MOZA Racing used Gamescom 2026 to show its Racing Lab AI Coach in one of its clearest demonstrations so far. Instead of limiting the presentation to telemetry graphs and automated laps, MOZA placed the system against Jarno Opmeer and Porsche Carrera Cup driver Dirk Schouten at Spa-Francorchamps, using a Porsche 991 GT3 R and two matching simulator setups.
The demonstration was designed to show more than outright lap time. Racing Lab combines telemetry analysis, AI-generated reference data and direct control of compatible MOZA hardware. The system can record a driver’s inputs, identify where time is lost and then reproduce its own steering and pedal behavior through the wheelbase and active pedals.
Racing Lab records detailed telemetry while the driver practices. The software analyzes lap time delta, speed, braking, throttle, gear selection and steering input. Drivers can then select a sector or corner and compare their own racing line and inputs against the AI reference, making it easier to identify exactly where performance starts to fall away.
MOZA is also developing Racing Lab as a more accessible coaching tool. Instead of requiring beginners to interpret every telemetry trace themselves, the AI can identify areas that deserve attention and direct the user toward the relevant data. This can highlight problems such as braking too early, releasing the pedal too quickly, carrying too little speed or delaying throttle application.
The most unusual feature is the AI’s ability to control the simulator hardware directly. Racing Lab can steer through a compatible MOZA wheelbase while the mBooster Active Pedals reproduce braking and throttle inputs. A driver can therefore feel when the AI brakes, how it releases pressure into a corner and how much steering input it applies.
MOZA sees this physical feedback as a way to make techniques such as trail braking easier to understand. Drivers can first observe an AI lap, then experience the same inputs through the hardware before attempting to reproduce them. Racing Lab also offers real-time telemetry overlays and audio braking cues to support drivers while they remain on track.
Human vs AI at Spa-Francorchamps
MOZA turned the technology demonstration into a direct Human vs AI challenge at Spa-Francorchamps. One simulator was controlled by the human driver, while a second rig showed the AI physically operating the steering wheel and pedals. This allowed spectators to compare both driving styles while watching Racing Lab control the virtual Porsche without human input.
Dirk Schouten took the first attempt. Although he regularly trains on simulators alongside his Porsche Carrera Cup and Supercup programs, he still needed time to adapt to the specific hardware and braking characteristics of the Gamescom setup. His lap times improved quickly as he became more comfortable with the car and equipment.
The AI established a pace around the 2:14.6 to 2:14.7 range. Schouten eventually reached approximately 2:16.8, reducing the gap to just over two seconds after only a small number of laps. His improvement showed how much simulator familiarity can influence performance, even for an experienced real-world racing driver.
Jarno Opmeer then took over the human simulator and produced a much closer fight from the beginning. He made the stronger start and kept the Racing Lab car behind during much of the opening lap. The AI eventually moved ahead, but Opmeer continued experimenting with different braking points, racing lines and track limits as he learned the setup.
Opmeer’s best displayed lap reached approximately 2:16.1, while the AI again ran in the mid-2:14 range. That left the three-time F1 Esports World Champion around one and a half seconds behind the fastest AI lap shown during the demonstration, although his pace continued to improve throughout the limited run.
What the Human vs AI Result Really Shows
The Gamescom result should not be treated as proof that Racing Lab can simply outperform professional drivers under equal conditions. The AI had already been trained for the selected Porsche and Spa combination, while Opmeer and Schouten had only a few laps to adapt to the simulator, pedals, force feedback and car.
Both drivers became noticeably faster as the challenge progressed. Racing Lab, by comparison, already had a trained reference for the chosen combination. The event was also a live Gamescom demonstration rather than a controlled benchmark, so differences in preparation and familiarity need to be considered when comparing the final lap times.
The AI also did not appear to repeat exactly the same lap every time. Its lap times changed between attempts and some racing lines differed from the approaches used by the human drivers. Opmeer and Schouten both commented on unusual choices from the AI, although the resulting pace remained consistently competitive.
MOZA later invited a Gamescom visitor named Felix to attempt the same challenge. He began around the 2:35 range but improved to approximately 2:27 after only a few laps. Racing Lab remained far ahead, but the improvement helped demonstrate how the system could also work as a reference for less experienced sim racers.
Racing Lab Is Designed as a Training Tool
The Human vs AI format naturally focused attention on which side could produce the fastest lap, but MOZA’s wider objective is driver development. Racing Lab is designed to explain where additional speed comes from rather than simply demonstrating that an AI can complete a fast lap.
The learning process can begin with an AI driving demonstration before moving into telemetry comparison and physical feedback through the wheel and pedals. Drivers can then gradually reproduce the same techniques themselves, using the AI reference to understand braking points, trail braking, racing lines and steering behavior.
This combination gives Racing Lab a different role from conventional telemetry software. Instead of only presenting data after a session, MOZA is trying to connect analysis with physical feedback. The driver can see what the AI does, compare the underlying inputs and then feel those inputs through the same hardware used during practice.
At Gamescom, Racing Lab ultimately produced faster laps than both Schouten and Opmeer during their short attempts. However, the more important demonstration was how the system generated that reference and communicated it back to the driver. For MOZA, the goal is not simply Human vs AI competition, but using AI to help human drivers understand how to become faster.
See you on the track!
This website uses affiliate links which may earn a commission at no additional cost to you.
