Fourier GR-2 vs Kepler Forerunner K2: Which Wins?

Choose the Kepler Forerunner K2 if you need a robot that lifts and carries for a full shift, because it handles 15 kg per arm and runs for 8 hours. Choose the Fourier GR-2 if you run research, teaching or rehabilitation projects, because it walks faster at 5 km/h, has finer 12-joint hands and ships with open developer tools. The K2 is the newer machine by one year, launched in 2025 against 2024 for the GR-2. For factory work the K2 is the stronger fit; for labs the GR-2 is.

The Forerunner K2 carries far more, at 15 kg per arm against 3 kg per arm continuous for the Fourier GR-2. Its maximum deadlift is 25 kg, while the GR-2 is rated at 6 kg. In practice the K2 can move a full parts bin or a heavy tool with one arm, while the GR-2 suits light items such as sample tubes, small boxes and medical supplies. If your task involves lifting anything heavier than a few kilos, the GR-2 will not do it, and the K2 is the only one of the two that fits.

The Forerunner K2 runs for 8 hours on a 2.33 kWh battery, while the Fourier GR-2 runs for 2 hours on a battery of about 2.3 kWh. Both packs can be swapped, so a spare battery keeps either robot working. The GR-2 recharges in 1.5 hours and the K2 in 2 hours. The gap matters most on a shift: the K2 can cover a morning without a stop, while the GR-2 needs a swap or a charge several times a day. For a lab demo the GR-2's two hours is usually enough.

The Fourier GR-2 has the more articulated hands on paper, with 12 joints per hand in its enhanced version and 6 touch-sensor arrays that let it feel an object's shape. The Forerunner K2 has 11 joints per hand and tactile fingertips, and its hands are built to grip loads of up to 15 kg. The GR-2 hand is made for delicate tasks such as handling lab tools or guiding a patient's arm. The K2 hand is made for firm, repeatable grasps on factory parts. Joint count favours the GR-2; grip strength favours the K2.

The Fourier GR-2 walks faster, at 5 km/h against up to 4 km/h for the Forerunner K2. Both are bipeds, meaning they walk on two legs, and both are designed for indoor floors. The 1 km/h gap is small in a lab or a showroom, but over a long plant corridor it adds up to a few extra minutes per trip. The K2 trades some speed for a heavier frame, at 75 kg against 63 kg, which helps it stay steady while carrying a load. Neither robot is meant for outdoor ground.

The Fourier GR-2 offers the more open toolkit, with ROS, Linux, NVIDIA Isaac Lab and MuJoCo support, which suits universities and teams that train their own motion models. The Forerunner K2 runs a Linux-based system that is ROS-compatible and adds proprietary programming interfaces from Kepler. Both learn through reinforcement learning, a method where the robot improves by trial and error in simulation. If you want to write and test your own behaviours, the GR-2 gives more room. If you want a working factory robot with less setup, the K2's package fits.

The Forerunner K2 is aimed at materials handling, sample processing, patrol and quality control, and it has worked in a Shanghai car plant, moving material bins. The Fourier GR-2 is listed for rehabilitation help, patient guidance, light medical item transport and laboratory work, and Fourier reports trials in guidance, research, manufacturing and rehabilitation. Both are rated IP54, meaning they resist dust and splashes, and both run indoors. Pick by job: heavy, repeated shifts point to the K2, while careful hand work near people points to the GR-2.

The Forerunner K2 is the better value on published numbers, with a listed price of $50,000 to $70,000 against USD 100,000 to 150,000 for the Fourier GR-2. For that lower price the K2 carries five times the load per arm and runs four times as long. The GR-2 costs more but gives finer hands and a more open software platform, which research teams may value above payload. A buyer who needs a worker gets more from the K2. A buyer who needs a research platform may still choose the GR-2.