Compounding pressure on Russia dominates the outlook: a federal deficit set to breach 10 trillion rubles collides with an advancing US sanctions bill, pre-election food-price interventions, cyber-espionage against EU sanctions networks, and continued grinding gains in Sumy-Kharkiv — even as the global AI buildout keeps accelerating through Alphabet's capex and looming US curbs on HBM memory exports to China.

The underlying base rate is genuinely high: APT29/APT28 collection against EU policy and economic-intelligence targets is near-constant, and CERT-EU/vendor attribution reporting is frequent, so 'some attributed campaign against sanctions/trade nodes within 90 days' is a soft bar. Both analysts converged (GPT 0.69, Claude 0.75) on espionage-over-disruption, which fits actor incentives — Moscow avoids Sandworm-style disruption that risks NATO escalation while it manages the confrontation. HOWEVER, I am applying a hard self-correction: my cybersecurity track record is FAILING (Brier 0.357, over by 32pp), and the cyber agent's own record is weak (Brier 0.36). What would make me wrong: (1) attribution lags beyond 90 days even if activity occurs; (2) reporting frames it generically rather than tying it to sanctions entities; (3) I am pattern-repeating my overprediction bias. I therefore compress from the ~0.75 base-rate signal down toward the Skeptic's 0.68 and settle at 0.66 — deliberately near the bottom of my defensible range for a high-base-rate event. Pillars: threat_actor_analysis, attack_surface.