technologybriefs
11:32in productionCh. 1 · The Promise/ 11:32 · ceiling 15 min
Robotics · Failures

Rethink Robotics

Rodney Brooks · 2008

Collaborative robots are not safer by design — they’re safer only when their limits are never tested.

Rethink Robotics launched Baxter in 2012 as a collaborative industrial robot designed to work safely beside humans — using physical stoppage and manual override, not AI. It released Sawyer in 2015 for finer tasks, then went global with cobots in 2016. It ceased operations in October 2018. Its assets were sold. Its later relaunch attempts failed within a year. Its safety mechanisms worked. Its business model did not.

RoboticsFailuresAutomationFailureHardwareAutomation across the network →
Chapters & takeaways4
  1. 1:04
    The Promise

    Rethink Robotics built Baxter and Sawyer as low-cost, safe, programmable cobots — not to replace workers, but to share space with them.

  2. 3:02
    How It Actually Worked

    Baxter’s safety relied on mechanical stoppage and manual override — not AI perception or adaptive response.

  3. 4:44
    The Collapse

    Global cobot rollout in 2016 preceded collapse: Rethink shut down in October 2018 after failing to achieve sustainable adoption.

  4. 6:43
    What Remained

    Asset sales — to HAHN Group and Hunan CoThink — confirmed dissolution, not transition; later ‘relaunches’ were asset rebrandings, not operational continuity.

Worth your time?

Yes. Study the whole thing.

2.5/ 5
What works
  • Physical safety implementation
  • Human-readable interface design
  • Low-barrier programming via demonstration
What does not
  • Deliver sustained commercial viability
  • Scale beyond pilot deployments
  • Integrate into complex or variable workflows
Study it if
  • Robotics engineers evaluating safety architecture
  • Operations managers assessing cobot TCO
  • Historians studying post-2010 automation narratives
Skip it if
  • Manufacturers seeking production-ready cobots
  • Software developers building on Intera 5
  • Investors assessing robotics startups
The written brief1 min read

What it is and the problem it solves

Rethink Robotics was a robotics company co-founded by Rodney Brooks and Ann Whittaker in 2008. It built collaborative industrial robots — Baxter (2012) and Sawyer (2015) — to work alongside humans in factories without cages or safety fencing.

How it works

Baxter used physical safety mechanisms: it stopped automatically on human contact and featured a manually accessible emergency off switch. It was programmable by demonstration, not coding. Sawyer used the same Intera 5 software platform and shared Baxter’s collaborative design ethos but targeted smaller, more precise tasks.

What works

Baxter’s physical safety features worked: automatic stopping on contact and a prominent manual off switch delivered verifiable, immediate de-energisation. Its design prioritised ease of setup and human-readable feedback — large eyes, arm motion cues, and intuitive teach-by-demonstration programming.

What does not

It did not sustain a viable business model. It failed to scale beyond niche adoption. Its cost projection — equivalent to a $4/hour worker — ignored real-world integration costs, maintenance, training, and software lock-in. Its cobots required factory environments with controlled layouts and predictable workflows, limiting flexibility.

What it changes

It established that industrial robots could be physically safe enough for unshielded human proximity — a regulatory and design precedent. It proved that collaborative robot interfaces could be simplified for non-programmers. It shifted early industry conversation from replacement to co-location — but did not deliver durable deployment at scale.

Is it worth your time

No. Rethink Robotics ceased operations in October 2018. Its assets were sold. Its attempted relaunches — Sawyer BLACK Edition (2019), relocation (2020), and September 2024 relaunch — all failed within a year. The company no longer exists as an operational entity.

Same field · Robotics4 of 26
10:52
Unmanned aerial vehicleUAVs are aircraft without onboard pilots. They solve access problems in hazardous or repetitive aerial tasks. Their mechanism relies on remote control or programmed autonomy, enabled by improved electronics and cheaper components. Military adoption was complete by the twenty-first century. Civilian use followed regulatory shifts: UAS terminology formalised in 2005; FAA civilian airspace permission came in 2006; DJI’s 2013 Phantom lowered the consumer barrier. But autonomy remains narrow: Ingenuity flew on Mars (2021–2024), yet no global standard governs lethal AI targeting—the Kargu 2’s 2020 Libya strike exposed that gap. Certification lags: EASA’s 2024 ETSO-C198 basis for Embention’s flight controller is the first of its kind. UAVs change who bears risk—and who decides when a machine may act.
9:15
Mobile robotA mobile robot is a locomotive, automatic machine—not fixed, not necessarily intelligent. It works by combining controller, sensors, actuators and power. It succeeds where movement and environment match. It fails when autonomy is assumed but not engineered. It changes infrastructure from static to relocatable. It is worth your time if you need machines that move—not just compute.
10:28
Unmanned surface vehicleUSVs are operational—but not systemic. They deliver real results in niche applications. They lack standardisation, interoperability, and regulatory grounding. Their value lies in removing humans from risk—not in replacing captains with code.
10:32
Self-driving truckSelf-driving trucks are a systems-level adaptation of autonomous technology to freight logistics. They rely on multi-sensor fusion and AI navigation, but their real-world deployment is bounded—not by capability, but by self-imposed safety thresholds and infrastructural control. Kodiak’s December 2024 launch on private lease roads is the first commercial driverless operation in the U.S., yet no autonomous truck has hauled freight without a human on public highways. What works is geofenced, industrial, or military convoy logic—not open-road autonomy.
Up next in Technology

RISC-V

UC Berkeley researchers · 2010 · 10:49

RISC-V is not an alternative CPU—it’s an alternative to the idea that instruction sets must be owned.

Play nextBack to Robotics
10:49