“Long live artificial intelligence, long live the Republic, and long live France!” proclaimed French President Emmanuel Macron at the International Summit on AI. Is this the new French trinity? The solemnity of the announcement and the majestic setting of the Grand Palais in Paris leave no doubt: AI seems to have become the ultimate grail sought after by the world’s powerful.

Aware of the Old Continent’s lag and pressured by the Americans and the Chinese, Emmanuel Macron called for a European “awakening” in the field of AI. France aims to show that it is backing words with action, promising €109 billion in private investments, notably to develop data centers.

“This is the equivalent for France of what the United States has announced with Stargate,” explained the French president. “Stargate,” a name that echoes Ronald Reagan’s Strategic Defense Initiative (SDI), dubbed “Star Wars,” which at the time pushed the Soviets into an expensive military spiral, contributing to their weakening.

As part of this new technological war, Donald Trump is counting on $500 billion, invested by OpenAI, the Japanese SoftBank, Oracle, and MGX, to secure his “America First” approach in the field of AI.

AI as a geopolitical weapon?

While security and ethics related to AI use were once again central topics in discussions at the Paris summit—and rightfully so—there should be no illusions: no major power intends to fall behind, least of all the Americans, who have traditionally been less concerned with these aspects than the Europeans.

This was, in fact, one of the key messages delivered by J.D. Vance, the U.S. Vice President, who made a lengthy case for AI deregulation, which he sees as the only guarantee for achieving the desired level of innovation.

J.D. Vance, who built his career in Silicon Valley, took on a threatening tone: “It would be a terrible mistake for your countries to tighten the screws on American tech companies”

In a highly aggressive speech, the Silicon Valley veteran took on a threatening tone: “It would be a terrible mistake for your countries to tighten the screws on American tech companies.” Even more striking, he asserted that the United States is the global leader in AI and intends to remain so by “blocking the path of adversaries reaching AI capabilities that could threaten (its) people.”

He did not hesitate to compare AI to weapons, “dangerous in the wrong hands, but incredible tools for freedom and prosperity when placed in the right hands.” These “wrong hands” being “authoritarian regimes.”

A barely veiled reference to China, which swiftly responded to Vance’s remarks. “We oppose ideological categorization, the broadening of the national security concept, and the politicization of economic, trade, and technological issues,” countered Chinese Foreign Ministry spokesperson Guo Jiakun, marking a contrast with the U.S. by declaring China’s commitment to “AI security.”

This Sino-American exchange highlights the tensions and growing American unease. And for good reason! A third disruptive player has recently shaken up the landscape: DeepSeek. Launched by a Chinese startup with a modest $6 million investment—minuscule compared to the massive Western investments—this free, open-source Chinese chatbot has proven to be particularly effective.

From the moment it was announced, it became one of the most downloaded applications on Apple’s App Store, causing a 17% drop in the stock price of graphics chip manufacturer Nvidia. More unexpectedly, it also impacted the shares of major energy companies, whose stocks had surged due to the anticipated high electricity consumption of data centers.

DeepSeek is indeed nearly 30 times less expensive than its most frequently cited competitor, the American ChatGPT

DeepSeek is indeed nearly 30 times less expensive than its most frequently cited competitor, the American ChatGPT. Despite being hit by U.S. export restrictions on high-end chips—supposedly the only ones capable of such data processing—the Chinese have still managed to develop an equally powerful tool. They have even gone a step further by requiring less computing power, thanks to smart optimization of their language model. The result: a highly efficient language model at a fraction of the cost.

Moving towards sovereign AI?

Proof that this Chinese advancement is worrying its competitors, the United States has launched an investigation to determine whether China is bypassing restrictions by acquiring Nvidia chips through intermediaries such as Singapore. After the war against Chinese 5G, are the Americans stepping up their game? Without venturing into value judgments, one conclusion can be drawn without speculation: the absolutely strategic nature of these new AI tools for all major global players is now undeniable.

But what about emerging countries? Are they doomed to be mere spectators while the traditional great powers battle it out with microprocessors and algorithms? One guest at the Paris AI Summit stood out more than the others: Indian Prime Minister Narendra Modi. His co-presidency alongside Emmanuel Macron can be seen as a commitment to France in an effort to counterbalance the influence of the Americans and the Chinese.

A giant in tech outsourcing, a major provider of skilled professionals—particularly in AI—and soon to be the world’s fifth-largest economy, India is advocating for global AI governance. The country is raising alarms about the risks emerging nations face, particularly in terms of job losses, and has announced its own efforts to develop a national AI model while calling for an international regulatory framework to oversee this technology.

But where does sovereignty fit into all of this? Whether it’s ChatGPT, DeepSeek, or Gemini, it’s important to remember that none of these AIs are politically agnostic or free from censorship

Other countries, technologically more modest, now have greater freedom of action, thanks in part to the example set by DeepSeek. The more affordable Chinese tool has demonstrated that it is possible to develop AI models requiring fewer expensive graphics chips, paving the way for smaller organizations to create their own advanced AI models using significantly less costly infrastructure. This shift is comparable in scale to the transition from supercomputers to personal computers, which ultimately led to the global democratization of computing and, eventually, to the smartphones we all carry in our pockets.

But what purpose could these national AIs serve, and where does sovereignty come into play? Whether it’s ChatGPT, DeepSeek, or Gemini, it’s important to remember that none of these AIs are politically agnostic or free from censorship. The diversification and proliferation of AI actors—including both private and public entities from emerging countries—therefore offer the best safeguard for a future where artificial intelligence operates with… as little artificiality as possible.

Artificial intelligence is to humanity in the 2020s what the invention of the printing press was in the mid-15th century. At least, that is what emerges from the various interviews we conducted with Moroccan experts on the subject. But are we truly ready… pic.twitter.com/J946iA1rzg

— TelQuel (@TelQuelOfficiel) March 3, 2025

Written in French by Zakaria Choukrallah, edited in English by Eric Nielson