FPT and ZEBOX target logistics AI deployment

FPT and ZEBOX have formed a three-year partnership to develop and deploy artificial intelligence, automation, and digital transformation projects across transport and logistics, combining FPT’s technology capability with an innovation network closely connected to the CMA CGM ecosystem. Shipping, port operations, supply chain management, logistics, and clean energy sit within the programme, which is intended to move selected ideas beyond demonstration and into operating environments where performance can be measured against defined commercial problems.

An early test comes through Challenge for Impact Vietnam, which has been running through 2026 and reaches its live final in Ho Chi Minh City on 15 October. Start-ups selected through the programme are working around operational challenges identified with CMA CGM Vietnam and other partners, including systems operation and terminal optimisation, with winning projects intended to progress towards funded proofs of concept after the final. That timetable gives the collaboration a defined conversion path from selection to deployment testing rather than another open-ended innovation programme built around the number of concepts it attracts.

Ports provide a particularly demanding environment for that model because apparently narrow optimisation problems sit inside a dense network of physical dependencies. Berth schedules, crane availability, yard density, gate flows, equipment positioning, container status, customs processes, labour, and late vessel arrivals all affect one another, so an algorithm that improves a planned sequence still has to survive the exceptions that dominate live terminal operations. A useful system must consequently connect with the software and operating responsibilities that control the movement of cargo, rather than producing recommendations that remain separate from the work they are supposed to improve.

Freight and supply chain applications face comparable integration difficulty, with relevant data spread across transport management systems, warehouse platforms, carrier portals, customer software, customs records, and locally maintained operational tools. Update cycles and identifiers often differ between those systems, while missing or inconsistent data can undermine an automated decision long before the quality of the underlying AI model becomes the limiting factor. FPT’s implementation capability gives the partnership resources to work on those less conspicuous interfaces, which are usually where a promising logistics pilot either becomes an operational product or stalls.

ZEBOX contributes access to start-ups, logistics operators, and defined industry problems, giving the programme a mechanism for matching technology with users that have a reason to test it. Over the three-year term, the partners intend to identify opportunities within the wider ecosystem, build proofs of concept, and support international deployment where projects demonstrate sufficient value. Moving from one site to several will demand considerably more than repeating the original installation, since data structures, local regulations, operating rules, languages, equipment, and customer requirements can all change while the core product is expected to remain supportable.

That scaling problem places governance alongside technical performance. Transport and logistics platforms contain sensitive shipment, customer, routing, pricing, and operational data, while port and terminal systems can sit close to nationally important infrastructure, so access controls, cybersecurity, auditability, and responsibility for automated decisions have to travel with the product as it expands. A pilot can operate with close supervision and a small group of specialists; a production deployment needs defined controls for the routine cases, the exceptions, and the point at which a person takes over.

Clean energy within the partnership’s remit broadens the operational scope without changing that requirement for measurable performance. Equipment utilisation, terminal sequencing, charging, routing, and energy consumption can be optimised together when the underlying data is available, potentially reducing idle movements or allowing electrified equipment to work around power constraints without sacrificing throughput. Projects still have to compete with conventional capacity investments for capital, however, which means reduced energy use will carry more weight when it arrives alongside improved utilisation, lower operating cost, or greater resilience.

The challenge format also changes the procurement sequence by defining an operational problem before a full technology purchase is made. Several approaches can be tested against the same requirement, weak concepts can be discarded before a large integration programme begins, and selected suppliers have a clearer route towards a commercial deployment if the proof of concept performs. The model only earns its keep when pilots convert into stable operations, making the proportion of projects that survive implementation more revealing than the volume of start-ups entering the accelerator.

Challenge for Impact Vietnam gives the new FPT-ZEBOX partnership a near-term opportunity to demonstrate that conversion. The live final on 15 October is one step in a programme that already passed through applications and selection, while funded proof-of-concept work is expected to follow for winning projects before the end of 2026. Successful trials would then leave the harder question for the remaining partnership term: whether systems proven against one logistics problem can be standardised enough for international deployment without stripping out the local detail that made them useful in the first place.