Perspectives

Are we turning a corner on the humanoid robot age?

23 February 2023

With the technology transition from autonomous vehicles to autonomous robots a logical one, humanoid robots could soon become a reality in our everyday lives.

Few subjects have captured the popular imagination like the possibilities of robots. For decades they have been the subjects of outlandish future predictions or cast as existential threats to humanity in Hollywood movies.

But, thanks to significant technological advances, falling costs and a greater understanding of the potential role in the factory of the future, robotics has now advanced to the point whereby it has begun to transform multiple industrial processes.

The use of collaborative robots (‘cobots’) as a manufacturing or precision tool is now widespread within the industrialised world. Advances in artificial intelligence (AI) mean we already see numerous digital processes being undertaken by robots. And thanks to the autonomous vehicle (AV), many can imagine a time in the near future when a robotic (driverless) car will take us to work.

Yet one significant frontier still to be breached is the acceptance and adoption of humanoid robots. Built to resemble and act like humans, and able to interact with human tools and environments, humanoid robots have thus far struggled to make it beyond the concept phase.

While floor-based mini-bots can perform simple tasks like vacuuming, sophisticated humanoid penetration of our most personal space is as yet a distant reality. Some researchers believe this is unlikely to change anytime soon: citing the required advances in platform agility and task complexity as major stumbling blocks standing in the way of a useful home-bot becoming a reality.1 

 

New horizons

Yet it pays to remember that not long ago, few of us could imagine plugging our car into a wall to charge it. Now, the rise of the electric vehicle (EV) is not only a major automotive leap forward,2 but their adoption is taken for granted as being a necessary step towards advancing the global energy transition.

The rapid rise and acceptance of game-changing developments like EVs and AVs does provide a test case in technical and product terms for the humanoid robot project.3 And as brands like Amazon4 and Dyson5 announce new forays into the consumer robot space, market futurists suggest a feasible time for humanoid mass production is finally approaching.6 It seems timely to ask, from a market size currently estimated at $US1.5 billion in 2022,7 where could robots be within our lifetime?

 

Timeline to a robot workforce

The humanoid robot penetration curve will be similar to that of battery electric vehicles

Source: Macquarie Research, September 2022

Research by Goldman Sachs suggests that a $US6 billion market for humanoid robots is achievable within the next 10 to 15 years.8 Initially, of course, demand will come from the industrial sector. Yet among its predictions, Goldman Sachs forecasts that by 2035, humanoids could help fulfil 2 per cent of global elderly care demand.9

 

A breakthrough product

Wendy Pan, machinery analyst for Macquarie Research in Japan, agrees that launching the humanoid robot for industrial use makes perfect sense, particularly as the product achieves greater sophistication. And, as her team explores in a recent investor report, this may well be just the tip of the iceberg.

The report predicts that home-based market demand for humanoid robots will act as the tipping point that helps propel the humanoid product category into a $US3 trillion market by 2050.10 Achieving such penetration in just 28 years would represent a major step forward for both industry and society.

According to Pan, it could mean that within the next three decades a humanoid robot could become as important to families as the car is today. “The car helped to shorten people’s commute time. I see the purpose as similar for humanoid robots: to shorten people's time spent on housework, making people's lives easier and more convenient,” she says. “I think it's a technology that will change our lives.”

 

Ahead of the curve

Tesla founder and CEO Elon Musk is among the technologists who agree with this sentiment. Speaking at the 2022 unveiling of Tesla’s humanoid prototype named Optimus, Musk predicted that once fit for purpose, the home-based humanoid robot would represent a “fundamental transformation of civilisation as we know it”.11

Progress made in the EV and AV space – which has advanced the affordability, availability and sophistication of advanced sensors, AI chips and batteries – is one of the reasons that Tesla’s initial prototype has been encouragingly quick to materialise, believes Pan. It also directly influenced Optimus’s projected pricing of $US20,000 per unit (which, if it proves correct, will see the market respond positively, she believes) and is paving the way for the broader commercialisation of humanoids.

 

An early-stage humanoid robot’s hardware costs

Estimated total hardware costs for an early-stage humanoid robots

Source: Macquarie Research, September 2022

Macquarie estimates total hardware costs for an early-stage humanoid robot at around $US40,000. Specifically, $US10,000 for sensors and chips, $US5,000 for inertial measurement units and torque sensors, $US10,000 for servomotors and motor drives, $US8,000 for precision reducers, $US2,000 for battery and battery control system and $US5,000 for all its other parts including body materials.12

Challenges ahead

Optimus is among four major prototypes that Macquarie Research has identified as setting the early running in the humanoid market: with the remaining three including Honda’s ASIMO, Boston Dynamics’ Atlas, and Xiaomi’s CyberOne.

In terms of their likely evolution, Pan and team predict that from now until the mid-2030s, humanoid bots within factories will start replacing repetitive or dangerous tasks performed by humans, as well as being adopted for use by those with physical challenges. Then from the late-2030s to the 2050s, humanoids should begin household penetration, with homes being the target market that truly drives mass production.

So why the likely time lag between factory and home penetration for humanoids? Simply put, the home is a far more complex and fragile environment.13 In a technical sense, engineers will be busy adapting the humanoid to cope with the degree of flexibility required to cope with a variety of surfaces, including carpets, tiles, wood and stairs. Not to mention the challenges provided by young children and household pets. Likewise, if robots are to cook, clean and assist in our daily medicine chores, their ability to judge materials and adjust grip strength accurately will be a key to their acceptance.14

 

Investor opportunity

For the next decade at least, the real business opportunity for investors eyeing the humanoid robot project will mostly be on the supply chain, especially around hardware, AI chip makers and sensor makers.


Comparison of major humanoid robot prototypes

Four types of humanoid robot are emerging, based on prototypes from four companies

Tesla

Model: Optimus/Tesla Bot
Height (cm): 173
Weight (kg): 57
Degree of freedom: 40
Speed (km/h): 8.0
Force (kg/hand): Extended arm can lift 4.5kg
Power source: Battery
Operating time: n/a
Actuators: Electronic


Xiaomi

Model: CyberOne
Height (cm): 177
Weight (kg): 52
Degree of freedom: 21
Speed (km/h): 3.6
Force (kg/hand): n/a (two fingers)
Power source: Battery
Operating time: n/a
Actuators: Electronic


Boston Dynamics

Model: Atlas
Height (cm): 150
Weight (kg): 89
Degree of freedom: 28
Speed (km/h): 9.0
Force (kg/hand): Can be either with or without fingers
Power source: Battery
Operating time: n/a
Actuators: Hydraulic


Honda

Model: ASIMO
Height (cm): 130
Weight (kg): 48
Degree of freedom: 57
Speed (km/h): 9.0
Force (kg/hand): Grasping force 0.5/hand (five fingers)
Power source: 51.8v Li-ion battery
Operating time: 1 hour
Actuators: Electronic (servomtor+harmonic speed reducer+Drive unit)

Source: Company data, Macquarie Research, September 2022


In addition, an increased focus will turn to the current missing links. “The bottleneck for the production of humanoids is in the actuators,” says Pan. “They should be the focus for humanoid makers or market investors.”

Over time, as the product becomes more technically sophisticated, debate over its mass viability will kick in. One significant hurdle is regulation. Given that humanoids are more complex than autonomous vehicles, there are a host of safety issues ranging from the prevention of mishaps to the alignment of responsibilities, should they occur. Likewise, ethical issues around their use will need to be addressed, while cybersecurity risks must be taken seriously.

The next question will be practicability. Or, in other words, how useful humanoids will be, and whether their use case represents true value for money. Likewise, given the product is in the home, just how patient the market will be with early faulty versions remains to be seen.

And finally, profitability will be assessed. While eventually the winning prototype may be hugely profitable, the potential losses prior to mass production could be challenging for the market to absorb.15

 

Signs of acceptance

Perhaps most encouragingly from a recent media survey of robot-watchers, is that the widespread predictions of robots having a negative impact on jobs, seem to be weakening. “Economists are revising their views on robots and jobs”, writes The Economist. Despite predictions, as global investment surges, evidence of automation-induced unemployment appears scant. Instead, the developed world faces a record 30 million unfilled vacancies across the OECD. “What seems clear at this stage is that the era of sweeping, gloomy narratives about automation is well and truly over.”16

Recent studies support the notion that automation and job growth will continue to coexist. Harvard University researchers suggest there is a “new view” of robots, predicting that “the direct effect of automation may be to increase employment at the firm level, not to reduce it.”17

Pan agrees that while numerous job categories have been replaced by machines over the past two centuries, new work categories have sprung up. “I don't think humanoids are especially harmful in terms of replacing jobs,” says Pan. “And in some areas where there is a real lack of labour nowadays, as in say the healthcare and aged care industries, I think humanoid robots will provide a labour source for those industries.”

In time, Pan believes humans and their robot partners will happily coexist. As the relationship evolves, she predicts that fear of humanoids will dissipate. “It's just a technology that’s helping people to achieve better things.”

This article was written using Macquarie Capital research prepared by Wendy Pan and Damian Thong in Tokyo. Subscribers can access the related research reports on the Macquarie Insights portal.

  1. Ayonga Hereid (Ohio State University), Why household robot servants are a lot harder to build than robotic vacuums and automated warehouse workers’, The Conversation, 7 September 2022, https://theconversation.com/
  2. 'The freeway to an electric vehicle future', BlackRock, https://www.blackrock.com/
  3. Wendy Pan, Damien Thong, ‘Humanoid robots - Lots of Bots – Opportunities and Challenges’, Macquarie Research, 19 September 2022, https://www.macquarieinsights.com/
  4. Cristina Criddle, James Fontanella-Khan, Antoine Gara, 'Amazon agrees to buy Roomba maker iRobot for $1.7bn', The Financial Times, 5 August 2022, https://www.ft.com/
  5. "Dyson gives tantalising glimpses of secret robot prototypes', Dyson, 25 May 2022, https://www.dyson.com.sg/
  6. 'Humanoid robots are getting close to reality', The Economist, 7 November 2022, https://www.economist.com/
  7. 'Humanoid Robot Market', Markets and Markets, April 2022, https://www.marketsandmarkets.com/
  8. 'Humanoid Robots: Sooner Than You Might Think', Goldman Sachs, 15 November 2022, https://www.goldmansachs.com/
  9. 'Humanoid Robots: Sooner Than You Might Think', Goldman Sachs, 15 November 2022, https://www.goldmansachs.com/
  10. Wendy Pan, Damien Thong, ‘Humanoid robots - Lots of Bots – Opportunities and Challenges’, Macquarie Research, 19 September 2022, https://www.macquarieinsights.com/
  11. https://www.washingtonpost.com/technology/2022/09/30/elon-musk-tesla-bot/
  12. Wendy Pan, Damien Thong, ‘Humanoid robots - Lots of Bots – Opportunities and Challenges’, Macquarie Research, 19 September 2022, https://www.macquarieinsights.com/
  13. Ayonga Hereid (Ohio State University), Why household robot servants are a lot harder to build than robotic vacuums and automated warehouse workers’, The Conversation, 7 September 2022, https://theconversation.com/
  14. Wendy Pan, Damien Thong, ‘Humanoid robots - Lots of Bots – Opportunities and Challenges’, Macquarie Research, 19 September 2022, https://www.macquarieinsights.com/
  15. Wendy Pan, Damien Thong, ‘Humanoid robots - Lots of Bots – Opportunities and Challenges’, Macquarie Research, 19 September 2022, https://www.macquarieinsights.com/
  16. 'Economists are revising their views on robots and jobs', The Economist, 22 January 2022, https://www.economist.com/
  17. Philippe Aghion, Céline Antonin, Simon Bunel, Xavier Jaravel, 'The Direct and Indirect Effects of Automation on Employment: A Survey of the Recent Literature', Harvard University, 2022, https://scholar.harvard.edu/