws the procurement lesson for NATO and Canada: vision-only terminal guidance is brittle; sensor fusion, human authorization, and edge-retrainable models are the robust answers. First paper in the NPSI Counter-Autonomy series. Released 12 July 2026.
No. 5 · v1.0 · May 2026 · For Discussion
Sovereign Compute North
An energy-for-memory swap, signed under Canadian law.
A sovereignty proposal for the Korea–Canada–allies decade. Canada has the cheapest low-carbon firm electricity in the G7; Korea controls the global memory chokepoint on which every modern AI accelerator depends; and the United States has stated under oath that it cannot guarantee allied citizens' data against its own extraterritorial reach. The paper fuses these facts into a single financeable architecture: a government-owned, allies-financed, Canadian-law-bound compute fabric — extended down to every Canadian municipality. Co-issued with Fit For Gov. Companion to Special Briefing No. 1, Zero Secrets. Released 27 May 2026.
No. 4 · v1.0 · May 2026 · For Discussion
The Addition Paradox
An energy thesis for Canada.
Why the energy transition did not happen, why that is acceptable, and why Canada is uniquely positioned to win the forty years that follow. A record year for clean-energy deployment was also a record year for fossil combustion: the global system did not transition, it expanded. The paper documents that expansion, names the chokepoint migration from molecule concentration to mineral concentration, and argues that Canada — alone among the G7 — holds simultaneously the cleanest barrel, the largest per-capita clean electricity, the highest-grade uranium reserves, and Pacific-and-Atlantic deepwater export capacity. The strategic preface to the three-rail architecture. Released 15 May 2026.
No. 3 · v1.0 · May 2026 · For Discussion
A Canada–Korea Pacific Defence-Industrial Corridor
Submarine procurement, industrial offsets, and the third rail of middle-power sovereignty.
A bilateral defence-industrial architecture pairing the Canadian Patrol Submarine Project with the Hanwha–Algoma Steel and Hanwha–APMA industrial-offset structures, anchored to the CKFTA institutional foundation, the October 2025 SDCP, and the 25 February 2026 Foreign and Defence (2+2) Ministerial Joint Statement, and integrated with the critical-minerals corridor that supplies the Korean HBM oligopoly powering the AI processor stack. Released May 2026, ahead of the 23 May 2026 ROK Navy operational demonstration at CFB Esquimalt and the June 2026 CPSP final-contractor decision. Completes the three-rail Pacific architecture (financial · energy & compute · defence-industrial) developed across NPSI Working Papers Nos. 1, 2, and 3.
No. 2 · v1.0 · May 2026 · For Discussion
A Canada–United States Energy and Compute Compact
Cooperative leverage in the AI-era continental grid.
A treaty-grade bilateral architecture pairing Canadian dispatchable generation with the U.S. grid through ultra-high-voltage transmission, siting AI training workloads on Canadian hydro-electric capacity, and preserving inference workloads at the U.S. urban edge. Structurally complementary to Working Paper No. 1; financed through the same supranational vehicle. Indigenous co-ownership designed in at issuance. Released May 2026, ahead of the CUSMA joint review of 1 July 2026.
No. 1 · v1.0 · May 2026 · For Discussion
The Bilateral Foundation
A thirty-year Canada–Korea sovereign bond as the architecture of middle-power independence.
Why Canada has the wrong number of friends, why the Bank of Canada–Bank of Korea swap line nobody talks about is the most important asset on the books, and how a thirty-year Canada–Korea co-issued sovereign bond — assembled from off-the-shelf instruments, priced as a spread over US Treasuries, with Indigenous co-ownership embedded at issuance — becomes the architecture of middle-power independence. The financial rail of the three-rail Pacific architecture. Indicative programme size US$25–40 billion.
Special Briefings
SB No. 3 · Accountability Audit · v1.0 · 26 July 2026 · Unclassified — All Sources Open
The Ledger With One Entry
Venezuelan oil revenue under United States custody: what left, what came back, and the accounting that was never published.
Since January 2026 the American government has controlled the sale of Venezuelan crude and the disposition of its proceeds. An estimated $13 billion has been raised; Venezuela’s own transparency portal records a single $300 million receipt, and no audit has been published. The briefing verifies the factual record, prices the gap, explains the two traps in American law that made an offshore account the rational choice — a structure that defeats Venezuela’s creditors and the United States Congress alike — and benchmarks the arrangement against the three comparable oil-custody regimes of the past quarter-century. The most accountable of them still lost $8.8 billion; this one is the least accountable of the four. Every material claim carries a confidence rating. Released 26 July 2026.
SB No. 2 · v1.0 · July 2026 · Unclassified — All Sources Open
The Three Doors
Under U.S. defense law, Canada has been domestic since 1992.
A Canadian miner's map to American federal money. The Defense Production Act has defined domestic source to include Canada since 1992 — Canadian companies qualify for U.S. defense-industrial grants, loans, and loan guarantees on the same footing as a firm in Ohio. The briefing maps the three doors — DPA Title III, Energy Act §18743, and EXIM's Supply Chain Resiliency Initiative — and the navigation moat that stalls applicants. Ten-panel briefing. Released 2 July 2026.
SB No. 1 · v1.0 · June 2026 · Unclassified — All Sources Open
Zero Secrets
Canada's data and AI sovereignty crisis.
Single-issue strategic assessments published when an exposure demands attention outside the working-paper cycle. Canada's data already answers to American law — under the CLOUD Act and FISA 702, jurisdiction follows corporate ownership, not server location — and the new generation of military AI converts that exposure into a complete intelligence picture. An audit of the federal sovereign-AI agenda against the law it operates under, four international models of what sovereign actually looks like, and a three-stage response. Companion to Working Paper No. 5, Sovereign Compute North. Released 11 June 2026.