How much should we invest in humanoid robotics?

to use-min

By Adam Dorr

What if it were possible for a country to increase the per-capita productivity of its economy by a factor of 2, 3, 5 or even 10 in the span of just a decade? What level of investment would that kind of result justify?

Committing just a single-digit percentage of GDP to the deployment of humanoid robotics would quickly yield a superabundance of goods and services, resulting in a staggering level of overall prosperity.

Investing in robot labor could be one the greatest uses of capital in human history, outperforming similar investments in electricity, running water and education.

Any investment decision is shaped by three basic considerations:

  1. What are the expected returns?
  2. What are the available alternatives?
  3. What is the cost of inaction?

Before we consider each of these questions, let’s set the stage with a reminder of some of the striking facts about humanoid robots that we’ve noted in our previous posts.

  • They will be manufacturable at less than $20,000 per unit at scale, compared to the cost of $300,000 or more of raising a human child in a wealthy country.
  • They will be built in days or hours, not 18+ years.
  • They can be trained and retrained instantly via software download, not over days or weeks or months like humans.
  • They will work up to three times more hours per day than humans.
  • They will require only a tiny fraction of supporting workplace expenditures that humans require (e.g., bathrooms, food, comfortable temperature, noise, air quality, parking spaces, HR and training, insurance, healthcare, child care, pensions, stock options, etc.).
  • They can be redeployed instantly to perform any new task without the social or administrative friction of shifting roles or responsibilities.
  • They will perform at a totally consistent level of quality, never compromised by boredom, fatigue, illness, drugs, etc.

What are the expected returns on investment in humanoid robotics?

Return on investment (ROI) are typically measured financially. But money is only a proxy for the benefits we are really concerned about here, which is the material prosperity created by the production of goods and services. For all its faults and limitations, GDP is a well-established indicator that measures productivity in terms of dollars, so we’ll use it here.

Consider the returns on investment in human workers in terms of GDP.

According to the National Center for Education Statistics, the United States spends almost $1 trillion per year on primary and secondary education and an additional $700 billion on post-secondary education at colleges and universities. The country then spends $100 billion on workplace training. This $1.8 trillion dollar annual investment in the U.S. workforce results in a GDP of $27.4 trillion. In greatly simplified terms, it could be considered a 1,500% ROI. Even though almost 40% of U.S. GDP returns to capital rather than to labor, the general relationship between total investment in a society’s education and training (i.e., its human capital) and the nation’s total resulting productive capability clearly holds.

The U.S. civilian labor force is currently estimated to be about 168 million people. If we take the prevailing estimate that labor’s share of GDP in the U.S. is 60%, then each working person in the labor force contributes an average of $98,000 to the country’s GDP. (About half of the population participates in the workforce.)

OK, so what about humanoid robots?

Here is a simplified ROI calculator that does a bit of arithmetic based on a handful of parameters and puts the results into the context of today’s GDP—in other words, how productive the economy of the U.S. is in 2024.

Simple Humanoid Robotics ROI Calculator
Cost per Year per Robot: $0
Total Investment per Year: $0 billion
Robot Boost to GDP: $0 billion
New GDP: $0 billion
Investment as % of New GDP: 0%
GDP Increase: 0%
Robotics ROI to Productivity: 0%

The calculator lets us dial in assumptions about how much robots will cost, how long they will last, and how productive they will be relative to humans. Then, based on how many millions of robots we want to add to the workforce, the widget calculates the total cost and returns on that investment in robots.

This tool is not intended to offer a rigorous analysis but simply to illustrate the ballpark picture. A quick play with the sliders shows that even if bots are fairly expensive (say, $20,000 each over their lifetime) and only last four years before needing to be recycled, the U.S. could nevertheless effectively double the size of its productive workforce for $840 billion per year—and in doing so boost its GDP by 60% up to $44 trillion. That’s obviously a huge amount of money to invest, but it’s less than half of what the U.S. currently invests in human education and training.

At scale and maturity, we ought to expect humanoid robots to become much cheaper, last longer and become considerably more productive per hour than human beings. When we dial in a cost of $5,000 each, a lifetime of six years and twice the productivity of a human being, we see we could triple the effective size of the workforce with 336 million robots by investing just $280 billion each year. In doing so, this huge increase in productive capacity would boost GDP by 240%. Again, that $280 billion isn’t a small number, but it’s just 0.3% of the resulting new GDP of $93 trillion. Put another way, $280 billion invested in humanoid robots yields $66 trillion in new productivity, for an ROI of 23,500%.

That’s an economy with one bot for every person.

Now imagine 2, 5 or even 10 robots for every person—never resting, never tiring, just working, producing. Go ahead and dial it into the calculator.

Committing a single-digit percentage of GDP to deploying humanoid robotics would result in a superabundance of goods, services and overall prosperity that is utterly staggering.

What are the available alternatives?

It’s difficult to imagine a better investment that societies can make over the next two decades than deploying as many humanoid robots as possible. However, other investments will also be required to support robots—especially in energy. A rough calculation—assuming each robot consumes 10 kilowatt-hours of electricity per day—shows that 100 million robots would need about 400 terawatt-hours of electricity each year, or roughly 10% of today’s total electricity consumption.

It’s tempting to think that societies must be cautious not to make national investments in robots that might draw resources away from other critical needs such as healthcare, education, and other social goods and services. But this zero-sum thinking is mistaken. The productivity gains from investment in humanoid robotics will directly translate into greater resources available for those other critical goods and services. Indeed, investing in humanoid robotics may well be the very best way to increase resources available for healthcare, education and so on.

The real question nations must ask themselves is: “Is there an alternative investment that could double or triple our GDP over the next 15 years?”

What are the costs and/or consequences of inaction?

GDP is a flawed measure of productivity. GDP and the productivity it represents have also been widely criticized as a faulty measure of human wellbeing. Nevertheless, it remains an incontrovertible fact that virtually all social indicators of wellbeing are strongly correlated with a society’s capacity to produce goods and services. One of the best things we can do to give everyone a better quality of life is to make society more economically productive. With enough prosperity, virtually every problem is solvable—even our most formidable challenges, like climate change.

The cost of failing to invest maximally in humanoid robotics would therefore be monumental, as measured in lost opportunity to increase productivity, prosperity and wellbeing—and the problem-solving capacity they represent. Rational policymaking ought to avoid such a disastrous outcome at all costs.

Furthermore, countries that fail to invest in humanoid robotics will also fall behind those that do—becoming losers in the race to comparative prosperity, and very likely faster than at any other time in human history. If a country like the U.S. were to fail to invest adequately in robotics, for example, it could find itself in 2040 with a GDP one-fifth the size of a rival that didn’t make the same mistake.

How much should countries invest in humanoid robotics?

The ROI on this technology is almost without precedent in human history—certainly comparable to the adoption of electricity, running water and literacy.

The bottom line is that every country needs to go all-in on humanoid robotics deployment, investing every penny they can muster, starting right now. The stakes are enormous and the returns on offer are so staggering as to nearly defy description.

 

This is the third blog in our series on humanoid robotics.

Read part 1: This time we are the horses

Read part 2: Near-zero cost labor: the disruptive economics of humanoid robots

Read more blogs on this topic

Sign up to our Newsletter