The alert arrives at 19:40 the evening before. Sud-Rail has extended its notice, the RER B will run a reduced service from the first train of the day, and the flight is still scheduled into Roissy at 08:15. In a group chat somewhere, three colleagues booked on it are arguing about whether to sort out the transfer now or deal with it at arrivals.
Most of the airport transfer advice they will find online describes a Paris that stopped existing this year. The bus everybody recommends as the strike-proof fallback made its last run in February. The taxi fares quoted in those guides were superseded on 1 February. And the comforting line about forty-eight hours of warning describes a legal mechanism that, this summer, has been producing notices measured in weeks.
The bus that used to be the answer stopped running in February
RoissyBus ran for the last time on 28 February 2026. Île-de-France Mobilités withdrew it after ridership fell roughly 40 percent from 2023 and its route by way of the périphérique made timekeeping close to fictional. Line 9517 replaced it the next morning, and it is not the same product. It runs between Argenteuil, Saint-Denis and Charles de Gaulle, every fifteen minutes at peak and every thirty off peak, from 05:20 to 00:30, on biomethane coaches, at €2.05 a journey or nothing extra on a Navigo pass.
Read that route again. It does not reach central Paris. The way into the city is Metro line 14 at Saint-Denis Pleyel, which cuts the strike calculation both ways at once. Line 14 is driverless, so it keeps running through most actions that paralyse a network built around drivers, and very few travellers know they now have that card in hand. It is also an interchange, with luggage, at a station that a platform-staff action can close. Interchanges are precisely where a disrupted morning falls apart.
The fare frame moved as well. Since 1 January 2025 a single Ticket Paris Région Aéroports covers both airports at €14, or €7 reduced, loaded onto a Navigo Easy card or bought in the Île-de-France Mobilités app rather than printed. An ordinary metro ticket is not valid to either airport, and an inspection produces a fine. It stays the most expensive small mistake made at the Gare du Nord barriers.
Three different disputes, one word
Travellers say strike and picture a train. Paris spent the first half of 2026 demonstrating how little that word narrows down.
On 10 June, a national SNCF strike cancelled about one TGV in three and six Eurostar services out of the Gare du Nord, which turned a rail problem into a road problem for anyone whose Paris trip begins at St Pancras. That chain, and what it does to a same-day connection, is mapped in our guide to Gare du Nord transfers for Eurostar passengers.
On 18 June, a joint front of CGT, CFDT, UNSA and SUD Aérien called a 24-hour action at Roissy, Orly and Le Bourget over access badges to secure zones. No train was involved. The effect landed on queues and boarding times, which is to say on the half of the journey that happens after the car has already delivered you.
Over the Ascension weekend in May, an air traffic control walkout cancelled up to 75 percent of departures at Orly and 65 percent at Roissy. No ground transfer solves that one. Worth knowing that controllers work to a five-day notice rather than the surface-transport rules, which is why an August dispute can appear with almost no runway.
The useful distinction is not between big strikes and small ones. It is between the parts of the journey you can decide and the parts you can only absorb. The link between the city and the terminal is the one part still under your control, and it happens to be the part most people leave to chance.
The RER B is the link that breaks
The RER B is run by two operators on one line. RATP holds the section south of Châtelet, SNCF Transilien holds the northern half including the airport branch. Two companies, two sets of labour disputes, two calendars, and a single track between you and the terminal. That structure is why the line shows up in almost every disrupted week in Île-de-France, whichever house is actually striking.
Sud-Rail has kept a rolling, deliberately localised notice live into early September 2026, and the Paris to Roissy section is named as the exposed point. Rolling is the word that matters. There is no date circled on a calendar, no clean forty-eight hour countdown, and no way to plan a fortnight ahead around an announcement that has already been made and simply has not resolved into anything yet. On a bad morning the CDG branch drops toward one train an hour against four to six normally, the platforms at Aéroport Charles de Gaulle 2 fill with several hours of accumulated passengers, and the trains that do arrive are already full when they get there.
None of this is dramatic if you are travelling light with an afternoon appointment. It is a different proposition at 07:00 with two cases and a 10:00 meeting in the 8th.
What each mode actually costs on a strike morning
| Mode | Normal day | Strike day | What actually moves |
|---|---|---|---|
| RER B, Ticket Paris Région Aéroports | €14 | €14 | Frequency and crowding, never the fare |
| Line 9517 plus Metro 14 | €2.05 | €2.05 | One interchange, and whether it stays open |
| Paris taxi, regulated flat fare | €56 Right Bank, €65 Left Bank | Identical | The rank queue, not the price |
| Ride-hailing app | Market rate | Multiplied | Price first, then whether a car accepts at all |
| Pre-booked private transfer | From €105 | From €105 | The road, not the availability |
The taxi line is the one that surprises people, and it deserves saying plainly. Paris airport flat fares are set by prefectoral order, and the current ones come from the order of 24 December 2025 that took effect on 1 February 2026: €56 from Charles de Gaulle to any Right Bank address, €65 to the Left Bank, €45 and €36 respectively from Orly. They apply in both directions, at any hour, with no night, Sunday or holiday supplement and no allowance for traffic. A strike cannot move that number. Legal extras exist and are modest, €4 for an immediate booking, €7 for one made in advance, €5.50 per passenger from the fifth onward. If price is the only thing being optimised, the regulated flat fare is the rational answer on a strike day and we will say so without flinching.
What a strike does to a taxi is queue it. Thousands of rail passengers reach the same conclusion within the same hour and walk to the same rank, and a fare that cannot rise has no mechanism for rationing demand. The wait absorbs what the price is forbidden to. Ride-hailing does the opposite and rations by price, which is why the same trip that sat comfortably under a hundred euros on Monday reads like a mistake on Thursday morning. Neither behaviour is scandalous. They are simply two different ways of running out of cars, and both are worth understanding before you need one. The full picture of what Paris ground transport costs across normal conditions is set out in our 2026 pricing guide.
A booked private transfer sits outside that mechanism for one unglamorous reason. At PrivateDrive the vehicle is assigned to your name before the morning begins, so there is nothing left for demand to reallocate. The fare was agreed at booking and does not reopen. What still moves is the road: elevated car traffic on a strike day is real, and a journey that normally takes forty-five minutes from the 8th to Roissy can want an hour.
What a strike notice tells you, and what it stopped telling you
The framework everyone half remembers comes from the law of 21 August 2007 on social dialogue and continuity of public service in land transport. Unions file at least five days ahead. Each employee who intends to join declares individually forty-eight hours in advance, so the operator can publish a service plan the evening before. That is the machinery behind the familiar promise that you will know two days out.
It still works, as far as it goes. What has drifted is what the notice actually contains. A rolling notice covering several weeks satisfies the legal requirement long before anyone can say which depots will be affected on which morning, and the summer of 2026 has been built almost entirely out of that kind of filing. The document exists. The information does not.
Parliament came close to changing it. A bill adopted by the Senate on 9 April 2024 passed the National Assembly's sustainable development committee on 14 January 2026 and reached the floor on 22 January. It would have capped notices at thirty days to kill off dormant filings, moved individual declaration from forty-eight to seventy-two hours, required strikers to join from the start of their shift rather than mid-service, and allowed the right to strike to be suspended during three morning and three evening peak hours for up to thirty days a year. Its own rapporteur put the odds of adoption at close to zero before the debate opened, and it did not become law. Anyone planning around a seventy-two hour warning is planning around a text that does not exist.
The practical consequence is a change of habit rather than of tool. The notice tells you a dispute is live. The service plan published the evening before tells you what will actually run. Only the second one is worth reorganising a morning around.
Departures and arrivals are not the same problem
Arriving late is expensive in the ordinary sense. You lose a morning, you pay three times what you meant to, you walk into a meeting having already had a bad day. Painful, recoverable. The arithmetic of a disrupted arrival, and what actually still runs when the network thins out, is worked through in our late arrival transfer guide.
Departing late is a different category, because the deadline does not negotiate. A missed departure out of CDG costs a rebooking fee at best and a written-off business fare at worst, and on a strike morning the flight you missed is the one everybody else is also trying to rebook onto. Two habits handle it. Add twenty to thirty minutes to the pickup time on any confirmed strike day, since road traffic rises as displaced commuters take their cars. And decide early rather than well, which is the same discipline that governs a pre-dawn departure and is laid out in the 4am playbook.
For companies moving people through Paris weekly, the exposure stops being about any individual trip. It becomes a question of how many meetings a year begin twenty minutes late and what that costs in outcomes nobody itemises, which is the argument we made in full in the hidden cost of unreliable corporate transport. Booking capacity in advance is not caution on those accounts. It is the cheapest line in the travel budget.
One timing note that matters more than it should. Private capacity in Paris tightens within hours of a strike being confirmed, exactly as hotel inventory does around a trade fair. The evening before is late. The week before is free.
The structural fix is dated. CDG Express opens on 28 March 2027, twenty minutes from the Gare de l'Est to the airport, €25 a ticket, operated by Hello Paris. It will be fast, it will be clean, and it will be a train, which means the same legal frame, the same unions and the same exposure as the line it is meant to relieve. A second rail link is a capacity answer. It is not a redundancy answer.
That distinction is the whole subject. Paris moved 107 million air passengers across its two main platforms in 2025, 72 million through Roissy and 34.9 million through Orly, up 3.4 percent and back to 99 percent of 2019. The city keeps adding ways to reach the airports, and every one of them so far runs on rails and depends on a service being voted rather than a vehicle being assigned. Our 2026 market study tracks where that leaves private ground transport as volumes climb.
Which is why the sensible response to strike season is not vigilance. Watching the notices more closely produces anxiety and no additional trains. The response is to settle the airport transfer at booking, when it is cheap and abundant, and to spend strike mornings reading the departure board like everybody else who has nothing left to arrange. Book the CDG run the week you book the flight, not the evening the alert lands.