An Alabama company has built one of the most ginormous robots the Army could ever want: a mini-tank that drives itself. Now to sell the Army on a bomb-fighting ‘bot the ground service actually once funded.

The machine in the video above is the Acer. Built by Mesa Robotics, it’s a tracked ground robot that weighs a whopping 4,500 lbs. and is about the size of a Bobcat loader or European electric coupe. Mesa considers it a one-stop shop for all the Army’s ground-robot needs, from stopping insurgent bombs to acting as a pack mule, with human control strictly optional. And yes, that is the soundtrack to Terminator 2 playing in the video.

What makes the Acer Terminator-esque is its artificial brain. An ex-Darpa employee named Steve Bruemmer designed a “Behavior Engine” that provides “intelligent instantaneous reactive responses to local environmental, sensor, and other data.” When Bruemmer’s 5D company won an Army contract for his Behavior Engine to help detect improvised explosive devices, he told Danger Room last month he was incorporating it into “a 4,500-pound vehicle that’s very promising,” but wouldn’t elaborate.

That vehicle is the Acer. “You could literally send it down the road and it’d run for 18 hours,” Mesa vice president Tim Cutshaw tells Danger Room.

But the Army’s effectively out of cash. It’s used a lot of ground robots at war — not always with good results — and may not see the need for an autonomous mini-tank.

The Acer’s origins start in 2003 with a forgotten Army program called the Robotic Combat Support System. The Army wanted a robot that could perform a task called “flail” — essentially tilling dirt to find bombs buried beneath, and then either disabling or detonating them without destroying the robot. Mesa got what Cutshaw pegs as a “couple million dollars” for the flail mission, and the Acer was born.

But it died young. “We were killed, initially, by FCS,” Cutshaw explains. That’s the Future Combat Systems, the Army’s big program to design and field futuristic vehicles, soldier communication tools — and robots, robots, robots. Long story short: FCS became a budget-buster, leading then-Defense Secretary Robert Gates to kill it in 2009.

Mesa saw an opportunity. If it could modify the Acer to do more than flail, maybe it’d have a chance at getting another shot at a fat Army contract. Shortly thereafter, Mesa execs met Bruemmer at a robotics conference and heard about his 5D Behavior Engine. “What the army wants is interoperability and autonomy,” Cutshaw says, “and 5D brings that to the table.”

So the new Acer was born. It still does flail: In the video, it digs up several roadside bombs and keeps moving as they detonate. When it’s not in flail mode, its ground-penetrating radar keeps scanning the dirt around it for an improvised explosive device. When it finds one, out comes the flail. Autonomously.

That’s not all. It’s a mule that carries a squad’s gear. Its flat top makes it a launchpad for small drones. (Though with only “a little bit of integration” with the drones, Cutshaw concedes, it’s not much more than a launchpad on treads.) And it opens up like a Transformer so a tinier robot, like a Packbot or a Small Unmanned Ground Vehicle (SUGV), can pop out.

But it might be too late.

First, the Army has years of unpleasant experience using robots in Afghanistan’s rugged conditions. Much smaller ‘bots, like the 125 lb. Talon ordnance-disposal ‘bot or the SUGV, either have difficulty with culverts or aren’t experienced with them. The far heavier Acer will be very difficult to get out of an Afghan ditch.

And the Army will be largely done with Afghanistan by 2014. Even if Mesa can sell Acer to the Army, the time it takes to buy and make the robotic tanks will mean it’ll miss the war. Then there’s cost: In the face of looming Pentagon budget cuts, the Acer’s $250,000 asking price might be too rich for the Army’s blood.

Still, Cutshaw is hoping to sell the Army on Acer’s autonomy and multiple functions. Hence the video, filmed last month — right around the time he took a meeting at Fort Leonard Wood to (inconclusively) pitch Acer. The next step, Cutshaw says, is “helping the Army figure out how to use this vehicle.” He may not get many more chances.