How difficult is it to build an "artificial sun"? Unveiling the core technologies and entrepreneurial opportunities of controlled nuclear fusion.

"AI and nuclear energy utilization are two great things that can change the face of human life in the 21st century." Sam Altman, the CEO of OpenAI, once asserted.

Artificial intelligence and nuclear energy are like two tracks leading to future civilization. One is reshaping the essence of intelligence, while the other is trying to tame the energy from the origin of the universe.

More importantly, these two tracks are converging and collaborating. On the one hand, AI is developing rapidly, and the demand for computing power has skyrocketed, which in turn has led to the problem of energy consumption. Nuclear energy, especially controlled nuclear fusion, is regarded as the key to solving this challenge. On the other hand, AI may contribute to the development of nuclear energy. It can not only control the nuclear reaction system but may also participate in the real - time control of the reaction itself. In June 2025, TerraPower, a nuclear power company, announced the completion of a $650 million financing round, and NVIDIA was one of the new investors.

Uncle Feng once mentioned in a podcast that if we can master controlled nuclear fusion, it means that we will no longer just passively receive the energy radiated by the sun but can actively produce energy. That is to say, humans will become the "sun" themselves.

So, how far is the journey to realize controlled nuclear fusion?

In this industry research, we will focus on the ultimate energy proposition of controlled nuclear fusion, discussing its development difficulties, industrial progress, cross - field impacts, as well as the entrepreneurial and investment opportunities behind it. We hope to provide a new perspective to understand how this technology will affect the future energy landscape, technological ecosystem, and even the direction of civilization.

/ 01 / What is a nuclear reaction?

Both nuclear fusion and nuclear fission essentially release huge amounts of energy through the transformation of elements. Different elements release different levels of energy during the reaction due to their mass differences.

Nuclear fusion, as the name implies, is the process in which light elements such as deuterium and tritium combine to form heavier elements, accompanied by the release of a huge amount of energy. Nuclear fission, such as the fission of uranium - 235, releases slightly less energy but is still one of the highest - energy - density methods currently mastered by humans.

In the early stage of the Big Bang, the entire universe consisted almost entirely of hydrogen. Subsequently, through nuclear reactions inside stars, elements such as helium, carbon, and oxygen emerged, and eventually, more heavy elements were formed. These heavy elements are the basis for the emergence of planets and life - they are mainly born at the end of a star's life cycle.

To achieve a controlled nuclear reaction, the purity and density of materials are crucial. Both fusion and fission require certain conditions to trigger a chain reaction. For example, the uranium enrichment carried out by Iran before was a process to increase the purity of uranium - 235, which is a key step in realizing the utilization of nuclear energy.

During the nuclear reaction process, in addition to generating new elements, various radioactive by - products such as α, β, and γ rays are also produced. These rays carry a large amount of energy. Apart from being finally converted into heat and electricity to become an available energy source for us, their radioactive properties can also be directly utilized to create value, such as being used in the medical field to treat cancer.

In short, the essence of nuclear energy is to release the huge energy contained in atomic nuclei by controlling the transformation of elements. How to efficiently control and utilize this energy is the key direction for the future development of energy technology.

/ 02 / Why is it the right time for controlled nuclear fusion?

On May 23rd, Trump signed a series of executive orders related to nuclear energy. Previously, his policies were more inclined to support traditional fossil fuels. This move has attracted a lot of attention and also raised many questions: Why now? Why skip new energy and go straight to nuclear fusion?

I. Nuclear fusion: A solution to the energy shortage in the AI era

Uncle Feng once mentioned in a podcast that a key background for the development of nuclear fusion is that the United States and the world may have realized that relying solely on traditional energy may not be sufficient to support the technological needs in the next few decades, especially in the face of the huge energy consumption problem caused by the rapid development of AI and the soaring demand for computing power.

Before Trump's nuclear energy executive order was issued, Elon Musk pointed out in a media interview that AI development would face a power bottleneck next year, and China's power infrastructure far surpasses that of the United States. Vicki Hollub, the CEO of Occidental Petroleum in the United States, once mentioned that about 97% of the world's currently produced oil was discovered in the 20th century. Since the world cannot update the existing crude oil reserves quickly enough, the oil market will face a supply shortage by the end of 2025.

In this context, the United States' choice has become clear. On the one hand, China has taken the lead in the new energy field. On the other hand, the conversion efficiency and sustainability of traditional energy are limited. Therefore, the United States may skip new energy and aim directly at controlled nuclear fusion as a solution.

Of course, this policy shift is not set in stone. Due to the influence of the two - party politics in the United States, many policy directions are subject to change. For example, the subsidy content in the chip bill signed during the Biden era was later reduced or even cancelled. Therefore, the public may still be on the sidelines regarding the sustainability of this nuclear fusion support policy.

II. Humans can view problems beyond the solar system

The energy that humans have been using has essentially been in the form of solar energy conversion. Whether it is coal, oil, wind energy, hydropower, or photovoltaics, these energies are essentially converted from solar energy, only with different conversion difficulties and application scopes.

Uncle Feng believes that if we can master controlled nuclear fusion, it means that we will no longer just passively receive the energy radiated by the sun but can actively produce energy. That is to say, humans will become the "sun" themselves.

By then, humans will have the ability to view problems beyond the solar system and move towards an interstellar civilization.

Looking back at history, every developed civilization at each stage was based on strong engineering capabilities. For example, China built the Great Wall, Egypt built the pyramids, and the United States built ships in large numbers during World War II.

In this round of nuclear industry competition, China may be one of the most promising countries. Because we can "achieve miracles through great efforts" and use our strong organizational and execution capabilities to promote the implementation of complex systems.

Developing controlled nuclear fusion is not simply about competing in population or resources. It requires strong organizational capabilities, engineering capabilities, scientific research capabilities, design capabilities, and management capabilities, etc. The whole process is like a precise chain - every link must be reliable; otherwise, the entire system will break down. The lack of core technology is often not because of a single weak point but because the whole chain is broken.

/ 03 / "Cooking dumplings in a paper pot" - How difficult is nuclear fusion?

Nuclear fusion bears the ultimate dream of human energy and civilization, but the difficulty of achieving it is beyond imagination.

The difficulty of achieving nuclear fusion is comparable to "cooking dumplings in a paper pot." The whole process must be controlled extremely precisely; otherwise, all previous efforts will be in vain. We need to make the water boil exactly (maintain a plasma at hundreds of millions of degrees Celsius), prevent the pot from breaking (existing materials are very fragile during the nuclear fusion process), and continuously add new dumplings (inject fuel).

I. Core elements for achieving nuclear fusion

The essence of nuclear fusion is to make light atomic nuclei (such as deuterium and tritium) fuse under extreme conditions, releasing a huge amount of energy. To achieve this, three core conditions are required: a high enough particle density, an extremely high temperature (usually up to hundreds of millions of degrees Celsius), and a long enough confinement time. The product of these three elements is called the "triple product" in physics. Only when this value is large enough can the fusion really "ignite."

In addition, there is a often - overlooked factor: volume. Since volume is the cube of the radius while the surface area is the square of the radius, as long as the volume is large enough, even if the unit energy output is not high, an impressive overall power can be accumulated.

Another key parameter to measure the practicality of nuclear fusion is the Q - value, which is the ratio of the output energy to the input energy required to maintain the fusion state.

Currently, people have been able to make the Q - value greater than 1, that is, the output energy is greater than the input energy. But this is just the first step. The real challenge lies in how to continuously and stably generate electricity, rather than just conducting a single - time, short - term experiment.

II. Main technical paths for nuclear fusion

Currently, the main technical paths for nuclear fusion can be roughly divided into two categories: inertial confinement and magnetic confinement.

The main technical routes of the first inertial confinement type include:

• Laser ignition: Using the shock wave of a laser to make a fuel pellet, usually containing deuterium and tritium, reach extremely high temperatures and pressures to trigger a nuclear fusion reaction;
• Field - reversed configuration: Directly using the fusion reaction to generate electricity instead of indirectly generating electricity by heating a fluid or driving a turbine;
• Z - pinch: Using a strong magnetic field to confine and pinch the plasma to achieve nuclear fusion.

The principle of these methods is to input a large amount of energy in a short time to compress the fuel. The advantage is that extremely high parameter conditions can be achieved in a short time.

For example, the National Ignition Facility (NIF) in the United States focuses on laser ignition, and Helion Energy, an American energy company, has achieved high parameters in the field - reversed configuration. However, the challenge of this route is that it can only generate one pulse each time, and it is difficult to achieve continuous discharge and energy utilization.

The second magnetic confinement type is currently the technical route closest to achieving continuous operation, including various forms such as tokamaks, stellarators, and magnetic mirrors. Although these forms have different structures, their basic principle is the same: confining the high - temperature plasma in a closed space through magnetic field configuration.

China's Experimental Advanced Superconducting Tokamak (EAST), the International Thermonuclear Experimental Reactor (ITER), and the Joint European Torus (JET) are all important representatives of this route.

Among them, ITER is currently the world's largest tokamak device, weighing 23,000 tons and nearly 30 meters high, equivalent to a ten - story building.

Construction site of ITER. Image source: ITER

How does a huge system like a tokamak start? We can briefly sort out the process:

First, the toroidal coils are energized to generate a strong toroidal magnetic field. Then, hydrogen isotope gas is injected and ionized through discharge to form a plasma. Charged particles rotate around the magnetic field lines in the magnetic field. However, this is not stable enough, and the plasma will expand outward. So, scientists add an axial magnetic field to shrink the "doughnut" inward. Finally, a vertical magnetic field is added to form a mirror compensation effect, with the ultimate goal of minimizing plasma escape. During this process, the plasma needs to be continuously heated to break through the fusion threshold temperature, and the magnetic field needs to be dynamically adjusted to maintain the stable confinement of the system.

Does it sound not very complicated? But the problem is that the magnetic field has a characteristic: it has no starting or ending point and is always a closed loop. In contrast, the electric field is divergent. The interaction between these two fields makes it impossible for us to achieve completely stable plasma confinement only by electromagnetic force. Therefore, scientists have been working hard to "play a game" - how to precisely control the magnetic field to minimize plasma escape.

We can use the "prison guard model" to understand the difficulty of controlling the magnetic field. The plasma is like prisoners in a prison who always want to escape. We hope to manage them with the least amount of manpower (control energy). However, these "prisoners" are charged, repel each other, and are affected by the magnetic field, making their dynamics extremely complex. What's more troublesome is that we can't even track every "escaper" in real - time and can only regulate them through macroscopic means. It's like building a prison where we hope the number of guards is far less than that of prisoners but still prevent them from escaping en masse.

In addition, simulation is a huge challenge. It is very difficult to fully simulate the boiling process of a pot of water, and the complexity of controlling the plasma far exceeds this. Every small disturbance may lead to a loss of control. For example, a sudden increase in local density in one place will trigger a chain reaction, similar to the spread of water waves, oscillating repeatedly in the entire system.

Therefore, nuclear fusion is not only a challenge to the physical limits but also an ultimate test of human engineering capabilities. In terms of materials, we need structures that can withstand extreme temperatures and radiation. In terms of the control system, we must achieve real - time and efficient monitoring and adjustment. In terms of engineering integration, multiple subsystems need to be integrated into a controllable and sustainable energy device.

/ 04 / Progress of the nuclear fusion industry

Despite the numerous challenges in the development of nuclear fusion, human exploration of it has never stopped. Currently, what stage has nuclear fusion development reached?

I. Exponential growth in nuclear fusion research

Since the 1960s, research on nuclear fusion has been continuously advancing globally.

According to the summary and statistics of Anthony J Webster from the UK Atomic Energy Authority, from the 1960s to the early 2000s, the triple product (the product of the three elements of nuclear fusion) approximately doubled every 1.8 years. Its growth rate (the purple line in the above figure) was faster than Moore's Law, which doubles every 2 years (the red line in the above figure), and also faster than that of particle accelerators, another star device in physical research, which doubles in energy level every three years (the green line in the above figure).

Behind this exponential growth is the result of collaborative breakthroughs in multiple technical fields such as system control, materials science, and structural design.

II. Improvement in human control ability over complex systems

If nuclear fusion is a precise physical experiment, it is also a challenge to the limits of the control system.

In the modern chip manufacturing process, plasma etching technology is widely used, which has allowed engineers to accumulate rich experience in plasma operation. In the field of nuclear fusion, this experience is being used to build a more precise and responsive control system.

A typical example is the dynamic control of magnet coils. Limited by the superconducting properties or physical parameters of the coils, rapid adjustment has always been a technical challenge. However, with the rapid development of power electronic devices, we can quickly adjust the coil current, making the entire magnetic field system more flexible and controllable. In other words, a system that was originally like an "incandescent lamp" that could only be adjusted slowly has become an "LED lamp" that can change rapidly.

This improvement in control ability means that people can not only initiate a nuclear fusion reaction but also intervene in the intermediate process in real - time to prevent the plasma from getting out of control, greatly enhancing the controllability of the system and enabling the reaction to run stably.

III. Magnets are the core of nuclear fusion, and China is rising in the nuclear fusion field