Cuộc đua công nghệ AI giữa Trung Quốc và Mỹ với hai chiến lược đối lập
• Trung Quốc đang triển khai hành lang điện toán phủ sóng 99% dân số, trải dài từ vùng duyên hải phát triển đến sa mạc Gobi, biên giới Siberia và Tibet
• Đến năm 2030, các trung tâm sẽ được kết nối bằng cáp quang tốc độ cao, cho phép các startup ở thành phố nhỏ 500.000 dân có thể xử lý AI với độ trễ dưới 3 mili giây
• Ngược lại, Mỹ tập trung 70% trung tâm dữ liệu thế giới tại Virginia phía Bắc, nơi các gã khổng lồ công nghệ như Microsoft, Google và Meta đặt trụ sở
• Chiến lược phân tán của Trung Quốc nhằm:
- Đảm bảo công bằng và giảm chênh lệch kinh tế giữa các vùng
- Tối ưu hóa việc sử dụng năng lượng xanh
- Tăng cường an toàn và giảm rủi ro địa chính trị
• Năm 2023, Trung Quốc xây dựng hơn 40.000km lưới điện cao thế, trong khi Mỹ chỉ xây dựng chưa đến 1% con số này
• Mỹ chiếm 32% công suất điện toán toàn cầu, đứng đầu thế giới. Trung Quốc đứng thứ 2 với 26%
• Tính đến tháng 6/2024, độ trễ truyền dữ liệu giữa đông-tây Trung Quốc đã giảm xuống 20 mili giây
📌 Trung Quốc đang thực hiện chiến lược phân tán hành lang điện toán phủ sóng 99% dân số, đối lập với Mỹ tập trung 70% trung tâm dữ liệu tại Virginia. Mỹ dẫn đầu với 32% công suất điện toán toàn cầu, Trung Quốc theo sau với 26%.
https://www.scmp.com/news/china/science/article/3283901/how-china-and-united-states-have-parted-ways-ai-power-race
How China and the United States have parted ways in AI power race
China’s hi-tech corridor is designed to send computer signals across the country, while America has gone in a different direction
As China embraces the coming AI era, a programme to set up a computing power corridor across the country will cover 99 per cent of the nation’s population, according to a key architect of the project.
It is the complete opposite of the United States’ approach.
In the US, most AI computing facilities are being built in northern Virginia, an area which already boasts 70 per cent of the world’s data centres.
Meanwhile China’s corridor, launched earlier this month, has facilities distributed across a wide region, from economically developed coastal areas to the western Gobi Desert, the northern borders to Siberia, and even taking in Tibet.
By 2030, these centres will be linked by high-speed optical fibres, forming a unified network. Even in a smaller city of around 500,000 people, a start-up will be able to leverage a nearby massive computing cluster to process AI tasks with a latency of under three milliseconds – faster than the refresh rate of a smartphone screen.
This approach is obviously more expensive and less convenient than building a centralised hub.
In an article in the journal E-Governance on October 21, Yu Shiyang, director of the Big Data Development Department at the State Information Centre of China, explained the rationale behind this strategy.
First, there is the idea of fairness. The AI revolution risks exacerbating the concentration of wealth in the hands of a few. Northern Virginia is already one of the wealthiest regions in America. Of the seven US counties with the highest household income, four are in this area.
Advertisement
“Most of the ultra-large data centres are concentrated in northern Virginia, where tech giants like Microsoft, Google and Meta have set up their bases,” Yu wrote.
China also grapples with economic imbalance, with eastern regions richer than the west.
“Optimising the distribution of computing resources, fostering a balanced digital industry layout and coordinating east-west industrial development can unlock new innovation and growth opportunities in vast regions like the west and northeast,” Yu said.
The second reason is efficiency.
Overly concentrated data centres hinder efficient energy use, particularly green power.
Due to energy shortages, Microsoft even plans to reopen the Three Mile Island nuclear plant not far from Northern Virginia, despite the nuclear meltdown which took place there in 1979.
According to a map in Yu’s article, China’s computing power corridor aligns closely with its ultra-high-voltage transmission network.
Advertisement
This ensures abundant power supply, including wind and solar energy from the Gobi and other deserts.
Some US industry insiders envy this.
“There are calls from them to learn from China,” Yu wrote.
Higher voltage means longer electricity transmission. Last year, China built more than 40,000km of high-voltage grids, some reaching 1,100 kilovolts in capacity. In contrast, the US built less than 1 per cent of that length, with a maximum voltage of 345 kilovolts.
Decentralised facilities could also be also safer, Yu said.
Northern Virginia’s proximity to the ocean poses risks. Some US security experts warn of potential destruction from natural disasters or attacks.
China has chosen lower-risk western regions as a strategic hinterland for its computing power corridor, according to Yu.
“Building data centres in strategically deep regions like Guizhou, Xinjiang and Tibet, which are remote and far from economic centres, reduces geopolitical security risks and enhances resilience and risk resistance in extreme situations,” he wrote.
China is ramping up R&D efforts to bring this megaproject to life.
Last year, the United States accounted for 32 per cent of global computing power, ranking first worldwide.
China ranked second with a share of around 26 per cent, in part due to US sanctions.
Beijing hosts World Robot Conference as China eyes expanding humanoid robot market
But China’s AI chip manufacturing is rapidly growing, thanks to the efforts of hi-tech companies such as Huawei Technologies.
These companies are also breaking world records in long-distance, large-capacity data transmission.
By June this year, data exchange latency between China’s east and west was reduced to 20 milliseconds, supporting large-scale AI training and task processing, according to Yu.
This has allowed Chinese companies to surpass the US in certain commercial applications.
For instance, while OpenAI’s Sora remains in the lab, some Chinese companies already offer similar text-to-video services to global users.
The impact of China’s computing power corridor will be felt around the world, according to Yu.
“High-capacity computing channels will be extended to countries and regions along the ‘belt and road’ in the future,” he wrote, referring to China’s Belt and Road Initiative.
“We will fully leverage the regional advantages of Xinjiang, Qinghai, Gansu, Inner Mongolia and other regions to export computing resources to Central Asia, West Asia and the Middle East.
“We will also promote regions such as Chongqing, Guizhou, Yunnan, Guangdong, Guangxi, Hainan and others to provide computing supply services to South Asia and Southeast Asia, and guide northeastern regions such as Heilongjiang to export computing capabilities to Northeast Asia.”
Thảo luận
Follow Us
Tin phổ biến
TAG
AI giáo dục
AI sinh-y-duoc
AI nghệ thuật
AI pháp lý-quản trị-chủ quyền
AI models
AI xã hội
AI prompts
AI kiến thức-khóa học
AI video
AI công nghiệp-lĩnh vực
AI edge
AI viễn thông
AI tools
AI chính phủ
AI GPT Store
AI cybersecurity
AI so sánh
AI đạo đức
AI tips
AI market
AI relations
AI quân sự
AI an toàn-an ninh-techwar
AI việc làm
AGI
AI doanh nghiệp
AI No-Low Code
OpenAI ChatGPT
AI môi trường-năng lượng
AI skill-talent
AI & công nghệ khác
AI nghiên cứu
AI chips-hardware-compute
AI vs con người
AI coding assistant
AI mở-nguồn mở
AI năng suất
AI startup-M&A
AI tương lai
AI báo chí
AI data
AI bản quyền
AI PC
AI riêng tư
AI deepfake-ảo giác-ANTT
AI ảnh-video-music-âm thanh
AI detect-test
AI minh bạch
AI nhỏ
AI hồi sinh
AI kiếm tiền
AI nông nghiệp-thực phẩm
AI ngân hàng-tài chính
AI giao thông
AI marketing-kinh doanh
AI smartphone
AI robotics-auto-agents
AI consumer devices
AI ecosystem
AI manufacturing
AI benchmark
Telecom
AI thành công-thất bại
National AI links
Digital
Semi-Cloud-DC-Green
Cybersecurity
Digital Legal
HTS