Reindustrialization is alive in America.Yet the recent backlash on artificial intelligence (AI) data centers reveals we have not learned from our offshoring mistakes: creating dependencies that will not be easily reversed.Whether or not AI data centers are built in America, AI is here to stay.
Either we choose to shape the trajectory of AI technology by embracing the AI data center buildout or cede this advantage to other nations that are eager to lead the AI race.China, for example, announced nearly $300 billion in data center investment through 2030, not including private sector spending.
India is offering a 20-year tax holiday for foreign cloud operators and the Republic of Korea is offering regulatory relief.Meanwhile, many Americans are calling for moratoriums.The national security consequences of who hosts AI data centers go beyond competition with China.
If we do not build AI data centers in the United States, we risk everything that is inside the data centers: the data, compute, and intellectual property (IP) that drive innovation, protect privacy, and enable sovereignty.We will also forfeit our ability to set the security standards for this critical infrastructure.LIZ PEEK: CHINA IS CHEERING AS THE LEFT BLOCKS AMERICA’S AI INFRASTRUCTUREIf AI data centers are the engines of AI innovation, data is the fuel.
A significant amount of data is required to make AI useful, all of which must pass through AI data centers.Data, therefore, must be handled carefully — especially for privacy reasons — and viewed as a national asset.Adversaries will view data routed through AI data centers as an opportunity to advance their own AI ambitions.
Data may be collected for espionage purposes or to accelerate adversaries’ own AI models, as we have seen China do time and again in cyberspace and through adversarial distillation.Adversaries may also conduct adversarial attacks during model training, which could, without our knowledge, degrade and manipulate data that ...