Intel CEO Lip-Bu Tan: “If we’re not careful, China will be ahead of us. DeepSeek was just a wake-up call.”
As Beijing prepares to unveil its next five-year industrial blueprint, Western technology leaders are warning that China’s push to convert AI breakthroughs into state-backed scale could redraw the balance of power.
This week, China’s leadership is expected to publish the outline of its 2026-2030 Five-Year Plan at the National People’s Congress, detailing how it intends to turn advances in artificial intelligence, robotics, space and advanced manufacturing into sustained industrial momentum.
The ambition is no longer limited to headline-grabbing model releases. Policymakers are expected to prioritize “AI-plus manufacturing,” deploy state-owned enterprises as anchor adopters and channel funding into sectors deemed strategically vital.
For Western executives, the message is unsettlingly clear.
Speaking last month at Cisco’s AI summit, Lip-Bu Tan, CEO of Intel, warned that assumptions about Beijing’s constraints may be outdated.
“China is just shortly behind us,” he said. “And if you’re not careful, they will be ahead of us.”
China’s new plan follows a year in which its AI developers stunned global markets. When the startup DeepSeek released a powerful model despite tight U.S. restrictions on advanced chips and lithography equipment, it triggered a reassessment of China’s capacity to innovate under pressure.
“DeepSeek was just a wake-up call,” Tan said.
Analysts expect Beijing to double down on embodied intelligence, the software and hardware systems powering humanoid robotics, as well as autonomous driving and space launch capabilities. The emphasis is increasingly on industrialization, embedding AI across manufacturing, logistics and energy networks at national scale.
Western observers note that China’s advantage may lie not only in engineering talent but in coordination. Approval processes for large infrastructure projects, including energy-intensive AI data centers, can move with striking speed.
“In China, if they decide to have it, they quickly can get all the approvals and get it done,” Tan said, contrasting that pace with the slower regulatory environment in the United States.
Export controls were intended to slow China’s climb. Yet Tan suggested the gap is narrower than many believe. He pointed to the depth of Chinese engineering capacity, including at firms such as Huawei, which he said employs “100 CPU architects, top notch.”
“You will be surprised,” he said, reflecting on how quickly Chinese developers have adapted to restricted access to Western design tools and advanced GPUs.
At the same time, the United States faces structural strains of its own. AI systems are devouring memory and power at unprecedented rates. “In terms of the AI, the biggest challenge… for a lot of my customers is memory,” Tan said. “There’s no relief until 2028.”
Demand for compute remains insatiable. “Almost every CEO, they call me, ‘Lip-Bu, can I have more?’” he said.
Yet capacity constraints and regulatory friction risk slowing deployment just as Beijing is attempting to accelerate it through centralized industrial policy.
“The fundamental research we’re starting to not continue,” Tan warned, expressing concern about declining U.S. investment and the migration of academic talent abroad. On open-source development, he was blunt: “Very knowledgeable people told me we are behind China now.”
China’s forthcoming plan will be scrutinized for how it intends to protect and weaponize supply chains, from rare earths to low-end semiconductors, at a time when technology controls are increasingly intertwined with geopolitics.
“If you’re not careful,” Tan said, “they will be ahead of us.”
Reuters contributed to this report.
