Cheap Chinese AI quietly takes over the global market

While Silicon Valley is spending hundreds of billions of dollars on extravagant AI models, China is quietly conquering the globe by employing a strategy of "distilling" technology to make AI affordable. 

The tech and investment world is witnessing two contrasting yet accurate portrayals of the overall landscape of artificial intelligence (AI). In the US, Anthropic – the creator of the hugely popular chatbot Claude – is angrily accusing three Chinese labs of creating thousands of fake accounts to secretly poach intellectual property from their system. Conversely, ByteDance is nonchalantly launching Seedance 2.0, a tool for creating smooth, Hollywood-quality videos from text at surprisingly low costs.

Leaving aside copyright disputes, from a financial perspective, this is a declaration of war: America may possess the smartest and most expensive machines, but China holds a far more dangerous weapon: transforming AI, a luxury, into an affordable commodity, infiltrating every corner of the global economy .

The art of "distillation" and the strategy of economies of scale over costs.

In business, the leader isn't necessarily the one who creates the most innovative product, but rather the one who can bring that product to the most people at the lowest cost. Instead of spending billions to build massive programming language models from scratch, Chinese developers are employing a sophisticated technique called "distillation." Simply put, they use their compact models to "learn" the output from advanced Western systems, then compress that vast amount of knowledge into a lighter version.

This approach offers a competitive price advantage. "Distilled" models don't require the most advanced chip arrays or exorbitant operating costs. They can be offered to businesses and general consumers at extremely affordable prices.

According to Anil Agarwal, CEO of InCruiter, industrial-scale distillation is no longer an academic curiosity, but a lever for accelerating competition. He emphasized that in AI, training from scratch is extremely expensive, and distillation helps mitigate that burden, clearly reflecting the "scalability first" doctrine.

Beijing's approach perfectly matches the formula for success it has applied in the telecommunications and digital payments industries: Global expansion with solutions significantly cheaper than Western competitors, thereby creating long-term technological dependence. These inexpensive models may not solve quantum physics problems, but they are perfectly suited for the commercial, logistical, and administrative tasks that millions of businesses worldwide are eager to implement every day.

Wall Street's $700 billion dilemma and the ecosystem shock.

China's shift toward low-cost AI is exposing a critical weakness in the investment model of American tech giants. According to CNBC, the Big Four—Amazon, Microsoft, Meta, and Alphabet—have outlined plans to spend up to $700 billion on AI infrastructure this year alone.

However, this massive influx of money is reigniting Wall Street's obsession with return on investment (ROI), which once caused the market capitalization of major software companies to evaporate by as much as $1 trillion during a short-term sell-off.

Karim Moussalem, Chief Investment Officer of Selwood Asset Management, frankly stated that the market is beginning to lose faith in "American exceptionalism." A lot of money is being thrown into the race, but the amount of capital spent far exceeds previous forecasts, while questions about actual profitability are growing.

The US is building pioneering capabilities and grappling with regulatory hurdles, while China is opting for massive deployment, collecting data from millions of users to continuously optimize. This feedback loop on a super-large scale creates tremendous momentum.

Rory Green, chief economist for China at TS Lombard, warns that the "technological shock" from the world's largest economy has only just begun and is directly threatening the US's AI monopoly.

Beijing is crafting the perfect formula: world-class technology from developed markets combined with the ultra-low costs of a massive supply chain. Backed by an $8.69 billion national AI fund and the "AI+" initiative, this strategy has the potential to create a global "China technology zone."

Mr. Green outlines a realistic scenario: In the next 5-10 years, the majority of developing countries will opt for China's inexpensive ecosystem, from 5G and energy batteries to AI, instead of expensive solutions from the US and Europe. When a price becomes too attractive, the concept of technological frontiers will be blurred by immediate economic benefits.

When AI enters the factory, it will reshape the structure of the economy.

Beijing's AI race isn't just hovering on social media platforms; it's deeply rooted in the real economy. At the National People's Congress (NPC) meeting in early March, Chinese leaders are expected to unveil the 15th Five-Year Plan (2026-2030), outlining a roadmap for transforming AI breakthroughs into industrial-scale technology. At the heart of this vision is the "AI + manufacturing" strategy and, in particular, "embodied intelligence"—the technology behind humanoid robots.

In reality, China's hardware prowess is creating a solid foundation. Recently, on national television, domestically produced humanoid robots demonstrated exceptional balance and agility. However, the large-scale integration of AI into manufacturing will create seismic shifts in corporate restructuring. According to expert analysis, this strategy will be led by giant state-owned corporations – entities with the capital-intensive resources to absorb the infrastructure deployment costs.

Shin Nakamura, President of Daiwa Steel Tube Industries, predicts a strong differentiation in the industrial structure. The gap between large corporations and small and medium-sized enterprises (SMEs) will widen, triggering a fierce wave of mergers and acquisitions across the Chinese market, as smaller companies lack the resources to keep up with technology integration.
Meanwhile, hardware constraints imposed by US export controls on advanced chips don't seem to be deterring the world's second-largest economy. Instead of giving up, Chinese businesses are becoming increasingly adept at optimizing algorithms to run on older chips or leveraging Huawei's massive chip clusters combined with cheap power sources.

This extreme adaptation demonstrates the philosophy that large-scale distribution and operational efficiency are sometimes more decisive than a slight theoretical technical superiority.

For the world, especially developing countries, the rise of the "Chinese AI model" presents a perplexing dilemma. On the one hand, low-cost tools enable businesses to rapidly transform digitally, solving the problem of tight budgets. But on the other hand, as the Freedom House Institute notes, it comes with a closed system of information control and management. Nevertheless, the allure of the economic solution remains irresistible.

The gap in technological prowess is narrowing at a dizzying pace, as Google DeepMind CEO Demis Hassabis once admitted. Chinese AI models are now only a few months behind their Western counterparts, a margin too fragile for the United States to rest easy on.

Source: Dan Tri Newspaper