Heerema and Amphitrite Prove Vessel-Specific AI Routing on the World's Largest Crane Vessel

Heerema Marine Contractors has cut 2.5 days off a North Atlantic transit of Sleipnir — the world's largest semi-submersible crane vessel — using a voyage-optimization pilot with French ocean-intelligence company Amphitrite, trimming 240 nautical miles of sailing distance and up to 18% of CO₂ emissions while holding higher average speeds without burning more fuel. How it works: Amphitrite built a machine-learning model of Sleipnir itself, trained on five years of the vessel's operational data, that predicts speed and power consumption under changing wind, wave and current conditions — vessel-specific performance modelling layered on high-resolution ocean intelligence, rather than generic weather routing. What it did in practice: outbound from Rotterdam to New York in punishing winter conditions, the system picked the optimal window to ride tidal currents through the English Channel, then recommended a more northerly line around a developing storm to catch favourable tailwinds. The return crossing was the clever part: instead of the shortest route, the model deliberately sent Sleipnir the long way — into the Gulf Stream — where favourable currents lifted speed over ground to 14.6 knots and beat the direct route on time despite the extra miles. Heerema is now using Amphitrite's Ocean Bulletin across its fleet for voyage planning and offshore project preparation. Why this matters for ship managers: transits are among the most energy-intensive phases of offshore operations, and this shows AI routing paying off on a complex, non-standard hull — not just container tonnage — with the vessel's own history doing the heavy lifting. The honest caveat: these are company-reported figures from a single pilot; watch whether the gains repeat across seasons and routes before treating them as a baseline. [Hellenic Shipping News · OE Digital · Marine Log]

This story is part of the Maritime AI Digest of 19 July 2026.