AgBot with Kongskilde Viberflex used 60 litres of diesel per hectare on Will Mumford's farm in Cambridgeshire. A wheeled tractor with plough needed 75 l/ha; the Challenger with Väderstad TopDown took 100 l/ha. The result came from a three-strip field trial Mumford set up to test whether the autonomous AgXeed machine could cut fuel use and soil damage.

Mumford runs 400 hectares near St Neots and is one of the first arable farmers in England to operate an AgBot. He also runs the AgXeed import business for England and Wales through his trading company ASC — a position that makes him a partisan witness in any fuel and soil comparison.

A new AgBot costs around £320,000; second-hand examples sell for about £150,000. About eight AgBots now work on UK farms, including a demo unit on his own land.

What did the trial compare?

Mumford divided a single field into three sections after harvest. One strip went under the AgBot pulling a Kongskilde Viberflex. Another took a conventional plough with a wheeled tractor. The third took a Challenger tracked tractor with a Väderstad TopDown cultivator. A neighbouring farmer who works as a soil specialist by day helped plan the test.

After one pass, the Viberflex produced a seedbed too shallow to match the other two treatments. The AgBot strip therefore needed a second pass. Even with that extra work, the autonomous setup still came out on top.

How much fuel did each setup burn?

• AgBot + Kongskilde Viberflex (two passes): 60 l/ha
• Wheeled tractor + plough: 75 l/ha
• Challenger + Väderstad TopDown: 100 l/ha

The AgBot used 40% less diesel than the Challenger combination. Mumford wants to isolate where the saving sits. "Is the AgBot now much more fuel-efficient than a Challenger, or does the Viberflex require much less draft power than the TopDown?" he asked. A follow-up trial will pair the AgBot with a Viberflex against a Challenger with a Viberflex.

What did the soil tests show?

The Challenger plus implement weighs 24 tonnes. Even on tracks, that weight loads onto a few rollers during headland turns, and Mumford saw a thinner crop stand in those zones. Penetrometer readings did not separate the three treatments cleanly, even though the wheat grew differently. "Still surprising, because the wheat did show a difference in growth," Mumford said.

The team then ran an infiltration test. They drove a 20cm pipe 7.5cm into the soil, filled it to the top with water and timed how long it took to drain.

• AgBot + Viberflex: 1 minute 36 seconds
• Ploughed strip: 1 minute 21 seconds
• Challenger + TopDown: more than one day

They repeated the test three times. The TopDown strip still failed to drain in under 24 hours. Mumford sold the TopDown afterwards. "It is a fantastic implement, but not for my farm," he said. "A light autonomous vehicle with a light implement protects the soil and helps take yield to a higher level."

What changes for next season?

The infiltration gap points to capped pores under the TopDown pass. Faster drainage on the ploughed strip fits with the loosening action of the mouldboard. The AgBot result suggests that implement weight, not tractor power, drove the damage on this soil.

Mumford registered 100 hectares as subsidised fallow, paid at £600/ha, after the scheme's budget was heavily oversubscribed. Winter wheat covers 250ha, with 20ha of spring barley, 4ha of sunflowers, 2ha of grain maize and 2ha of millet. He would rather enter more land into the fallow scheme than grow wheat.

The next trial will resolve the fuel question before he commits more hectares to autonomous cultivation. Watch the AgXeed importer's social channels and the Cereals event for the first results.