ATC Was Condemned In 2015 But It's Still Flying The Country

At around one o'clock this Tuesday afternoon, the flight data processing system at NATS' (National Air Traffic Services) Swanwick centre in Hampshire stopped working properly. Controllers did what they are trained to do when the machinery misbehaves. They cut the number of aircraft allowed into the sky.

Departures were held at Heathrow, Gatwick, Manchester, Stansted, and Edinburgh. Inbound aircraft went into holding patterns or diverted. Engineers applied a fix at about 4.40pm, which sounds like a swift recovery until you understand what four hours does to an aviation network running at capacity.

Aeroplanes ended the afternoon in the wrong countries. Crews ended it out of hours.

By Wednesday teatime the aviation tracking firm Cirium counted more than 2,000 flights cancelled or grounded across the two days. British Airways alone scrapped nearly 550 services affecting over 85,000 of its passengers. Ryanair reported 65,000 delayed and demanded the resignation of the NATS chief executive. Wizz Air pronounced the organisation unfit for purpose. Somewhere around 330,000 people had their journeys wrecked.

Yesterday the Transport Secretary, Heidi Alexander, stood at the despatch box and apologised wholeheartedly to passengers. She confirmed the fault sat at Swanwick; ruled out a cyber attack by Iran or Israel; ordered an independent review by the Civil Aviation Authority reporting within six months; gave NATS one week to explain itself; and delivered the sentence which will follow this episode around for years:

I do not believe this issue was unavoidable.

Then the House proceeded to the second reading of the Civil Aviation (Consumer Protection and Regulatory Reform) Bill.

The timing was nobody's fault. The contents are.

Parliament Saw This Coming

The Civil Aviation (Consumer Protection and Regulatory Reform) Bill arrived in the Commons as Bill 132 after the Lords finished with it in July, having been introduced there on 14th May 2026. Its long title promises provision for the protection of purchasers and users of air transport, provision about airspace change, air traffic and air navigation services, airport slots, a power for the CAA to make rules, and provision about aviation offences.

In other words, this nonsense has been known about for some time and did not arrive out of the blue this week:

A Bill to make provision for the protection of purchasers and users of air transport and airport services; to make provision about airspace change, air traffic and air navigation services and airport slots and schedules; to confer power on the Civil Aviation Authority to make rules; to make provision about aviation offences; and for connected purposes.

Clause 9 amends the Transport Act 2000 to alter the procedure for modifying air traffic services licence conditions and to adjust who must be consulted. Clause 10 changes who may be charged for air navigation services and permits different charging methods. Both are economic plumbing.

Clause 12 is the constitutionally "interesting" one.

It invents a new statutory creature called "CAA rules", letting the regulator make substantial aviation law which until now lived in Air Navigation Orders or in retained EU aviation legislation.

The CAA must normally consult the Secretary of State and anyone else it considers appropriate although a fast-track exists for urgent safety deficiencies, minor changes, and restatements. The Lords Delegated Powers and Regulatory Reform Committee counted seventeen delegated powers in a Bill of thirteen clauses, and described one of its purposes as delegating aviation safety and operations rulemaking to the CAA.

Peers noticed.

During committee stage the objection was put plainly: a body which is already the safety regulator, the economic regulator, and the airspace regulator was being handed the pen as well. The Lords extracted one concession, forcing certain restrictions on CAA rule-making through the affirmative procedure rather than the negative one.

The Commons Library briefing confirms the rest of the package: passenger rights, mandatory dispute resolution, accessibility, slots, airspace modernisation, direct CAA fining powers, and the migration of safety rulemaking to the regulator.

What the Bill does not contain is any requirement on the operator of Britain's national air traffic system to:

• Maintain diverse redundant flight processing capability to demonstrate a maximum tolerable outage;
• Hit a recovery time objective;
• Sustain a minimum throughput while degraded, or
• Submit its architecture to independent technical testing.

There was one moment when Parliament came close to touching the incentives. During the Lords stages, an amendment would have allowed airlines to recover passenger compensation costs where disruption was not their fault, explicitly including air traffic control failures.

It was defeated.

An obscure technical amendment on 13th July looks rather different on 10th September.

Resilience Was Already A Licence Condition

The tempting conclusion to arrive at after another idiotic British infrastructure failure is somebody forgot to write a rule. Not here. NATS En Route plc holds its licence under the Transport Act 2000, and Condition 2 requires it to secure its core services on a continuing basis and to develop and maintain the assets, personnel, and systems needed to do so.

Since 2018 the licence has gone further, requiring an actual Resilience Plan covering resilience, contingency, and business continuity, reviewed at least every twenty-four months, capable of independent review at the CAA's direction, and certified by the board.

The licence even defines resilience: the ability of assets, networks, people, and procedures to anticipate, prevent, absorb, and adapt to disruption and recover normal service rapidly.

Britain did not lack a resilience duty. She had a resilience duty, a resilience plan, a resilience definition, a board certificate, and a biennial review cycle.

On Tuesday it also had no aeroplanes.

A Backup Which Could Fail Safely

Three years ago in 2023, NATS' NERL flight plan processing system (FPRSA-R) failed catastrophically. The independent review of the August 2023 collapse, led by Jeff Halliwell and published by the CAA in November 2024, contains the detail which should have ended any complacency about British air traffic resilience.

On 28th August 2023 the primary flight plan processor met a flight plan it could not handle and stopped.

The secondary system took over, applied essentially the same logic to the same plan, and stopped roughly twenty seconds later. Automated processing collapsed from somewhere between 700 and 900 flight plans an hour to about sixty, done by hand.

Nobody was endangered. That is the point, and it is also the problem.

The contingency arrangement performed beautifully as a safety mechanism and delivered under a tenth of the country's operating capacity.

The panel found the fallback could not maintain flight operations without restrictions.

Worse, the review found the CAA's own trigger for scrutinising a new or modified NATS system had traditionally been safety. If a contingency mode was judged safe, its effect on airspace capacity did not necessarily prompt deeper regulatory examination. A system capable of taking the country from 800 flight plans an hour to sixty could satisfy the governing test, because the sixty were handled impeccably.

The panel also looked at how much technical oversight actually happened.

In 2023, NERL submitted 101 change notifications. The CAA audited thirty-eight of them.

The panel questioned whether the regulator had the specialist staff to understand systems of that complexity and recommended reviewing both headcount and the sampling method. The CAA's second progress update, published in December 2025, reports revised recruitment arrangements and then concedes challenges remain in recruiting specialist roles.

Ryanair's verdict this week: the primary system is unreliable and the backup useless.

Fifteen Years Past Its Replacement Date

The Halliwell review buried its most damaging finding in the language of programme management. The ageing FPRSA-R and NAS platform at the centre of the 2023 failure is due to be replaced by a system called “Information through European Collaboration” (iTEC). Deployment is currently targeted for 2030. The previous target was 2015. Earlier plans had envisaged replacement during the 2000s.

The iTEC programme is now scheduled to be deployed in 2030, which is a significant delay from its target of 2015, which itself represented a delay from earlier plans which aimed for a replacement system to go-live in the 2000s. Delays to the iTEC programme were highlighted in the Egis report commissioned by the CAA in 2023 to review NERL’s proposed forward capital investment programme in regulatory period NR23. Given the age and criticality of the existing system, a replacement is an important factor in reducing the likelihood of flight data processing problems in the future.

The review sets out the consequences without much comfort.

1. Delay increases technical complexity and instability.
2. Keeping the legacy estate alive becomes progressively harder.
3. The staff who understand it are expected to have retired by around 2030, which is the same year the replacement is supposed to arrive.

Then comes the sentence which ought to have generated a select committee inquiry on its own. NATS' own investigation described the complexity of its systems and the constant stream of upgrades as making the maintenance of an overall current system map "effectively impossible". NATS classified this as a "minor" finding. The independent panel disagreed and reclassified it as major.

In its Major Investigation Report into the incident, under the minor finding Mi2, NERL states that “The complexity of the system architecture across NERL - and its regular changes and upgrades - results in any attempt to maintain up-to-date overall system mapping becoming effectively impossible”. The Panel disagrees with the assessment of this issue as being minor and instead suggests this is a major finding.

Britain therefore runs its national airspace on a production system whose complete architecture nobody can reliably model; whose replacement has drifted by roughly fifteen years; and whose institutional memory is walking out of the door on pension terms.

£65 Million Damages, £1.8 Million Consequence

The most scandalous number in this entire affair is the gap between what the 2023 failure cost our country and what it cost the company responsible. The Department for Transport's own written statement put the total cost to industry and passengers at between £75 million and £100 million.

Airlines absorbed roughly £65 million of it.

NERL's regulatory penalty came to about £1.8 million, and even that was not a bespoke sanction for felling British aviation.

It fell out of the existing performance mechanism because delay targets were missed.

The panel compared this unfavourably with electricity network regulation, where operators face materially larger financial exposure when the lights go out. Virgin Atlantic, in its submission on future price controls, described airlines as the insurer of last resort, because the framework does not push consequential failure costs back to the operator.

The panel accordingly recommended strengthening the financial incentives on NERL to deliver a resilient service as well as a safe one, and ensuring business plan regulation covered resilience and consumer impact rather than safety and efficiency alone.

The CAA's December 2025 update records these as embedded into its operational regulatory oversight practices. Methodology and business plan guidance in 2026. Consultations in 2027. Final regulatory decision in 2028.

A catastrophe in 2023 produced a 2024 finding of inadequate financial incentives, which the regulator marked as substantially addressed in 2025, and which will be decided in 2028.

The system fell over again in 2026, roughly halfway through the paperwork.

Dividends Paid Out, Inside Rots Out

NATS is not a normal company and not a government department either: it is a Blairite public-private partnership. The Crown holds 49 per cent. A consortium of airlines holds 42. Employees hold the remainder.

It is a regulated monopoly supplying critical national infrastructure with private shareholders attached, which is a structure Britain has tried before with water. It tends to produce a familiar argument about where the money went.

NATS paid £175 million in dividends in 2025.

Its group accounts for the year ended 31st March 2025 record statutory profit before tax of £234.3 million.

Martin Rolfe, chief executive for eleven years, spent Wednesday morning on the radio noting NATS has invested well over a billion pounds in technology since part-privatisation and enjoys one of the best reputations in the world for safety and resilience.

A claim he made on the second day of the third significant failure in three years.

Nobody should assert dividends caused Tuesday's outage. The relationship between distributions and engineering capability is not that simple, and Ryanair's version of the accusation is self-interested.

A decade and a half of slippage on a critical replacement programme does sit alongside a sustained record of distributions, and the two have never been examined together in public.

Somebody should line up dividends, executive pay, capital expenditure, technology spend, and iTEC spend, year by year, from 2010 onward. If the answer is uninteresting, the operator gains an unusually strong defence. If it is not, the government owns 49 per cent of the problem.

Regulating Outcomes, Neglecting Machines

Downing Street's response to the collapse of the national flight processing system was to draw attention to a Bill strengthening passenger rights around cancellations, delays, and support. Because of course. A bill which has been sitting there for months is a handy thing to point to. Someone needs to do something about this, as always. The time to do it was a decade ago.

Stronger rights are worth having. They do not keep a computer running.

The pattern is by now well worn in British public life.

1. A critical system ages.
2. The replacement programme slips.
3. A regulator grows around the problem, generating plans, certificates, review cycles, and progress updates.
4. The operator demonstrates formal compliance throughout.
5. The system fails anyway.
6. An inquiry follows.
7. Recommendations are accepted, embedded, and scheduled into a multi-year implementation programme.
8. Another failure arrives before the programme concludes.

Aviation safety doctrine is genuinely world class at failing gracefully. Nobody died on Tuesday, or in July 2025, or in August 2023, and that is not a small thing. The doctrine for not failing at all, for treating a monopoly flight processing system as national infrastructure with hard availability obligations and real money attached to breaching them, has never been written.

Safety was allowed to stand in for resilience, and the regulator's attention followed the criterion it was given.

MPs were invited to approve wider CAA rule-making powers, a mandatory dispute resolution scheme, direct fining powers over airlines, and improved accessibility duties, the morning after the infrastructure the CAA supervises stopped the country flying.

Before granting a regulator new authority, it is customary to ask why it did not use the authority it already had.

The licence condition existed. The resilience plan existed. The independent review existed. The recommendations existed. Thirty-two of thirty-four were reported closed or embedded nine months ago.

An adversary shutting down British air traffic control would be a defence story, and Whitehall would know exactly which committee to convene.

Britain shutting down British air traffic control, using nothing more sophisticated than a legacy system nobody got round to replacing, is a story about whether the state can still operate the machinery it regulates.

The Bill will pass. The CAA will report in six months. NR28 will conclude in 2028. iTEC may arrive in 2030, assuming the people who understand the system it replaces are persuaded to postpone retirement.

In the meantime the country's skies depend on software, yet again, whose retirement was first pencilled in while Tony Blair was in Downing Street.