At the Davos event, Elon Musk repeatedly mentioned one word: abundance.
On January 23, 2026, in Davos, during a conversation with BlackRock CEO Larry Fink, Musk condensed all his visions for the future into one word: abundance.
In his view, it's not a pipe dream to enable everyone to have a high standard of living.
How to achieve it?
Next year, the Optimus robot will go on sale; within 3 - 5 years, there will be more robots than humans.
The end result: Everyone will be able to call upon a robot assistant at will, and there will be so many goods and services that "you won't even know what else to ask the robot to do."
But can this path really be successful?
Section 1 | Abundance Is Not a Pipe Dream, Musk Gives a Timeline
During this conversation, Musk said:
In the future, everyone will own a humanoid robot.
This is not a vision for ten years later; it's a product that will go on sale next year. Robots will take care of children, accompany the elderly, do housework, drive cars, and deliver goods. It's no longer just a factory robotic arm; it's a second you.
To what extent will there be abundance?
"There will be so many goods and services that you won't even know what else to ask the robot to do."
In other words, shortages won't last forever.
What's the timeline? He clearly stated:
- By the end of this year, Optimus will complete complex tasks in factories;
- By the end of next year, it will be sold to the public;
- Within 3 - 5 years, there will be more robots than humans.
And it's not just about robots.
AI also has a timeline. Musk predicts that by the end of this year or next year, an AI smarter than any human will emerge; by 2030, AI will be smarter than all humans combined.
What does this mean?
Musk gave a formula:
"Economic output = Average productivity per robot × Number of robots"
When AI becomes smart enough and there are enough robots, the global economy will experience explosive growth.
An ordinary person will also be able to have a robot assistant at their beck and call.
The timeline has been set, the products are being manufactured, and the intelligence curve of AI is rising steeply.
Section 2 | The Real Challenge Is Not Technology, but Electricity
The host asked: Can all this really happen? Are there any limiting factors such as energy, chips, factories, and costs?
Musk gave the same answer again:
"Electricity."
The development of AI is not hindered by the model or the chips; it's hindered by the power supply capacity.
The chip production capacity has been growing exponentially, but the expansion speed of the power grid is only 3 - 4% per year. Musk said, soon there will be a situation where there are chips but no electricity to use. This might happen this year.
How much power is needed?
Musk calculated: A solar field measuring 100 miles by 100 miles can power the entire United States. What does 100 miles by 100 miles mean? It's just a small corner of Utah or Nevada.
What about Europe? The uninhabited areas of Spain and Sicily are enough.
It's not difficult technically, but the reality is: no one is building on this scale. The electricity consumption for AI training has increased tenfold in the past two years, and the gap is getting bigger.
Whoever can solve the power problem can make AI truly a reality.
He cited the example of China: China deploys over 1000 gigawatts of solar power annually. With battery energy storage, it can provide about 250 gigawatts of steady - state power. The average electricity consumption in the United States is 500 gigawatts. That means, just with solar power, China can achieve half of the annual electricity output of the United States.
Electricity has become the real battlefield in the AI competition.
Musk didn't wait for policy changes. Tesla and SpaceX are each building a solar manufacturing capacity of 100 gigawatts per year, which will probably take three years.
This competition has begun.
Section 3 | How Can Robots Be Deployed First When Electricity Is in Short Supply?
It takes 5 - 10 years to expand the power grid, and 2 - 3 years to build a space data center.
So what about now?
Musk's strategy is to get the robots up and running first.
Why is it feasible? Because the technology is already mature.
He said that autonomous driving is essentially a solved problem. Insurance companies have recognized its safety and offer half - price insurance to car owners using FSD. Robotaxis will be widely promoted in the United States by the end of this year.
And this technology can be directly applied to robots.
The vision + neural network technology verified by Tesla in cars can be directly transferred to the Optimus robot. It has the same ability to "understand the world" and the same rapid iteration rhythm.
Moreover, the energy consumption of robots is much lower than that of AI data centers: The electricity consumption of one million Optimus robots is only equivalent to that of a medium - sized data center. But the human labor they can replace is in the millions.
This is why Musk dares to say "go on sale next year": Instead of waiting for sufficient electricity to build robots, he uses the low - energy - consumption characteristic of robots to solve the labor problem first.
The expansion of electricity supply takes ten years, but robots can operate under the existing power grid.
Section 4 | The Energy Source for Abundance Lies Beyond Earth
The host asked Musk: What exactly supports so many AIs, robots, rockets, and factories?
He said: The most ideal place for AI to operate is not on Earth, but in space.
Why?
- There is no weather in space, and the sunlight is stable and abundant.
- Without the atmosphere, the efficiency of solar energy is five times higher than on the ground.
- The shaded side is extremely cold, with a built - in cooling system, eliminating huge cooling costs.
What's the timeline? Within two to three years, the first batch of solar - powered AI data centers will be sent into orbit.
Why can it be so fast?
Because the Starship will achieve full reusability this year. This is the first time in rocket history. Previously, the Falcon 9 could only recover the booster stage, and the upper stage would burn up, which was equivalent to throwing away a medium - sized jet after each flight. Now the entire Starship can be reused.
How much will the cost drop?
It will be less than $100 per pound, cheaper than air freight.
Only when the cost of entering space drops by 100 times can the space data center turn from a concept into a reality.
This rewrites the competition rules of the AI industry: Whoever deploys the space data center first will have all - weather, high - efficiency, and low - cost AI capabilities. It doesn't occupy land, doesn't strain the power grid, covers the globe, and generates its own power.
On the ground, people are competing for the speed of power grid expansion, while in space, there is infinite energy.
This is not just a lead of a few years; it's a change of the playing field.
Conclusion | Curiosity and Optimism
Why is it Musk who is talking about these things?
He said he has a "philosophy of curiosity": He wants to understand the meaning of life and turn science fiction into scientific facts.
From SpaceX to Tesla, from robots to space, he's doing the same thing: Turning the distant future into a calculable timeline.
At the end of the conversation, he said:
It's better to be an optimist even if you're wrong than to be a pessimist even if you're right.
Perhaps this is why he keeps talking about "abundance": not because it will definitely be achieved, but because it's worth believing in.
📮 Original link:
https://www.youtube.com/watch?v=hXb1k59w3M8
https://www.reuters.com/business/davos/musk-speak-davos-thursday-2026-01-22/?utm_source=chatgpt.com
https://www.wired.com/story/elon-musk-sure-made-lots-of-predictions-at-davos?utm_source=chatgpt.com
https://www.businessinsider.com/davos-wef-live-updates-jan-22-2026-1?utm_source=chatgpt.com
This article is from the WeChat official account "AI Deep Researcher", author: AI Deep Researcher, published by 36Kr with authorization.