Every few weeks we read of a founder raising, say, half a million dollars in seed capital and announcing they are going to transform shipping with AI. The pitch typically has a ship silhouette, a nod to the IMO, and a promise that the platform pays for itself. The product might even be real. But too often the boring infrastructure underneath is missing, because the founder cannot afford to build it.

I write from the investor seat, but I have sat in the seafarer’s, ship owner-manager’s, and charterer’s offices, and built maritime tech from those seats. Most of what gets pitched as the next AI wave looks plausible from a technologist’s chair but falls apart the moment operational reality lands on it.

AI has collapsed the cost and the expertise needed to build a software stack. Two founders with AI coding assistants can do what teams used to, and a half-million seed will buy a persuasive MVP for early adopters. I welcome that. Frugal AI on the engineering side is healthy. The trouble is the maritime business built around the product has not got cheaper. Data acquisition, integration cycles, regulatory engagement, security audits, the time it takes to land reference customers, none of those have moved. AI helps with the build cost. It helps far less with the infrastructure or governance cost, and there is a limit to how frugal anyone can and should be on those.

This is the gap between what founders believe customers will pay for and what capital will underwrite. Closing it is what separates credible maritime AI businesses from another wave of orphaned MVP platforms.

My counsel to founders: have answers to these six questions, because institutional capital will ask them.

First, data provenance. Whose data are you using, on what licence, with what residency, and what happens if your largest data partner walks. Most founders cannot answer the second half. The data is the operator’s data. It is captive feedstock for someone else, who can switch it off.

Second, model lineage. What are you training, on what dataset, with what auditable changes over time. If the answer is that you took a freely available model and fine-tuned it once on customer data from three years ago, you are not running an AI business. You are running a consultancy with an AI badge.

Third, an accountable decision system. Maritime has one. It is called the master, and the ISM Code is the template. The human stays responsible. AI is an input, not a substitute. If your product positions itself as removing the decision maker rather than supporting one, the underwriters will price your real exposure before the regulator ever sees it.

Fourth, cyber and information security posture. A proper programme with named accountabilities, a working incident response plan, and a story that stands up to a serious third-party audit. If your security narrative is one slide and a vague promise to get it sorted, your customers will pay the price.

Fifth, commercial terms that survive scale. Pricing that holds when a customer onboards twenty vessels. Implementation contracts that do not quietly become a services business. A gross margin path that does not rely on the founder doing the integration work herself.

Sixth, customers willing to be referenced. Not logos. Phone numbers. I will call them.

None of this is exotic. It is what every regulated industry already does, and shipping is the regulated industry that pretends it is not one until something goes wrong.

Founders are not the only ones with homework. Most maritime operators are already running more AI in their business than they account for. The routing vendor, the fuel optimisation platform, the CCTV provider, the bunker analyser, the insurer’s loss model, or even the ‘personal AI toolkit’ that their colleagues bring to work, ashore and at sea. Many vendors have machine learning built into the product, or quietly route data to one of the big AI providers in the background. Thetius found last year that 81% of maritime organisations were running AI pilots while only 11% had formal policies for scaling them, and the unaccounted AI sits on top of that. Every maritime board should ask its CIO to map the AI already inside the supply chain.

There is a Goldilocks zone here. AI that aids a human decision is insurable, defensible, and scalable. AI that operates without one is not, and pretending otherwise to win a sale will be untangled by your P&I club well before it is untangled by the IMO. There is a reason your car insurer gives you a discount for a black box but is not yet selling cover for a driverless taxi. Same gradient on ships, at higher stakes.

Both directions showed up in the past fortnight. A leading maritime data platform raised a billion-dollar growth round. A young AI inspection business merged into a larger fleet management group that already had what it lacked. The aggregator platforms and the point solutions they absorb will succeed with AI in maritime for the same reason. Both need to invest disproportionately in the foundations.

Twenty years in this seat has taught me that the businesses that scale are not the ones with the snazziest UI or the slickest pitch. They are the ones whose founders combine the know-how with the know-who, who have done the boring infrastructure work, and who have built a commercial model that does not need rescuing in year three.

If you are a founder, build the unglamorous infrastructure first. If you are an operator, find out what AI you are already running before you buy any more. If you are an investor, ask the boring questions early.

AI in maritime is not a get-rich-quick scheme. By and large, capital will follow readiness, not the pitch. Founders who do the work will still be scaling serious businesses in five years. The ones who hoped no one would ask the questions will not.

——-

Manish Singh is a third-generation seafarer, a ship manager and a marine technologist. He is CEO of Maris Investments, a maritime technology and data platform built in partnership with GTCR. He was previously CEO of Ocean Technologies Group and Group Managing Director of V.Group.

Photo and content by Manish Singh

———–
Click here to sign up for our free monthly ‘De-Mystifying Digitalisation and AI in Maritime’ newsletter and get the latest cross-sector expert insights, news, interviews, videos and case studies straight to your inbox.