€1 Crous meals: university restaurants overwhelmed, how to avoid shortages and food waste
Since France opened its €1 Crous meal to all students, university restaurants have been serving 48% more meals than a year ago, when the government had planned for 10%. Queues, dishes running out, staff under pressure: beyond the overcrowding, the reform raises a question for the entire food service sector. How do you produce the right amount when demand changes overnight?
In Lille, some university restaurants served 700 meals every lunchtime before the reform. They now serve almost 1,000. On 24 September, the Cnous, the national body that oversees France's student services network (the Crous), asked each Crous to set a “maximum sustainable capacity” for its food outlets, even if that means turning students away beyond it. A first for a measure that is putting the Crous' logistics under strain.
The €1 Crous meal: a measure whose cost was underestimated
Opening the €1 meal to all students was expected to cost €90 million a year. It will require €143 million in 2027, because attendance has grown five times faster than forecast.
Since 4 May 2026, the full €1 meal is no longer reserved for grant holders and students in financial hardship. The number of eligible students rises from fewer than 700,000 to more than 3 million. Students without a grant previously paid €3.30, while a meal costs between €8 and €9 to produce according to figures cited in the French National Assembly, with the State covering the difference.
The network was already under strain. It served 42.8 million meals in 2023-2024, then 44.2 million in 2025, in around 940 university restaurants across 186 cities. To absorb this continuous growth, 3,157 seats were added at the start of the 2024 academic year and 2,613 in 2025.
What does it cost each year? The measure had been estimated at €90 million, based on a 10% rise in attendance. Reality far exceeded that scenario: up 22% in the spring according to the Cnous, then up 48% since the start of the academic year compared with September 2025, i.e. 1.3 million additional meals. The Minister for Higher Education announced that funding would rise to €143 million in the 2027 draft budget, warning that if the budget were not adopted, the government could no longer guarantee the measure for all students.
Why are university restaurants overcrowded?
University restaurants are overcrowded because attendance has jumped by 48% without a matching increase in seats, staff and production capacity.
The national figure of 48% is an average. It combines restaurants that are full from 11:45 am with others where attendance has barely changed. The increase depends on how close a restaurant is to lecture halls, on timetables, on competing food options nearby, and on whether it opens in the evening. In Paris, only one university restaurant remains open in the evening, which concentrates demand on a single site.
The operational consequences are visible. The Lille Crous has committed €1 million, ten additional staff and new seats. Staff took action on 22 September to denounce a pace of work that has become unsustainable. Students, for their part, report missing meals because they ran out of time or because dishes were no longer available at the end of service.
Contacted by Jabu at the end of September, one Crous told us its teams would remain under heavy pressure until the All Saints' holidays in late October:
“The €1 meal for all students is keeping us very busy at the moment.” A Crous, September 2026
These pressures have one thing in common: the volumes to be produced are decided several days, sometimes several weeks, in advance, based on historical data that the reform has made obsolete. A restaurant that orders its supplies based on last year's attendance is now structurally wrong, one way or the other.
Shortages and food waste: the double risk facing the Crous
The same forecasting error produces either a shortage or food waste. Across the Crous network, each percentage point of overproduction costs around €1.3 million a year.
Faced with uncertainty, a food service manager has only two options. Produce generously so that no one is turned away, at the cost of dishes that are prepared but never served. Or cut production sharply to stay within budget, at the risk of running out at the end of service. Either way, the cause is the same: a forecast that no longer matches the reality of the site and the day.
University catering now faces both risks at once. Overwhelmed sites run out of food, while stable sites that raised their orders as a precaution produce surpluses. Across a network, the two errors do not cancel each other out: a missing tray in Lille does not make up for a tray of food thrown away in Toulouse. They add up.
The financial stakes are easy to measure. Based on the 44.2 million meals served in 2025, each percentage point of overproduction represents about 442,000 portions prepared and not served. Using a conservative value of €3 per portion, well below the full production cost of €8 to €9, one percentage point of error costs around €1.3 million a year. With attendance up by almost half, this figure is a floor.
Field measurements confirm the order of magnitude. In 2024, the management of the Crous Nord-Pas-de-Calais estimated food waste losses at around €29,000 per restaurant per year. The weighing systems installed in several university restaurants mainly measure what is left on trays. Dishes that are prepared but never served depend on another lever, further upstream: forecasting.
The cost of shortages is less visible but politically heavier. A student who leaves without a meal is a public service not delivered, and a cost-of-living measure that misses its target. Emergency restocking, substitute menus and overtime in turn push up the cost of a meal.
How predictive AI helps forecast attendance in university restaurants
Predictive AI estimates, for each restaurant and each service, the number of meals that will actually be served, up to thirty days in advance. Orders and production can then follow real demand rather than past data.
When demand shifts, historical data alone is no longer enough. You need a forecast that learns fast, site by site, and that tells lunch from dinner, Tuesday from Friday, exam week from holiday week. This is Jabu's core business: forecasting attendance for every restaurant and every service, up to thirty days in advance, so that orders and production follow real demand.
Jabu's model combines each site's till data with the factors that drive attendance:
the academic calendar (start of term, exams, holidays)
public holidays and long weekends
the weather
the menus on offer
local events
It recalibrates continuously as new data comes in. A break like the one on 4 May is therefore absorbed within a few weeks rather than a year. Nothing needs to be installed in the institution's systems: data is exchanged via API or secure file transfer, and hosted in France.
Results measured at clients with profiles similar to university catering, i.e. high-traffic sites with variable attendance, give an order of magnitude:
Network of government staff restaurants (5 sites, around 2,000 meals a day)
Overproduction reduced by 60%
Underproduction stable
Corporate restaurant (around 1,700 meals a day)
Overproduction reduced by 47%
Underproduction divided by 3
Method: gap between the quantities forecast and the meals actually served, measured restaurant by restaurant, day by day and service by service, with no offsetting between sites or between services. In both cases, the gains appeared within the first ten days of service.
The second case illustrates an often overlooked point: forecasting well is not only about producing less. On this site, forecasting reduced both surpluses and shortages, because it shifted volumes to the days and services where demand actually was.
What about your restaurants? Our team will show you how Jabu forecasts attendance, site by site and service by service, and what it would change for your orders.
2027 budget: why capping access will not be enough
Capping access protects staff, but it does not solve the underlying equation. The 2027 budget allocates €143 million to the measure. Every point of accuracy gained on production frees up part of it to serve more students, rather than paying for dishes that are never eaten. With public finances under pressure, it is also an argument for the measure's long-term sustainability.
The €1 meal is a wake-up call for the entire food service sector. School canteens after a pricing reform, corporate restaurants facing remote work, hospitals and care homes with fluctuating headcounts: everywhere, the same rule applies. Changing demand requires a forecast that can keep up.
Frequently asked questions about the €1 Crous meal
Who is eligible for the €1 Crous meal?
Since 4 May 2026, all student card holders, whether or not they receive a grant, as well as apprentices and work-study students, PhD students and civic service volunteers. Payment must be made with an active Izly account.
How many €1 meals can a student have per day?
One €1 meal per service. A student can therefore have two a day, at lunch and dinner, in university restaurants that offer both services.
What does a €1 meal include?
A main dish and up to two sides: starter, cheese, dessert or fruit. Bread and water are included. Other drinks are not, and extras are charged depending on the restaurant.
Why are university restaurants overcrowded?
Attendance has risen by 48% since the start of the academic year compared with September 2025, while the government expected 10%. Seats, staff and production capacity have not kept pace.
Can a university restaurant turn students away?
It is now possible. On 24 September 2026, the Cnous asked each Crous to set a “maximum sustainable capacity” for its food outlets, even if that means refusing entry beyond it.
How much does the €1 meal cost the State?
The measure had been costed at €90 million a year, based on a 10% rise in attendance. The 2027 draft budget allocates €143 million. A meal costs between €8 and €9 to produce, and the State covers the difference with the price paid by the student.
How can shortages and food waste be avoided when attendance changes?
By forecasting attendance restaurant by restaurant and service by service, rather than relying on last year's figures. That is what Jabu's AI does, up to thirty days in advance: discover the Jabu solution.
Is attendance at your restaurant hard to predict? Jabu analyses your historical attendance data free of charge and shows you, within a week, the gap between your forecasts and the meals actually served.
About Jabu
Jabu develops artificial intelligence that forecasts attendance for the food service sector: school and university restaurants, healthcare facilities, government staff restaurants and corporate restaurants.
Winner of the 2026 Innovative Eco-Company Trophy (PEXE and ADEME), Waste and Circular Economy category
GreenTech Innovation label from the French Ministry of Ecological Transition
Software publisher listed with UGAP, the French public procurement agency
Contributor to the Agenda 2030 working groups against food waste, presented to the French National Assembly in February 2026
Sources (in French)
Service Public, French government portal: the €1 meal open to all students
French National Assembly: average production cost of a Crous meal
Wikipedia (FR): university restaurants and the network's outlets
France 3 Hauts-de-France: a university restaurant cuts food waste with AI (2025)
Jabu client results: internal measurements, September 2026
