Gotrade News - China is preparing a $295 billion AI data center plan. The blueprint aims to lock out Nvidia and favor domestic Huawei chips.
This surge in AI infrastructure spending is intensifying a global buildout race. US investors are now focusing on networking suppliers and data center developers.
Key Takeaways
- China will direct roughly 2 trillion yuan toward a nationwide AI data center network.
- Meta launched a $115 million program to train data center construction workers.
- CoreWeave is exploring a euro bond sale to fund its infrastructure expansion.
According to Yahoo Finance, China will channel about 2 trillion yuan over five years. That figure equals roughly $295 billion for a national computing network.
The plan reportedly requires at least 80% of technology from domestic suppliers like Huawei. That mandate effectively squeezes Nvidia and AMD out of a major market.
As reported by Capacity, the NDRC planning body is drafting the project blueprint. State firms such as China Mobile would operate most of the computing hubs.
In the United States, Meta launched its $115 million America's Workforce Academy. The no-cost program trains skilled tradespeople to build AI data centers.
Per CBS News, the program guarantees job placement for graduates. It will open in Louisiana, Ohio, Indiana, and Texas.
Meta is also reportedly in talks to lease a CleanSpark data center facility. According to Insider Monkey, the Georgia site holds about 250 megawatts of capacity.
CoreWeave (CRWV) is exploring its first euro-denominated bond sale. JPMorgan is organizing investor calls in Europe to gauge demand for the deal.
The company has raised more than $20 billion in funding during 2026. That capital supports nearly 50 data centers across North America and Europe.
CoreWeave supplies AI compute capacity to large customers including OpenAI and Meta. The euro deal would give European investors rare direct exposure to AI debt.
Rivals like Nebius Group (NBIS) are racing to add similar GPU capacity. The competition reflects soaring demand for rentable AI compute.
This buildout is also lifting demand for networking and optical gear. Suppliers such as Arista Networks (ANET) stand to benefit directly.
McKinsey estimates the global data center buildout could reach $7 trillion by 2030. That projection underscores the scale of the current infrastructure race.
Meta's academy targets welders, fiber technicians, and electricians in trades. Graduates earn an industry credential alongside a guaranteed job offer.
China's plan forms part of a broader "Six Networks" infrastructure program. Beijing wants to anchor long-term domestic AI development with state capital.
For retail investors, the AI theme is shifting from chips toward physical buildout. Power, networking, and labor are emerging as the new focal points.
Sources
