“If We Stop Trying to Look Like Humans, the Cost of Robots Could Drop to One-Tenth”

When it comes to humanoid robots, opinions generally fall into two categories: “Wow! This is the future!” or “It’s still too expensive to be practical.”

Both perspectives are valid. However, there is a more important question to consider.

“Why do robots even need to take on a human form in the first place?”

The new robot “Eno” from the French startup Genesis AI is intriguing. This robot has abandoned the goal of replicating the human “form” and instead focuses solely on mimicking human “abilities” — such as grasping, transporting, and making judgments.

What does this mean? The costs associated with replicating human-like features — a human-like face, smooth bipedal movement, and precise movements of five fingers — are enormous. This is why Tesla’s Optimus is estimated to cost over $20,000 (around 3 million yen); the cost of “human-like features” is significant.

Eno takes the opposite approach. It doesn’t need to look like a human. It could have wheels or three arms. As long as it can perform the necessary tasks, that’s sufficient. As a result, manufacturing costs could potentially decrease significantly.

This leads us to consider the question: “So, how much should the monthly rental fee be to break even?” Let’s calculate this for three industries: warehousing, agriculture, and nursing care.

—

Warehousing: A Monthly Fee of 150,000 Yen Matches Labor Costs, but the Real Value Lies in “Night Operations”

Let’s take the example of picking operations in a warehouse.

First, let’s break down the current labor costs:

• Hourly wage: 1,100 yen (minimum wage level for 2024)
• Working hours: 8 hours a day, 22 days a month
• Monthly labor cost: approximately 193,600 yen (including social insurance, around 230,000 yen)
• Picking speed: approximately 80 to 120 items per hour (averaging 100 items)
• Monthly processing volume: approximately 17,600 items

Now, let’s introduce a robot. Assuming the picking speed is 1.5 times that of a human (150 items/hour), this is a conservative estimate.

The robot’s strength isn’t just speed; it can operate 24 hours a day, 365 days a year.

• Operating 16 hours a day (after deducting 8 hours for maintenance)
• Operating 30 days a month
• Monthly processing volume: 150 items × 16 hours × 30 days = 72,000 items

In comparison to a human’s monthly 17,600 items, one robot can handle 72,000 items — that’s about four times more.

This means that if the monthly rental fee is four times the labor cost — below approximately 920,000 yen — the robot would win in terms of picking cost per item.

However, let’s think more realistically. Can a small to medium-sized warehouse operator afford 920,000 yen a month? No, they cannot.

Let’s reverse engineer this. If the monthly rental fee is 150,000 yen, it would be cheaper than the labor cost for one person (230,000 yen including social insurance). Moreover, it can operate at night, eliminating the major headache for local warehouse operators during busy periods when “people are hard to find.”

150,000 yen a month. This is the “break-even line” for the warehousing industry.

By the way, there is absolutely no need for it to resemble a human. If it’s moving between shelves in a warehouse to pick items, a wheeled box robot with one arm would suffice. Bipedal movement and a human-like face are unnecessary. This is why costs can be reduced.

—

Agriculture: A Monthly Fee of 100,000 Yen Eliminates “Labor Shortages During Harvest”

Labor shortages in agriculture are severe, especially during the harvest season. Local farmers lament every year that “no one shows up.”

Let’s take a look at the current situation:

• Hourly wage for harvesting work: 1,200 to 1,500 yen (for short-term part-time workers and technical interns)
• Working hours: 6 hours a day, 20 days a month (subject to weather variations)
• Monthly labor cost: approximately 144,000 to 180,000 yen (per person)
• For technical interns, including housing, travel, and management costs, the effective monthly cost is around 250,000 to 300,000 yen.

This is crucial. The labor costs in agriculture are not visible just from the surface hourly wage. Arranging housing, transportation, and fees for managing organizations can lead to effective costs being 1.5 to 2 times the hourly wage.

So, what about robots?

Agricultural robots do not need to take on a human form; in fact, that would be a hindrance. For strawberry harvesting, having a camera and arm positioned low would suffice. For rice harvesting, there is already the combine harvester, which is a “prototype of a robot.”

If we pursue the design philosophy of Genesis AI’s Eno — focusing on the “optimal shape for the task” — the costs of agricultural robots can decrease even further.

Let’s reverse engineer this.

If the effective cost of one technical intern is 250,000 yen, and one robot can handle the workload of two interns, then it can break even at a monthly fee of below 500,000 yen.

However, let’s consider the reality for small farmers. Would a farmer with an annual revenue of 20 to 30 million yen rent a robot for 500,000 yen a month? It’s tough.

100,000 yen a month. This changes the conversation.

For 100,000 yen a month, they could rent it just for the three months of the harvest season. This totals 300,000 yen, which is one-third of the effective cost of hiring one technical intern for three months (750,000 to 900,000 yen). Plus, it eliminates the uncertainty of whether “they will come back next year.”

100,000 yen a month, rented only during peak season. This is the “break-even line” for agriculture.

Again, there is no need for it to resemble a human. Instead, it just needs wheels to run through the fields and a soft gripper that won’t damage the crops. If it were human-shaped, it would actually be more difficult to use in a muddy field.

—

Nursing Care: A Monthly Fee of 250,000 Yen Fills the “Night Shift Gap,” but Full Replacement is Unrealistic

Nursing care has slightly different circumstances.

First, let’s look at the labor costs in nursing care:

• Average monthly salary for nursing staff: approximately 250,000 to 300,000 yen (including night shift allowances)
• Effective cost including social insurance and benefits: approximately 350,000 to 400,000 yen per person
• The cost of securing dedicated night shift staff is even higher.

It is currently unrealistic for robots to “do everything” in nursing care. Bathing assistance, toileting assistance, and communication are tasks that only humans can perform — at least for now.

However, tasks such as “night monitoring,” “mobility assistance,” and “record-keeping” are different.

The biggest burden on night shift staff is the waiting time, which is often filled with the thought, “Nothing may happen, but if it does, I have to respond.” What if we could assign this waiting time to a robot?

• Monthly cost for one night shift staff member (including night shift allowance): approximately 350,000 yen
• The robot handles night monitoring and emergency alerts
• If we could reduce the night shift staff from two to one, the savings would be: 350,000 yen a month

If the rental fee is 250,000 yen a month, that results in a net savings of 100,000 yen.

Moreover, the fundamental issue in the nursing care industry is the inability to recruit staff. Beyond cost reduction, the value of filling the gap of “not having anyone to work night shifts” is significant.

Once again, there is almost no need for the robot to take on a human form. For monitoring, a cluster of cameras and sensors would suffice. For mobility assistance, a powered cart would be adequate. In fact, if it resembles a human too closely, there is a risk of confusing elderly individuals.

250,000 yen a month. If limited to filling the night shift gap, this is the “break-even line” for nursing care.

—

By Reducing the Cost of “Resembling Humans,” We Enter a Price Range Accessible to Small and Medium Enterprises

To summarize:

In all three industries, the conclusion is that “a human form is unnecessary.”

This is not coincidental. The cost of “resembling humans” does not create any value for the field. For warehouse shelves, it doesn’t matter if the robot has a handsome face. For strawberries, there is no need for the harvesting hand to have five fingers.

What Genesis AI’s Eno has demonstrated is the structure that shows, “If we discard human-like features, the cost of robots can drop dramatically.” If Tesla’s Optimus costs 3 million yen, and a robot that discards human-like features can be produced for 500,000 yen, then the monthly rental fee falls within the range of 100,000 to 250,000 yen.

This price range is something that small and medium enterprises in rural areas can consider when saying, “Shall we try renting one unit?”

—

So, What Should We Do?

This is not a call to buy robots immediately.

There are three things that need to be done:

1. Accurately Understand Your Company’s “Hourly Equivalent Cost”
Social insurance, recruitment costs, training costs, turnover costs. The effective cost is 1.5 to 2 times the surface hourly wage. Without understanding this, you cannot compare it with robots.

2. Identify “Tasks That Don’t Need to Be Done by Humans”
There is no need to replace all operations with robots. Night monitoring, simple picking, repetitive harvesting tasks. There are always “tasks that don’t need to be done by humans” in any field.

3. Keep Track of Rental Model Information
Instead of purchasing, consider renting. More models that can be tested with monthly subscriptions will emerge in the future. When functionally specialized robots like Genesis AI’s become available for 100,000 to 250,000 yen a month, be prepared to make a quick decision.

The true revolution of humanoid robots will occur not by “looking like humans,” but by “stopping the effort to look like humans.” Costs will decrease, making them accessible to small and medium enterprises. At that time, companies that are prepared will differentiate themselves from those that are not.

The question is simple. If there were a robot that could work 24 hours a day for 150,000 yen a month at your site, what tasks would you assign it?

The company that has that answer will seize the next five years.