Follow
Follow
Follow
Follow
News
Log In
Follow
Follow
Follow
Follow
News
By Rex M. Lee | Security Advisor | Tech Journalist | My Smart Privacy
After spending the past three-plus years working with Smart Africa on cybersecurity, telecom, internet, and AI infrastructure initiatives across Africa, I realized that Chinese telecom and technology companies were responsible for much of the infrastructure buildout throughout the continent through China’s Belt and Road Initiative.
I had the opportunity to attend and speak at cybersecurity conferences in Accra, Ghana; Dakar, Senegal; the Cape Verde Islands; and Morocco, including the GC3B Global Cybersecurity Conference and Regional Cybersecurity Week Conference.
Through collaboration with more than 30 member nations and EU partner organizations involved with Smart Africa, I came to understand that the global AI race is no longer just about chatbots, algorithms, or semiconductors.
It is now an infrastructure war centered on data centers, electrical grids, telecom networks, cloud ecosystems, and the industrial supply chains powering artificial intelligence.
As I investigated the telecom, internet, and AI infrastructure being built out across Africa, I realized that much of the Chinese infrastructure was also integrated with U.S.-manufactured technologies, including AI IP, operating systems (Android, iOS, and Windows), apps, platforms, AI chatbots, cloud ecosystems, and other technologies developed by NVIDIA, IBM, Google, Apple, Microsoft, Amazon Web Services, and other U.S. tech giants.
After all, there are effectively only three primary onramps to the global internet ecosystem: Android, iOS, and Windows.
This means Chinese infrastructure, app, AI, and platform developers remain dependent on Google, Apple, and Microsoft to participate in global internet trade and commerce.
Additionally, without AI GPU chips developed by NVIDIA and AI IP originating from companies such as IBM, many Chinese telecom, internet, AI, app, platform, and chatbot developers may never have entered the AI race to begin with.
Developers worldwide remain dependent on Google, Apple, and Microsoft ecosystems for the development, operation, and distribution of AI-infused apps, platforms, social media networks, and chatbots because Android, iOS, and Windows continue to dominate global internet access and computing environments.
Without Android, iOS, or Windows ecosystems, companies such as ByteDance (TikTok), Alibaba (Qwen AI), Baidu AI, DeepSeek, Tencent (WeChat), Temu, and many other Chinese AI app and platform developers would not have achieved global scale.
Due to the dominance of Google Android, Apple iOS, and Microsoft Windows ecosystems, AI data centers must be designed to support all three operating systems, along with the AI-infused apps, platforms, social media networks, chatbots, digital currency systems, and other evolving digital technologies that operate within those environments — including the next generation of quantum computing technologies.
Unless there is a major shift away from dependency on Android, iOS, and Windows toward alternative operating systems and ecosystems — such as Debian Linux-based platforms capable of supporting billions of devices globally at scale — Chinese AI and quantum developers, like developers worldwide, will likely remain dependent on Google, Apple, and Microsoft ecosystems. This means U.S. AI and quantum IP, software frameworks, and development environments will continue being shared globally, including with Chinese developers and manufacturers.
It also raises broader concerns regarding security, safety, and privacy because Android, iOS, and Windows ecosystems support extensive surveillance, data mining, targeted advertising, AI-infused apps, social media platforms, and chatbot technologies tied to business models rooted in Surveillance Capitalism.
After having hands on experience working with Smart Africa, it made me ponder the question: “Is the U.S. in an AI race with China, or are the U.S. and China trading partners?” — which became the title of a report I produced for NTD News in 2025.
Governments and media outlets continue framing AI as a strategic competition between the United States and China, often presenting it as a battle over compute power, national security, and technological dominance.
The U.S. AI/data-center buildout is being marketed as a national-security race against China, while parts of the buildout still depend on Chinese-linked supply chains—especially electrical infrastructure.
However, after researching the global AI data center buildout and drawing from firsthand experience working with Smart Africa and global telecom stakeholders, I believe the conversation is missing a critical reality:
China may already be winning the global infrastructure race underpinning artificial intelligence.
The Infrastructure Behind AI
Direct China exposure: roughly 11% of total U.S. data-center equipment imports by value in 2025, down from 41% in 2020. But that hides the bigger problem.
Power infrastructure exposure is much higher: U.S. data-center power-stack imports—transformers, switchgear, lithium-ion batteries—reached about $77.1B in 2025. China supplied 59% of U.S. lithium-ion battery imports, and direct Chinese transformer share reportedly fell to 10%, but with major concerns about rerouting through Vietnam and Thailand.
The most vulnerable area is not the servers alone; it is the power chain: transformers, switchgear, batteries, inverters, cabling, control electronics, and grid-interconnection equipment. DOE also confirms large power transformers are essential grid components with long replacement lead times and resilience concerns.
Artificial intelligence does not exist in isolation. AI requires an enormous physical infrastructure ecosystem consisting of:
While the United States leads in many areas of AI software, semiconductor design, and cloud services, China has spent years scaling something equally important: global infrastructure dominance.
The Global Huawei and ZTE Footprint
Over 170 countries built portions of their telecom, internet, and AI infrastructure using technologies from Huawei and ZTE. Much of this infrastructure integrates with U.S.-origin AI IP, software frameworks, and operating system ecosystems.
This means American-developed AI technologies and software ecosystems are indirectly helping Chinese infrastructure vendors compete against Western manufacturers such as Nokia and Ericsson on a global scale.
This is not theoretical. I observed these dynamics firsthand while working with Smart Africa, including collaboration involving Nokia and other telecom stakeholders across emerging markets.
In many developing regions, Chinese vendors succeeded because they offered:
The result is a global digital infrastructure footprint deeply influenced by Chinese hardware and supply chain ecosystems.
Below are the noted Chinese manufacturers:
Servers / AI Compute / Cloud Infrastructure
The Data Center Contradiction
The contradiction becomes even more apparent when examining the global AI data center expansion.
The United States continues warning about the strategic threat posed by China’s AI ambitions while portions of the infrastructure supporting the AI boom remain dependent on Chinese-linked manufacturing and supply chains.
Even as direct Chinese server infrastructure usage in the United States declines, AI data centers still rely heavily on Chinese-linked:
Globally, the dependency is even greater.
Below are the noted manufacturers:
Power / Battery / Grid Infrastructure
Semiconductor / AI Accelerator Ecosystem
ODM / Manufacturing / Supply Chain
The important distinction is:
This is why many analysts say the “decoupling” narrative is incomplete. Even when U.S. hyperscalers avoid direct Huawei servers, portions of the electrical, battery, manufacturing, or component stack may still originate from China-linked supply chains.
The global AI/data-center buildout is now one of the largest infrastructure expansions in modern history, projected to reach roughly $7 trillion by 2030.
What is happening globally is essentially a new “AI industrial map,” where countries are competing for:
Meanwhile, U.S. companies continue selling:
into Chinese AI and data center ecosystems.
In other words, both nations remain interconnected through the underlying AI infrastructure stack despite public narratives centered on decoupling and strategic competition.
Major Countries Building AI/Data Centers
United States
The largest AI infrastructure market globally, accounting for roughly 77% of AI infrastructure spending in Q4 2025.
Major builders:
Primary regions:
Texas alone is seeing enormous growth tied to AI power demand.
Middle East (Fastest Growth Region)
The Gulf states are aggressively building “sovereign AI” infrastructure. IDC reported Middle East & Africa AI infrastructure growth above 500% YoY in late 2025.
Saudi Arabia
Major builders:
Goal:
United Arab Emirates
Major builders:
The UAE is positioning itself as an AI hub between Europe, Asia, and Africa.
Europe
Europe is rapidly scaling due to AI sovereignty concerns.
Germany
Large expansions from:
Portugal
Microsoft reportedly committed around $10B toward AI infrastructure.
United Kingdom
Large hyperscaler and sovereign-cloud buildouts:
Italy
Growing as a Mediterranean AI hub.
Nordic Countries
These regions are favored because of:
Asia-Pacific
China
Still the world’s #2 AI infrastructure market despite export controls.
Major builders:
China is simultaneously:
India
Rapidly emerging AI/data-center market:
Builders:
India is becoming a major “AI sovereignty” nation.
Japan
Focused on:
Singapore
Critical AI and cloud hub for Southeast Asia despite land/power constraints.
Australia
Sydney and Melbourne are expanding heavily due to Asia-Pacific demand.
Latin America
Brazil
Emerging hyperscale hub:
Builders:
Brazil is attracting both U.S. and Chinese interest.
Who Actually Controls Many “Foreign” Data Centers?
One major overlooked issue:
A significant share of non-U.S. data centers are still operated by U.S. firms.
A 2025 study of 775 non-U.S. data-center projects estimated that U.S. companies operate about 48% of non-U.S. projects by investment value.
Meaning:
That is why many governments are now trying to build:
Bigger Strategic Trend
AI data centers are becoming:
The global competition is no longer just about software.
It is now about:
Globally, dependency on Chinese vendors for data-center and AI infrastructure is significantly larger than in the United States.
Outside the U.S., Chinese firms are deeply embedded across:
The reason is simple:
Chinese vendors often provide:
Areas Where Chinese Vendors Dominate Globally
Batteries / Energy Storage
China dominates global lithium-ion battery manufacturing.
Chinese companies such as:
control large portions of the global battery supply chain used in:
Globally, China controls roughly:
That means even Western-built data centers often rely on Chinese battery ecosystems.
Telecom & Fiber Infrastructure
Chinese firms have massive global penetration:
Particularly across:
Huawei alone helped build telecom infrastructure in over 170 countries before U.S. restrictions intensified.
Many AI/data centers rely on:
Electrical Infrastructure
China is one of the world’s largest exporters of:
Chinese companies like:
are heavily used globally in:
Cloud & AI Infrastructure Expansion
Chinese hyperscalers are expanding globally:
especially in:
These firms often bundle:
Africa: One of the Largest Chinese Infrastructure Footprints
Africa is one of the clearest examples.
Chinese companies have helped finance/build:
Huawei and ZTE became dominant because they provided:
This often created long-term infrastructure dependency.
Latin America
Chinese influence is also expanding rapidly through:
Countries including:
have substantial Chinese telecom and infrastructure exposure.
Europe
Europe has reduced direct Huawei telecom deployment in some countries, but dependency still exists indirectly through:
Germany in particular remains economically intertwined with Chinese industrial supply chains.
The Core Reality
Even when countries claim to be “de-risking” from China:
So the real issue is not simply:
“Is Huawei equipment installed?”
The larger issue is:
“How much of the underlying industrial ecosystem depends on Chinese manufacturing capacity?”
Globally, the answer is:
A very large percentage.
Especially for:
Despite U.S. export controls, a significant number of American companies still provide products, software, cloud architecture, tools, or indirect infrastructure support tied to Chinese data-center and AI ecosystems.
The relationship is complex because:
Major U.S. Companies Historically or Currently Involved
NVIDIA
NVIDIA has been one of the most important suppliers to China’s AI ecosystem.
Chinese firms historically purchased:
Customers reportedly included:
Even with restrictions, China remains a major market for NVIDIA because Chinese AI infrastructure is heavily dependent on CUDA and NVIDIA’s AI software ecosystem.
Oracle
Oracle has operated in China through partnerships and cloud arrangements.
Oracle products historically used in Chinese enterprise/data-center environments include:
Oracle Cloud infrastructure has also been tied to multinational operations inside China.
Microsoft
Microsoft has maintained one of the deepest enterprise presences in China among U.S. tech firms.
Key dependencies include:
Microsoft also partnered with local operators for Azure China cloud regions.
Many Chinese AI developers still build on:
Google services are restricted in mainland China, but Google technologies remain globally foundational.
Chinese developers still commonly rely on:
Android remains one of the largest indirect U.S. software/IP dependencies globally, including in China-derived app ecosystems distributed worldwide.
Amazon Web Services
AWS historically operated in China through local partnerships.
AWS-related infrastructure and developer ecosystems have influenced:
Intel
Intel CPUs still power enormous portions of China’s server infrastructure.
Despite China’s push toward domestic chips:
AMD
AMD server processors and accelerators are also used within portions of China’s AI/server market.
Broadcom
Broadcom networking chips and infrastructure technologies are embedded throughout global networking ecosystems, including Chinese enterprise environments.
VMware
VMware virtualization technologies historically played a major role in enterprise data-center architecture globally, including China.
The Bigger Reality: U.S. IP Dependency
Even when China builds domestic infrastructure, much of the global digital ecosystem still depends on U.S.-origin:
Examples:
This is why China is aggressively pursuing:
Strategic Tension
The contradiction many analysts point to is:
The U.S. frames AI competition with China as a national-security issue while:
At the same time:
AI Is Now an Energy and Infrastructure War
The next phase of AI dominance may not be determined solely by who develops the best chatbot or large language model.
It may instead be determined by:
China recognized this years ago and aggressively expanded its infrastructure influence across Africa, Asia, Latin America, and the Middle East through telecom, smart city, cloud, and industrial infrastructure initiatives.
The West is now attempting to catch up while simultaneously trying to reduce supply chain dependency.
National Security Questions Moving Forward
The AI race is increasingly exposing contradictions within the global technology ecosystem.
How can nations claim to be in a strategic AI competition while simultaneously relying on interconnected infrastructure, manufacturing, and software ecosystems?
How secure are AI systems when portions of the underlying infrastructure stack depend on geopolitical competitors?
And perhaps most importantly:
Can true AI sovereignty exist without infrastructure sovereignty?
As AI continues reshaping national security, defense, critical infrastructure, and the global economy, these are questions policymakers, technologists, and industry leaders can no longer afford to ignore.
Rex M. Lee is a security advisor, tech journalist, former OTA app and platform developer, and author of the Electronic Bill of Rights initiative.
His work focuses on surveillance capitalism, AI governance, digital infrastructure, and tech-based hybrid warfare.
See Rex’s Interviews, reports, and docuseries on TechStorm, PODTV, Rokus, and Amazon Fire TV at: CyberTalk TV with Rex M. Lee