On Friday evening, China Central Television’s most-watched news program broadcast footage that had never been shown publicly. The South China Morning Post reports that the Chip Fundamental Technology Research Laboratory at Huawei’s Lianqiu Lake campus in Shanghai appeared on Xinwen Lianbo, the state-run bulletin that reaches more than 200 million viewers nightly. The segment showed Huawei founder Ren Zhengfei hosting Vice-Premier Ding Xuexiang at the facility. The broadcast did not disclose what the laboratory is developing. It did not need to. The timing said everything the technical details did not.
President Donald Trump arrives in Beijing on Tuesday for a three-day state visit, his first trip to China in his second term. He is bringing a delegation of high-profile American executives, including Elon Musk and Tim Cook. The agenda includes trade, the Iran war, Taiwan, and semiconductors. Beijing chose to put Huawei’s most secretive chip research facility on national television exactly 48 hours before the American delegation lands. As noted by The Next Web, the audience was not the Chinese public. The audience was the delegation.
The Scale of Lianqiu Lake
Lianqiu Lake is Huawei’s largest research and development center in the world. The campus covers 2,600 acres in Jinze, a town in Shanghai’s Qingpu district, and cost 10 billion yuan (roughly $1.4 billion) to build. Construction took three years. The facility comprises eight blocks, 104 buildings, more than 40,000 offices, and an internal railway system. It is larger than Apple Park and Microsoft’s Redmond campus combined. Huawei expects it to house 35,000 researchers working on semiconductors, wireless networks, smartphones, autonomous vehicles, and energy systems.
The chip laboratory sits inside this massive complex. Its existence was known to intelligence analysts and semiconductor industry watchers but had never been shown on Chinese state media. Ren Zhengfei, who rarely appears in public, personally guided the vice-premier through the facility. Ding Xuexiang oversees China’s science and technology policy. The visit was not merely an inspection; it was a highly visible state endorsement of Huawei’s role as the national champion of semiconductor independence.
(Related: US Lawmakers Target DUV Lithography, Moving to Close the Last Major Chip Export Loophole to China)
The Silicon Behind the Signal
Huawei has been on the U.S. trade blacklist since 2019, barred from buying advanced chips or the equipment to make them. The export controls were designed to cripple the company’s ability to produce competitive semiconductors. Seven years later, Huawei’s AI chip revenue is projected to reach $12 billion in 2026, a 60 percent increase from 2025. The company is targeting 1.6 million Ascend dies across its product line this year.
The Ascend 910C, fabricated by SMIC on a 7-nanometer process that does not use extreme ultraviolet (EUV) lithography, delivers roughly 60 percent of the inference performance of Nvidia’s H100. The newer Ascend 920, built on SMIC’s 6-nanometer node, produces 900 teraflops and four terabytes per second of memory bandwidth. The Ascend 950PR entered mass production in March 2026. Huawei has orders for nearly 800,000 of the new chips this year, on top of the similar volume of older chips it already planned to ship.
When DeepSeek released its V4 models in late April, Huawei’s newest Ascend chips received day-zero adaptation. DeepSeek spent months rewriting its core code to work with Huawei’s CANN framework, moving away from the Nvidia CUDA ecosystem that has underpinned AI development for two decades. The chip lab that appeared on CCTV is not producing theoretical research. It is producing the silicon that runs China’s most advanced AI models.
A Negotiating Tactic Disguised as News
Huawei spent 96.9 billion yuan on research and development in the first half of 2025, representing 22.7 percent of revenue, a record proportion that caused net profit to fall 32 percent. The company has invested in more than 60 Chinese semiconductor firms through Hubble, a wholly owned investment platform established in 2019, the same year the U.S. blacklisted it. Ren Zhengfei has stated he is leading a network of more than 2,000 Chinese companies working toward 70 percent semiconductor self-sufficiency across the entire value chain by 2028.
The R&D intensity is extraordinary by any measure. Huawei is spending more than a fifth of its revenue on chip research, even though it cannot access the world’s most advanced manufacturing tools. SMIC’s most capable process, the 7-nanometer node used for the Ascend 910C, was achieved by repurposing older deep-ultraviolet lithography equipment in ways the machines were not designed to operate. Yield rates are lower than TSMC’s. Costs are higher. The chips are less powerful than Nvidia’s latest generation. None of that matters to the message Beijing broadcast on Friday night.
The broadcast is a negotiating tactic disguised as a news segment. China’s semiconductor industry remains behind the leading edge. The performance gap between Huawei’s best chip and Nvidia’s best chip is measured in generations, not increments. But the strategic question that Trump’s delegation will confront in Beijing is not whether China can match American semiconductor technology. It is whether the export controls designed to prevent China from trying have instead consolidated an entire national semiconductor ecosystem under a single company that the U.S. government cannot reach.
(Related: DeepSeek V4 Goes Fully Open Source; Huawei Confirms Ascend Chip Compatibility in Separate Statement)
Nvidia’s Jensen Huang has previously called the prospect of DeepSeek running on Huawei chips a “horrible outcome” for America. That outcome is no longer a prospect. It is the configuration that produced China’s most powerful open-source AI model. The chip lab that was not supposed to exist is building the hardware that runs the AI that competes with the AI that runs on the chips that China is not supposed to have.