Offshore Racing 2025: When AI Surpasses F1 and Reinvents Sailing

A SailGP F50 catamaran generates more than 35,000 data points per second — more than a Formula 1 car. Right. Now here's the part that makes you smile: four engineers are enough to manage the entire global fleet. Four. Where F1 deploys around twenty per car.

This ratio — frankly absurd — sums up the revolution shaking offshore racing in 2025. The world's oldest sport (the one of wind, currents, and intuition) has become, without most people noticing, the most advanced maritime laboratory on the planet.

SailGP and America's Cup: AC75s and F50s become "extreme IoT devices"

Warren Jones, CTO of SailGP, doesn't mince words: "These boats on the water? I call them 'extreme IoT devices.' We capture a huge amount of information from the boat and then turn it into metrics that we can understand and make better decisions with."

The numbers are staggering. Each F50 carries 125 sensors; across the fleet, that's 240,000 data points per second flowing to Oracle Cloud Infrastructure. End-to-end latency? 150 milliseconds. The time it takes to blink — and the tactical decision has already been made. Or at least suggested.

On the America's Cup side, the escalation is even more spectacular. American Magic's AC75 carries around 300 sensors measuring 4,000 distinct variables. The data team has grown from 2 to 6 analysts, a sign that the volume of information now exceeds raw human processing capacity.

At Orient Express Racing Team, Bruno Dubois explains: "As soon as we arrived on April 6, the hydraulics, electrical, and mechatronics departments took over the boat. This precision work took nearly seven weeks." More than 200 sensors installed over 46,000 hours of construction, by 65 builders. A shipyard, or an aerospace project? The question is seriously worth asking.

And then there's this constraint that changes everything: AC75s are banned from physical testing — no towing tank, no wind tunnel. Result: the race is played out entirely in digital twins. Steve Collie, aerodynamics coordinator for Emirates Team New Zealand, confirms: "We continue to run simulations until the final days of racing in October to extract every fraction of performance." When you can't touch the water, you push the limits of computation.

Predictive AI detecting the invisible: nearly 400 million readings per race

Here's the striking figure: nearly 400 million hydraulic readings analyzed by Oracle's AI at each SailGP race. The system detects an average of one fault every two days of sailing.

Result? Zero hydraulic failures in racing since deployment.

Zero.

Scott Babbage, Head Data Analyst at SailGP, describes the daily operation with disarming pragmatism: "Anomaly detection takes about six or seven hours to do a deep analysis. We wake up in the morning, and it tells us what our technical team needs to replace. It's very efficient." And adds: "Scale is the advantage here. We have a small team of engineers who look after a lot of boats. When you think about Formula 1, they probably have 20 engineers working on each car. We have a team of four engineers who look after all our boats."

For the 2025 season, SailGP takes an additional step with AI-driven anti-crash systems and smart cameras onboard. The goal? Not only to prevent mechanical failures, but to anticipate collisions between boats flying at over 50 knots just meters apart.

Across the maritime industry, these technologies are producing measurable results: nearly 70% reduction in unplanned failures, up to 15% reduction in fuel consumption, with ROI typically under two years.

Adaptive foils, CFD, and digital twins: simulation becomes reality

The jump from 6 foiling IMOCAs in the 2016 Vendée Globe to 19 in 2020 illustrates an irreversible shift. In four years, the foil went from technological advantage to competitive prerequisite. But this democratization masks a far more complex digital arms race.

IMOCA rules strictly limit foils to a second degree of freedom (rotation) of maximum 5 degrees and prohibit any automatic control during racing. In other words: the sailor remains master of the flight. But all the intelligence lies in the upstream design — in the thousands of hours of CFD (Computational Fluid Dynamics) that determine the optimal profile before the carbon is even molded.

The case of Gitana 18 — the Ultim trimaran of Team Banque Populaire — pushes the cursor even further: around 500 sensors onboard, custom-developed autopilots for oceanic flights at 55 knots. At that speed (over 100 km/h on water), the slightest piloting error comes at a price. The autopilot doesn't replace the skipper; it gives them those few tenths of a second of extra reaction time that separate the world record from shipwreck.

Anderson Reggio, former testing manager for American Magic, makes the diagnosis: "The amount of data you have on an AC75 is enormous and no single person can process it in one night and produce accurate insights." The digital twin is no longer a luxury. It's a structural necessity. This quest for permanent optimization is also found on the IMOCA circuits — and if you want to follow the adventure of these machines closely and directly support a skipper, Yannick Bestaven's Maître Coq V offers several sponsorship packages on Spencer. You can even sail on this boat through a sponsorship on Spencer.

WindSight IQ, LiveLine, and augmented spectacle: when AI transforms the fan experience

How do you make a novice understand why a boat tacks to starboard rather than port?

By showing them the wind. Literally.

The WindSight IQ system, developed by Capgemini for the America's Cup, relies on 3 LUMIBIRD LiDARs positioned on Barcelona's waterfront. Range: 12 kilometers. Accuracy: under 0.1 m/s. Refresh rate: every second. For the first time, the wind — that invisible element that decides everything — becomes visible on screen.

Grant Dalton, CEO of the America's Cup Event, sums up the challenge: "Four elements determine the winner of the race: boat design, team, water, and wind. Before and during the race, commentators will now be able to see real-time wind patterns and explain to viewers the options available to competing yachts."

On the SailGP side, the LiveLine system processes 53 billion data points per race day to produce graphic overlays with 2-centimeter accuracy, broadcast in 212 territories. The key figure validating the strategy? 84% of fans say these technologies improve their viewing experience. And most importantly: 77% of Gen Z prefer watching sports away from live venues — from a screen, enriched with data.

Sailing, an insider's sport? Not for much longer.

From Harken sensors to Cariboni systems: the high-precision hardware ecosystem

Behind the algorithms, there's metal. Silicon. Miniaturized steel.

Harken pressure sensors illustrate the level of requirement: 414 bar maximum pressure (equivalent to 6,000 psi), for only 90 grams on the scale. Output in industrial 4-20 mA signal, compatible with any data acquisition system. It's these micro-components, invisible to the eye, that enable hundreds of millions of readings per race.

The software ecosystem is structuring in parallel. Platforms like Kwindoo, SAP Sailing Analytics, Oracle Racing Analytics, or Sailmon are emerging to democratize access to performance data. What was reserved for America's Cup teams five years ago is beginning to trickle down to amateur and semi-professional offshore racing.

Industrial transfers: from Vendée Globe to CMA CGM cargo ships

100% of the 2024 Vendée Globe fleet was equipped with Watt&Sea hydrogenerators. A technology born from offshore racing in 2008, which became standard in sixteen years. The pattern repeats, always identical: competition invents, cruising adopts, industry deploys.

AdrenaShip has been optimizing routes for CMA CGM and Brittany Ferries since 2014. Intelligent weather routing — honed on transatlantic races — now saves commercial fleets significant amounts of fuel. Adoption figures confirm the acceleration: 276 organizations were using maritime AI in 2023; there are 420 in 2025. A progression of over 50% in two years.

Predictive maintenance (the same technology that detects hydraulic anomalies on F50s) reduces breakdowns by nearly 70% on merchant vessels and decreases fuel consumption by up to 15%. When a shipowner calculates the return on investment, the answer comes in under two years. The environmental argument and the economic argument converge — which, in shipping, doesn't happen often.

Vision 2030: smart foils, generative AI, and partial autonomy

Where will we be in five years?

Current trends paint a fairly clear horizon. Digital twins and adaptive foils will become the norm across all oceanic classes, not just exceptional prototypes. Generative AI — which writes text and produces images today — will tomorrow design hulls, appendages, rigs. As in automotive, where optimal shapes already emerge from algorithmic processes that the human eye would never have imagined.

The convergence between offshore racing and commercial maritime will accelerate further. What equips an AC75 today will be found on a container ship in three to five years. Performance gains are measured in percentages; environmental gains, in millions of tons of CO₂. To follow this evolution in real time — and discover the dates of the next major technological showdowns — visit the Spencer calendar.

The paradox is striking. Sailing — that millennia-old sport where you harness the wind with lines and canvas — has become the absolute cutting edge of innovation in artificial intelligence applied to transportation. Four engineers, 150 milliseconds of latency, nearly 400 million readings per race. And always, at the end of the chain, a crew that must read the water, sense the gust, commit the boat at the right moment.

AI doesn't replace the sailor. It gives them eyes where they were blind.