Human vs AI at 250 km/h: guess who won by 1.58s
TUM retained the A2RL title in Abu Dhabi as six fully driverless cars raced head-to-head. Unimore set fastest lap. Daniil Kvyat stayed 1.58s ahead of the AI. Watch the highlights and what it means for UAE tech.
Abu Dhabi just hosted the largest autonomous car race yet. Six driverless machines went wheel-to-wheel at Yas Marina in a world-first Grand Final, and Germany’s TUM came out on top again.
Unimore’s raw pace was real, but contact while lapping swung the race. And in the human vs AI face-off, Daniil Kvyat stayed just ahead. For the UAE, it’s proof that autonomy research is getting faster, safer and more public.
KEY TAKEAWAYS
- Six fully autonomous cars raced in a world-first A2RL Grand Final at Yas Marina; TUM defended the title.
- Unimore led early, then collided while lapping traffic; they still took Fastest Lap.
- Daniil Kvyat beat the champion car’s AI by 1.58s in a 10-lap showcase.
- Eleven teams competed across Season 2 for a US$2.25m prize pool.
- Over 8,000 fans filled the North Grandstand; the race capped Abu Dhabi Autonomous Week.
What happened on the night
A six-car autonomous Grand Final delivered close racing, a lead change and a decisive incident.
- TUM retained the title, with TII Racing second and PoliMOVE third.
- Unimore passed TUM early, then collided while lapping Constructor and dropped out of contention.
- Cars topped 250 km/h and ran a 20-lap race on the North Circuit.
- Unimore earned the Fastest Lap award for the event.
Germany’s TUM started from pole and had to absorb a hard charge from Italy’s Unimore, who grabbed the lead at Turn 6 before lap two ended. The battle reached over 250 km/h, with gaps of under a second.
At mid-distance, Unimore clipped the back of Constructor while lapping, sending both cars off and handing the lead back to TUM. TUM held on to win the championship, while Unimore still walked away with Fastest Lap.
- Related reading: our Grand Final preview with timings and finalists on tbreak: A2RL Grand Final in Abu Dhabi: time, teams, how to watch.
Human vs AI: how close is it now
A 10-lap chase showed how small the gap has become at Yas Marina.
- Kvyat’s best lap: 57.57s; the AI “HAILEY” best: 59.15s. Gap: 1.58s.
- Rolling start with a 10-second head start for the AI.
- Finish was nose-to-tail for the crowd.
Former F1 driver Daniil Kvyat returned to face the champion team’s autonomous car. He had ten laps to close a rolling-start gap. He did, with a best of 57.57 seconds versus the AI’s 59.15, leaving a 1.58-second edge. Given the gulf was around 10 seconds only 18 months ago, the pace of improvement is obvious to anyone watching.
- Also see: our quick brief on the du 5G+ private network powering A2RL.
Why this matters for autonomy
Season 2 shows how competition squeezes better decisions out of code.
- A2RL positions itself as “science in the public domain.”
- Qualifying saw AI lap times match or beat human benchmarks.
- ASPIRE says 18 months of SIM sprints plus track testing drove gains.
Pushing robots to race at the limit forces better perception, planning and control. The league calls this “science in the public domain,” where the work is stress-tested in front of a crowd. Teams have gone from minutes off the human pace to fractions of a second either side. ASPIRE credits a mix of aggressive simulation sprints and heavy real-world testing for the leap.
Backgrounder: our qualifying wrap with pole position and lap records is here: A2RL S2 qualifying: TUM on pole, six-car final set.
The field, the stakes and the setting
Abu Dhabi staged the biggest autonomous car race to date with a packed stand.
- Eleven international teams competed for a US$2.25m purse.
- Six teams made the Grand Final: TUM, Unimore, Kinetiz, TII Racing, PoliMOVE, Constructor.
- Over 8,000 fans filled the North Grandstand.
- Date and place: 17 November 2025, Yas Marina Circuit, Abu Dhabi.
Season 2 drew 11 teams into weeks of testing, heats, and strategy work, all for a multi-million-dollar prize pool. The Grand Final featured six cars sharing the same track and fighting for position at speed, a genuine first. With a full North Grandstand and a busy fan zone, the UAE got a very public look at where safety-critical autonomy is heading next.
Community, partners and what’s next
The race anchored Abu Dhabi Autonomous Week and a wider skills push.
- ADAW bundled summits, exhibitions and RoboCup Asia-Pacific.
- The A2RL STEM Competition put 140 UAE students on AWS DeepRacer cars; UAE University and SABIS RAK won.
- Backers ranged from du infra and AD Ports to AWS and Abu Dhabi Mobility.
A2RL doubled as a public lab and a talent pipeline. The STEM Competition mirrors the big race using 1/18-scale cars, building practical skills for students across all seven emirates. Around the track, a long list of partners supported the event, from network and broadcast to car and data systems. It all sat inside Abu Dhabi Autonomous Week, which tied the racing to research and industry.
- More on the grassroots side: A2RL STEM Competition returns with 140 UAE students.
FAQ
Who won A2RL Season 2 in Abu Dhabi?
TUM retained the championship in the six-car Grand Final at Yas Marina.
Was this really the first six-car autonomous race?
Yes. The Grand Final was billed as the world’s first six-car autonomous head-to-head.
How close is AI to a pro driver now?
In the 10-lap showcase, Daniil Kvyat’s best lap was 57.57s and the AI’s best was 59.15s, a 1.58s gap.
How fast were the cars and how long was the race?
Racing touched over 250 km/h across 20 laps on the North Circuit.
Who runs A2RL?
A2RL is organised by ASPIRE under Abu Dhabi’s Advanced Technology Research Council, aiming to accelerate autonomous systems R&D.
Written by
Subscribe to our newsletter
Subscribe to our newsletter to get the latest updates and news
