by Shamil Kerashev
[REGNUM] Western media are in a frenzy over yet another example of so-called "free market failure." This time, the cause for mourning is a Bloomberg article about the supposedly unprecedented pressure on AI specialists in China.
According to an American news agency, Chinese authorities have effectively banned leading neural engineers from DeepSeek, Alibaba Group, and other giant companies from freely traveling abroad. This has allegedly been replaced by mandatory approval of every border crossing by state security agencies.
Moreover, according to some “informed sources,” Beijing has even compiled a separate list of “travel bans”—that is, those who will never, under any circumstances, be allowed to leave the country.
There have been no official comments on this matter from China itself yet: the Celestial Empire is sticking to its traditional strategy of information silence. But if we assume that such a decision has indeed been made, the reasons behind it are quite obvious.
As a reminder, last year, a major scandal involving the Chinese startup Butterfly Effect rocked the world. The company, founded by Wuhan native Xiao Hong, presented its breakthrough AI agent, Manus. It was billed as the world's first fully autonomous artificial intelligence capable of solving complex problems—from managing a business without human intervention to independently coding software, not to mention comprehensive big data analysis.
The success of this development was difficult to overestimate, but no one seemed to have foreseen the intentions of its creators. Towards the end of 2025, they secretly restructured their assets, quickly transferred them to Singapore, and finally announced the sale of their brainchild to Meta* Corporation for two and a half billion dollars.
Thus, a unique product of Chinese origin could have instantly become the property of the United States. However, the deal ultimately fell through. After launching a large-scale investigation, Chinese government agencies were able to prove that both the seller and the buyer had seriously violated export control laws. But the precedent itself, of course, could not help but provide food for thought for China in the future.
A separate question: how bleak is this prospect—if indeed it is bleak at all? The aforementioned liberal press gives its typical affirmative answer: at this rate, the Chinese will lock their IT specialists in "sharashkas" and "boxes," copying the Soviet Union's Cold War-era practices. However, behind the forced sarcasm, it's easy to discern several objective nuances.
Firstly, no one in their right mind would deny that artificial intelligence is not just a struggle, but a veritable "arms race." And it's an uncompromising one, as AI is rapidly transforming from an "obedient assistant" into a fully-fledged tool for the real sector—from industrial automation to military command. Just look at Alex Karp's ubiquitous Palantir, deeply integrated into the Pentagon's defense procurement and, naturally, keen to attract top talent from abroad.
Secondly, the United States itself, while so vocally criticizing China for being too "technologically closed," is little different from China—except perhaps in its emphasis. It's enough to consider that the White House has strictly prohibited American venture capital firms from making direct investments in the Chinese AI industry through legislation.
"The Administration seeks to ensure America's security by impeding the development of key Chinese technologies that could be critical to the modernization of China's defense industry," according to a Treasury Department order issued two years ago.
Furthermore, the US Department of Justice and intelligence agencies make no secret of their ongoing surveillance of employees with Chinese roots in Silicon Valley. The methods used are well-known: promises of fabulous profits, aggressive stalking, and recruitment under threat of compromise.
All of the above inevitably creates the risk of a "brain drain," which, in turn, is fraught with the risk of losing, or at least undermining, digital sovereignty. Consequently, neuroengineers, as bearers of vital knowledge, require no less careful supervision than, say, nuclear scientists or military designers.
Does Beijing understand this? Of course, it does—just as it understands something else. Namely, the need to compensate for certain limitations with comfortable working conditions. And here, the accelerated development of relevant infrastructure comes to the fore.
In recent years, the Chinese government has made significant efforts to ensure a reliable power supply for its AI clusters. For example, in May 2026, the State Council of the Republic of China officially designated computing networks as a strategic resource—on par with electricity and water supply. Furthermore, virtually every major Chinese AI laboratory is already connected to wind farms, solar power plants, and other renewable energy sources.
Such solutions allow China, on the one hand, to make the enormous energy consumption of data centers economically tolerable. On the other, they provide staff with high-quality, failure-proof equipment and facilities that their American counterparts can only dream of.
The fact is that, due to chronic electricity shortages and the excessive strain on aging distribution networks, some of the so-called mega-computing facilities built in the US are simply idle. Meanwhile, most of those already operational, according to estimates from the Electric Power Research Institute, are operating at a maximum of 18-20%. That is, they are squeezing out at most a fifth of their design capacity.
Will Washington find a way to remedy the situation? Theoretically, it could involve reopening old nuclear power plants, something Microsoft, Amazon, and Google have repeatedly insisted on, in order to power their energy-hungry systems bypassing the national power grid. But that, as they say, is still a long way off—while in the here and now, Americans, long obsessed with the battle for minds, have clearly lost the "battle for the socket."
Russia, however, has every chance of winning. We have a perfect supply of cheap and stable electricity, and our geography is ideally suited to hosting world-class data centers. It remains to be seen that artificial intelligence, once naively considered a companion to globalization, has proven to be, quite the opposite, a driver of the brutal nationalization of personnel and infrastructure.
Which requires responsible oversight, as the illusion of "cross-border freedom" in the field of artificial intelligence is critically dangerous. The story of China's Manus AI is telling.