TRACE-DS Intelligence Report
Great-power war stays unlikely; Europe's eastern flank is already a lethal grey zone
Why read the 2026–27 elections as one story: they decide whether Europe meets Russian pressure on its eastern flank as one bloc or through partial coalitions. We put a great-power war before 2030 at 8% and a multi-theatre war at 3%, while lethal Russian strikes already reach the waters off Romania.
Executive summary
Fault Lines 2026–27 reads this cycle of elections from the top down: the spheres of the United States, Russia, China and the European Union, the alignments between them, three scored dimensions for 2030, the leaders who react to each result, and the signals on the ground. It is written for security, risk and intelligence teams, including GSOC operators, at organisations with staff, sites, suppliers or investments on Europe's eastern flank, with Romania as the lens.
The central judgement: a war between great powers stays very unlikely before 2030 (8%, and 3% for a war across more than one theatre), but the eastern flank is already a lethal grey zone. On 23 October a drone sank the cargo ship Khoryv about 13 nautical miles off Constanța; one sailor died and Romania's president described it as part of Russia's escalation. Romania accounts for 17 of the 35 dated Russian incursion events into NATO and EU territory we counted in 2026.
The report scores three questions separately. Who bargains with whom: no grand bargain is likely (61%), with a US–China accommodation at 19% and a US–Russia accommodation at 12%. Does Europe hold together: partial cohesion is the most likely outcome (54%), and paralysis rises from 9% to 22% if the Rassemblement National wins the French presidency in May 2027, which we put at 57%. Does fighting break out: an attributed lethal Russian strike inside NATO jurisdiction is three times as likely as a direct NATO–Russia military clash by end-2027 (18% against 6%) and twice as likely by 2030 (29% against 14%).
Every probability is a structured probability judgement with a definition, a deadline and a resolution source on the forecast board, built from audited base rates and a source-independence audit of the claims that drive the numbers. An operational annex separates physical exposure triggers from political escalation triggers, so that whether a forecast resolves never sets a safety threshold. The report covers 486 sources and comes in English and Romanian.
Key findings
- 1
A great-power war before 2030 is very unlikely: 8%, and 3% for a war across more than one theatre.
Rivalry is more heavily armed but restrained at the top: summits, a trade truce and a paused Taiwan arms package show Washington and Beijing preferring to bargain.
- 2
An attributed lethal Russian strike inside NATO jurisdiction is three times as likely as a NATO–Russia military clash by end-2027 (18% against 6%), and twice as likely by 2030 (29% against 14%).
The nearer risk on the eastern flank is strikes on shipping and border areas, not a clash between armed forces; the Khoryv sinking on 23 October was just outside Romania's territorial sea.
- 3
Europe most likely acts through partial cohesion (54%); paralysis rises from 9% to 22% if the RN wins France in May 2027.
France's presidential election is the single largest hinge for whether Europe holds its line on sanctions and Ukraine.
- 4
No grand bargain between the great powers is the most likely 2030 outcome (61%).
A US–China accommodation (19%) or a US–Russia accommodation (12%) would redraw spheres over the heads of Europe, Taiwan and Ukraine; the window for either narrows after January 2029.
- 5
Romania is NATO's most-hit frontline state in 2026: 17 of 35 dated incursion events.
Organisations exposed to Constanța, Sulina and the Danube delta should set safety posture on location, strike frequency and casualties, not on official attribution.
Subjects covered
- US midterm elections 2026
- French presidential election 2027
- NATO eastern flank
- Black Sea security
- Romania
- Russia hybrid warfare
- US–China relations
- Taiwan
- European Union defence
- Analysis of Competing Hypotheses
The full report covers domain analysis, priority questions, recommendations, and source assessment. It is available below after a short form.