Cover: AI voyage optimisation and fuel savings checklist
AI in Maritime

AI for Voyage Optimisation and Fuel Savings: What the Tools Do and What to Check

Maritime.Studio Team November 16, 2026

What "AI voyage optimisation" actually means

When vendors talk about AI for voyage optimisation, they usually mean software that combines weather forecasts, vessel performance data and routing algorithms, some of it machine-learning based, to recommend a speed, trim or route that uses less fuel for a given arrival time. This sits alongside older, non-AI weather routing services that shipping companies have used for decades. The "AI" part is mostly in how the model learns from a vessel's own historical performance data rather than relying only on fixed hydrodynamic tables.

The problem these tools are built to solve

A large share of the inefficiency in commercial shipping comes from a well-documented pattern the industry calls "Sail Fast, Then Wait" (SFTW): a vessel sails at a speed set by the charter party, then waits at anchor because the berth or terminal is not ready. BIMCO submitted an information paper on this to the IMO's Marine Environment Protection Committee (MEPC 82/INF.32) on 26 July 2024, arguing that SFTW is a systemic inefficiency rather than a one-off problem, and proposing the Blue Visby Solution (BVS) as a technical and contractual fix: adjusting a vessel's speed so it arrives closer to when the berth is actually ready, instead of racing to anchor and waiting there.

What the numbers actually say, and why they do not agree

This is where it gets harder to state a single fact. Savings estimates for Blue Visby-style speed optimisation vary a lot depending on who is reporting them and how they measured it: a November 2025 Blue Visby presentation cites an average of around 16% fuel and CO2 savings per voyage; Bureau Veritas has put the figure at roughly 15%; NAPA, a technology partner, has cited prototype trial figures closer to 29%; and an earlier 2023 BIMCO/Nautilus Labs white paper estimated 15-20% emissions reductions if applied industry-wide. None of these should be read as one settled number. They come from different fleets, different routes and different baselines, so the honest summary is that double-digit percentage savings are plausible in the right trade, but the exact figure depends heavily on the voyage and the comparison used.

Why classification societies are getting involved in verification

Because vendor claims vary so widely, independent verification is becoming part of the conversation. ClassNK has joined a project with Evergreen Marine Corporation, Samsung Heavy Industries and Weathernews specifically "to objectively verify reductions in fuel consumption and greenhouse gas emissions" from AI-based voyage optimisation, using a real Evergreen commercial vessel fitted with Samsung's autonomous ship system and Weathernews weather data. ClassNK's role is to review the verification methodology and issue a Statement of Fact once the trial is complete. The project itself exists because the effectiveness of voyage optimisation technology is known to vary with weather and sea conditions, which is exactly why a standardised verification method is needed rather than taking a vendor's own number at face value.

A short checklist before trusting a savings claim

  • Ask whether the number comes from a simulation, a controlled trial, or ordinary commercial operation
  • Ask what the baseline voyage was compared against, and whether it was a representative route or a best-case example
  • Ask whether an independent party, such as a class society, an academic study, or a charterer's own data, has reviewed the methodology
  • Ask how weather and sea conditions were accounted for, since the same software can show very different savings on a calm crossing versus a monsoon-season passage
  • Check whether the saving is against fuel burn only, or also accounts for extra time at sea, demurrage, or waiting time, since a slower but steadier voyage can look efficient on fuel while creating other costs

Where this fits with existing IMO work

None of this exists in a vacuum. The IMO has long promoted Just-In-Time (JIT) arrival as a concept: timing a vessel's speed so it arrives when the berth is actually available, rather than racing in and anchoring. A 2023 study for the Swedish Lighthouse programme at Chalmers University looked specifically at how AI and ship data analytics could support IMO's JIT arrival guidance, reporting meaningful energy-saving potential from AI-assisted approaches, again without a single universal percentage. AI voyage optimisation tools are best understood as one way of implementing JIT arrival in practice, not a separate concept competing with it.

What this means for masters, operators and charterers

AI routing software is a recommendation engine, not a decision-maker. The master and the company's safety management system still carry responsibility for the actual speed and route chosen. For operators evaluating a vendor, the practical question is less "does AI work" and more "can this vendor show me verified numbers on a route like mine, and who checked them." For charterers and owners negotiating charter parties, BIMCO's own SFTW paper is a useful starting point, since it explicitly frames the commercial incentives, who pays for fuel, who benefits from waiting, as part of the problem, not just the technology.

Bottom line

AI voyage optimisation is a real and active area of maritime technology, grounded in a genuine inefficiency that BIMCO itself has flagged to the IMO. The tools can plausibly save meaningful fuel on the right routes, but no single, independently verified industry-wide percentage exists yet, which is exactly why class societies like ClassNK are now building formal verification projects around it. Treat any specific savings number as voyage-specific until you have seen how it was measured.

Talk to us

Maritime.Studio works with maritime companies on the digital and technical side of projects like this, from evaluating vendor claims to building the training that helps crews and shore teams use new tools correctly. See our Technical Consulting services or get in touch.