Three hundred million passengers a year by 2034. That is the number Ryanair keeps pointing at, and on August 12 Europe’s largest airline signed a five-year data and AI partnership with Google Cloud to help it get there. The rollout begins with Gemini Enterprise and Google Workspace going out to 35,000 employees.
What the airline actually bought
Gemini Enterprise is Google Cloud’s agentic AI platform, built to connect an organisation’s data, automate workflows and let teams assemble custom agents. Ryanair’s stated targets for it are automated decision-making, flight crew logistics and corporate productivity. Crew scheduling is the one to watch. It is a brutal constraint problem involving duty limits, base assignments, licensing, sickness and weather, and airlines have been throwing optimisation software at it for decades with mixed results.
The deal also pulls in Google DeepMind models, specifically AlphaEvolve and WeatherNext, for fleet operations and maintenance scheduling. Weather forecasting sits upstream of almost every expensive decision an airline makes in a day, from fuel loads to diversions to how many spare crews to position and where. A genuinely better forecast is worth real money to an operator running Ryanair’s turnaround times.
Two clouds, on purpose
Ryanair already runs on Amazon Web Services and extended that relationship for another five years alongside the Google announcement. The airline is deliberately moving to a dual-cloud architecture so critical services can shift between providers during an outage.
That is the quietly interesting decision. Multi-cloud has been discussed for a decade and mostly avoided, because running two stacks costs more and duplicates engineering effort. Airlines have a specific reason to pay for it anyway. A long outage at a carrier operating thousands of daily flights does not simply stop bookings, it strands aircraft and crews in the wrong places, and the recovery takes days. Redundancy stops being an abstraction when a cloud region going dark means passengers sleeping in terminals.
Against the wider spend
Gartner forecasts worldwide AI spending will reach $2.59 trillion in 2026, up 47 percent year on year, driven by infrastructure, software, services and agentic workloads. Money is not the constraint. Getting anything into production is.
Roughly a third of organisations report having begun scaling AI across the enterprise, which leaves most of them stuck between a successful pilot and a deployment anyone depends on. Research from Writer put the share of organisations facing meaningful adoption challenges at 79 percent in 2026, a double-digit increase on the previous year. More companies are trying, and more of them are hitting the wall.
Why this deal is a useful test
Ryanair makes a good subject for exactly that reason. The airline is famously cost-obsessed, operates on thin per-seat margins, and has no cultural patience for technology that does not pay for itself. It is not signing a five-year agreement in order to appear in a keynote. If Gemini Enterprise cannot cut crew disruption costs or maintenance delays, that will show up in the airline’s own operating numbers, which it reports every quarter and discusses in unusually blunt terms.
Most enterprise AI announcements are impossible to grade. This one is attached to an operator whose performance metrics are public, granular and closely watched. Punctuality, cancellations, cost per passenger. If the agents are working, those numbers should eventually say so.
The unglamorous half
Lost behind the DeepMind headlines is that a large share of this deal is Google Workspace for 35,000 people. Mail, documents, shared drives, and Gemini sitting inside all of them. That is not the part anyone leads a press release with, and it is probably where the first measurable returns show up, because it needs no model tuning and no integration project.
Ryanair’s workforce is mostly not sitting at desks. Cabin crew, pilots, ground handlers and engineers use software on phones and tablets in short bursts between tasks. Any assistant built for them has to survive that context, which is a harder design problem than a chat window on a laptop. Enterprise AI rollouts get announced in seats deployed. The number that matters is seats used more than once.
The next twelve months
Watch for whether Ryanair starts naming specific operational wins on earnings calls rather than describing the partnership in general terms. Watch too whether other European carriers follow with hyperscaler deals of their own, because airline IT tends to move in herds once one large operator proves a pattern is survivable.
The 300 million passenger target implies a substantial increase in volume without a matching increase in headcount. Something has to absorb that gap. Ryanair has now placed a large, dated bet on what.
For more coverage of AI in business, visit Mylistingo.
