PROFESSION LOGISTIQUE N°10 – APRIL-MAY-JUNE 2026
The Appel d’aiR program, led by the AI Cargo Foundation and supported by energy savings certificates, has enabled the aggregation and analysis of logistics flows from hundreds of French companies on a neutral and confidential platform. From this unprecedented data collection emerged the project for a national rail loop with seven terminals, designed to offer 42 combined transport services in areas where the market currently offers no credible alternative to road transport. Bernard Guilbot, director of the Appel d’aiR program, describes its driving forces and ambitions.
At the heart of the Appel d’aiR program lies a simple yet fundamental observation: the data needed to identify the potential for shift ing freight from road to rail or river transport does not exist, or exists only in the form of broad regional statistics that are too aggregated to be operationally useful.
There are mountains of statistics… but they are, at the end of the day, ‘just’ statistics: for example, you’re told that X million tons leave the Lorraine region bound for the PACA region. However, you’re never told what’s in them, or how they’re transported. We only have a fairly general idea of the major fl ows moving from region to region,” summarizes Bernard Guilbot, director of the Appel d’aiR program and Honorary President of the AI Cargo Foundation. The lack of a consolidated view of actual logistics fl ows has, until now, prevented any meaningful consolidation—a condition that is nevertheless essential for a rail service to be economically viable.
It was in this context that the Appel d’aiR program was launched in 2022, eligible for the Energy Saving Certificates
(CEE) scheme and led by the AI Cargo Foundation, a non-profit association under the French Law of 1901 originating from the Hub France IA. Its objective: to establish a sovereign, neutral platform acting as a trusted third party between shippers who do not wish to disclose their commercial data. Set to conclude on December 31, 2025, the program has raised awareness among some 700 companies and actively engaged approximately 400 of them in the initiative. Among them are major manufacturers such as Michelin, Renault, and Danone, already heavy users of rail freight and seeking return loads to balance their trains and reduce costs.
CUMULUS: Data Driving Modal Shift
The program’s central tool is CUMULUS, an analytics platform that allows a shipper to upload their transport flow file—a shipment history or transport plan—and automatically identify routes that can be shifted to rail or river transport, with a corresponding carbon footprint for each. In practice, a shipper enters the departure point, destination, date, and tonnage for each shipment. The tool then identifies the nearest stations with existing services, calculates the pre- and post-transportation road legs, and lists the rail operators to contact. A shipper who has entered up to 650,000 lines into the tool illustrates the system’s industrial capacity. But the real innovation lies in pooling: when no service exists on a given route, CUMULUS aggregates the flows from multiple shippers to assess whether the critical mass would allow for the creation of one. In doing so, it preserves the anonymity of each participant: the flows are traced back to the departure and arrival stations, erasing all commercial information regarding the actual origins and destinations. Bernard Guilbot illustrates the process as follows: “We simply
display the departure station as Poitiers and the arrival station as Colmar. And we ask the shippers if they agree
to participate in a working group with rail operators to create this train?”
The question of competing shippers sharing the same train inevitably arises. Bernard Guilbot’s response is pragmatic: while some players are reluctant to share a train with their main competitor, others see no issue with it, arguing that competition centers on the product, price, or brand reputation—certainly not on the mode of transport. Large corporations, in particular, quickly grasped the value of the system in solving a well-known problem in the rail industry: according to Fret SNCF, 85% of trains return empty. Finding return loads allows costs to be cut in half—a lever far more powerful than any rate negotiation.
A unique national loop
The Lotus project was born from this logic of consolidating and balancing traffic flows. The idea, which Bernard
Guilbot has been championing since 2010, is to design a “combined metro”-style rail service, organized as a loop and operated in both clockwise and counterclockwise directions with daily departures. CUMULUS confirmed that the volumes, frequency, and regularity necessary for the viability of such a service were indeed present in the collected data.
The selected loop connects seven terminals: Paris, Dijon, Lyon, Miramas (near Aix-en-Provence), Toulouse, Bordeaux, and Tours. It enables the offering of 42 distinct services, several of which operate on routes currently completely absent from the market. Bernard Guilbot specifically mentions Dijon-Tours, Lyon-Toulouse, and Paris-Lyon (“Between Paris and Lyon, there are very few services, even though there is considerable volume,” he notes). The loop structure is precisely what makes these routes viable: regions that, taken individually, do not have sufficient volume benefit from transit traffic along their routes, creating the essential massification effect. In terms of volume, the CUMULUS analysis identified— after filtering out flows that are too seasonal or irregular—the equivalent of 109,000 trucks contributing to the loop, or 50 million ton-kilometers. The project’s goal is to shift one-third of these fl ows, which would represent approximately 36,000 trucks removed from the roads.
The carbon footprint is particularly favorable: rail freight emits 7 grams of CO₂ per ton-kilometer, compared to 86 grams for road transport—a ratio of approximately 1 to 9—as confirmed by data from ADEME and several sector-specific sources. Based on this ratio, shifting one-third of the identifi ed freight fl ows would result in savings of several million tons of CO₂.
Infrastructure Constraints
However, the project’s operational viability faces several constraints related to the condition of the French rail
network. The fi rst relates to the intensive renovation program carried out by SNCF Réseau, whose nighttime work primarily impacts freight: freight trains, 90% of which operate at night, are the first to be affected by construction disruptions, as passenger traffic is, by nature, given priority.
The future construction of the Bordeaux-Toulouse highspeed line will impose an additional constraint: during the construction period, estimated to last several years, freight trains will no longer be able to travel along this route, neither toward Toulouse via the west nor toward Bordeaux via the south.
The second obstacle is technical: combined transport relies primarily on swap bodies or semi-trailers loaded onto flatcars. However, standard P400-type semi-trailers (whose name refers to a height of 400 centimeters), favored by road hauliers for their versatility, cannot travel on the Atlantic route due to insufficient clearance. Bernard Guilbot emphasizes this point: bringing the entire network up to P400 gauge, particularly the Atlantic corridor up to Bayonne, is a key requirement for the competitiveness of combined transport. Freight coming from Spain bound for Poland is currently forced to travel all the way back to Paris to access a compatible network.
A third expected driver of improvement: the deployment of ERTMS (European Rail Traffic Management System), an on-board digital signaling system that improves train scheduling. Its large-scale deployment would increase network capacity by approximately 30%. Added to this is the issue of train length and tonnage: most lines are limited to 1,800 tons, while a few specific corridors allow trains of up to 3,000 tons. Increasing this capacity would reduce the cost per ton transported without increasing the number of train paths.
On the issue of electrification, Bernard Guilbot puts things into perspective: the main lines are well electrified, but terminal services to industrial sidings or combined transport hubs often remain diesel-powered. The introduction of dual-mode diesel-electric locomotives, however, provides a partial solution by eliminating the need to change locomotives during transitions between the electrified network and local services.
Toward a phased rollout
Faced with these constraints, the Lotus project team is adopting a phased strategy. The process for requesting train paths—the time slots allocated by SNCF Réseau to freight trains—requires a 15-month lead time between submission and actual use. This timeline, combined with uncertainties related to construction work on the Bordeaux Toulouse line, necessitates separating the testing phase from the full-scale launch of the complete loop.
An initial series of trials could be launched as early as before summer 2026, building on existing rail services but operated in line with the loop’s logic: balanced flows. The objective of these tests is twofold: to familiarize shippers who have never used combined transport and its practical constraints, and to test the responsiveness of recipients—particularly in the retail sector—to delivery times that are slightly different from those of road transport.
Mass retail is precisely one of the project’s main logistical challenges. Its requirements regarding delivery times—
oft en very early in the morning—are difficult to reconcile with the unloading times of a train arriving at the station at 6 a.m., redistributed by truck, and delivered no earlier than 9 a.m. This is why the project involves Club Déméter, which brings together manufacturers and mass retailers, in an eff ort to ease these usual constraints: shifting from 24-hour delivery to 48-hour delivery, and adjusting delivery windows. “If we don’t get them involved or if they refuse to participate, some may find themselves caught in a never-ending cycle, with penalties as the result,” summarizes Mr. Guilbot.
A European Dimension
Beyond the French market, the Appel d’aiR program aims to become a benchmark on a European scale. CUMULUS currently lists 2,896 combined transport services operating across Europe, all integrated and updated in its database. This continental coverage (unique, according to Bernard Guilbot) allows shippers engaged in import-export operations to test their logistics flows over a much broader scope than just the national territory.
By way of comparison, the Routescanner platform, developed by the Port of Rotterdam, offers a feature similar to CUMULUS’s primary use case—namely, identifying an existing service on a route—but does not provide mass analysis of flow files, consolidation logic between shippers, or the design of prospective services.
This differentiated positioning constitutes the main commercial value-added that the AI Cargo Foundation intends to promote in the coming months, relying on foreign consulting fi rms to roll out the solution across several European countries.
Thanks to the AI Cargo project, the French logistics sector is gaining a prototype of what a logistics data infrastructure of general interest could look like: neutral, sovereign, and capable of making visible flows that the fragmentation of players had previously condemned to invisibility.
