Formula 1’s 2026 power units have placed software-controlled energy deployment at the centre of car performance, leaving drivers increasingly dependent on computer systems to determine when electrical power is delivered and recovered during a lap.
The consequences became particularly visible at Spa-Francorchamps, where George Russell struggled to understand why his Mercedes was losing straight-line performance compared with his rookie team-mate. The difference was not attributed to a conventional set-up choice or an immediately identifiable mechanical failure. Post-weekend diagnostics instead traced the problem to an energy-deployment glitch after La Source, compromising Russell’s acceleration towards the Kemmel Straight and contributing to his contact with Lewis Hamilton.
The incident illustrated one of the defining technical challenges of the current season. Under the new rules, a driver can execute the required inputs correctly and still lose substantial performance because the car’s software has calculated a different deployment pattern.
How AI-driven energy deployment controls F1 performance
The 2026 power units divide propulsion between the internal-combustion engine and a much larger electrical contribution than in the previous generation. The approximate balance has been described as 53 per cent combustion power and 47 per cent electric power, making battery deployment fundamental to acceleration, straight-line speed and overtaking.
Managing that energy is no longer principally a manual task for the driver. Control electronics process information including wind conditions, available grip, tyre condition, driver inputs and previous laps before deciding when electrical energy should be deployed or harvested.
Rather than responding only to what has already happened, the system attempts to anticipate the demands of the lap ahead. It constructs a predictive energy model and distributes the available electrical power accordingly. Drivers retain access to an override function, but the car’s underlying deployment strategy remains governed by software.
The complexity of the process was already evident during the Belgian Grand Prix, when Ferrari and Mercedes explained why the 2026 power units had become so difficult to manage. At circuits featuring long full-throttle sections, an imperfect calculation can leave a car without sufficient electrical assistance at the most costly point of the lap.
Oscar Piastri highlights Formula 1’s growing software dependence
Oscar Piastri acknowledged after the Belgian round that coding and elements of artificial intelligence had already played an important role during the previous hybrid era. In his assessment, however, the degree of autonomy introduced for 2026 had significantly increased the computer’s influence over performance.
Qualifying can therefore depend not only on how accurately a driver performs but also on whether the energy-management system behaves as expected. Teams have encountered straight-line speed differences between cars running nominally identical specifications, with neither hardware nor set-up providing an obvious explanation for the variation.
Those discrepancies underline why understanding harvesting, deployment, clipping and derating has become essential to interpreting modern Formula 1. A car can reach the end of its intended electrical allocation before the end of a straight, creating a sudden performance loss even when the combustion engine continues operating normally.
Ferrari have also explored different methods of extracting performance from the system. During the Hungarian Grand Prix weekend, the Scuderia evaluated a Mercedes-inspired qualifying energy approach, highlighting how deployment programming has become another major development battleground alongside aerodynamics and mechanical set-up.
Energy calculations are influencing overtaking decisions
The impact is not confined to qualifying. Drivers must now consider whether completing an overtake will consume so much battery energy that they become vulnerable immediately afterwards.
Max Verstappen reportedly chose not to complete a pass at Suzuka because doing so would have drained his available electrical energy and exposed him to being overtaken again on the following straight. The decision demonstrated how a driver may have the opportunity to attack but still be constrained by the deployment consequences calculated for the rest of the lap.
Lando Norris has also indicated that the system can accelerate a car beyond a rival without the driver initially intending to complete the move. That was how he characterised his pass on Lewis Hamilton at Spa, suggesting that electrical deployment can influence not only whether an overtake is possible but also when it happens.
These situations create a difficult regulatory question. Article 19.1 of the FIA rules establishes the principle that the driver must operate the car alone and unaided, yet sophisticated energy-management software now makes decisions that directly affect acceleration and track position.
The driver still controls the steering, braking and throttle inputs, but the amount of power produced in response can depend heavily on calculations performed by the car. The distinction between managing a complex power unit and receiving prohibited assistance has consequently become increasingly difficult to define.
Why the FIA is unlikely to remove software-controlled deployment
The current technical regulations are intended to remain in force through 2030. Adjustments planned for 2027 and 2028 are expected to move the propulsion balance closer to 60 per cent combustion and 40 per cent electrical power, reducing some of the extreme energy-management demands created by the original concept.
Those revisions, however, are not expected to remove the software architecture governing deployment. The computer-controlled energy managers will remain, meaning teams and drivers must continue refining their understanding of how the systems interpret each lap.
The FIA has already faced difficulty securing agreement on changes to the regulations. Nikolas Tombazis explained that manufacturers had prevented earlier solutions from being adopted, despite the governing body anticipating several of the problems that later emerged. Formula 1 has since agreed a longer-term adjustment to the power-unit balance, but the fundamental dependence on deployment software remains intact.
Enforcing the driver-alone principle more strictly would require the FIA to identify an acceptable boundary within energy-management programs developed by teams and manufacturers over several years. With those systems embedded in the architecture of the power units, a major intervention before the next regulation cycle appears unlikely.
Formula 1’s commercial future is also moving towards AI
The technical shift is being accompanied by rapid commercial investment in artificial intelligence. Eight new AI partnership agreements were reportedly signed across Formula 1’s 11 teams during the first half of 2026.
Williams is using Anthropic’s Claude model, while Red Bull’s partnership with Oracle has developed towards an agent-based strategy system capable of identifying and presenting race information autonomously. McLaren’s relationship with Google has meanwhile expanded beyond promoting Pixel hardware to incorporate Gemini.
Technology sponsorship across Formula 1 was estimated at $769 million last season, representing a year-on-year increase of 41 per cent. Total sponsorship spending is projected to exceed $3 billion for the first time in 2026, with technology companies contributing significantly to that growth.
Artificial intelligence is therefore becoming more influential on both sides of the garage wall. It is helping teams process information and shape strategy while increasingly sophisticated software inside the car determines how electrical power is distributed.
F1 drivers remain responsible for decisions made by their computers
Formula 1 has traditionally presented the driver as the decisive element within an exceptionally advanced machine. That principle has not disappeared, but the 2026 power units have made it harder to separate driver performance from software performance.
A competitor can prepare correctly, execute the lap precisely and still lose acceleration because the deployment model has misjudged the required energy. Equally, the system may create an overtaking opportunity or initiate an acceleration advantage before the driver expected to act.
Drivers must ultimately adapt because they remain responsible for the outcome on track, even when the decisive loss or gain was generated by computer logic. As Formula 1 continues through the current regulation cycle, the championship must confront an increasingly important question: whether the person inside the cockpit is still controlling every meaningful aspect of the car’s performance, or supervising a machine that has begun making some of those decisions for them.
Leave a Reply