Over the past decade, the United States has become the overwhelming global leader in the creation of and private investment in AI startups, according to S&P Global.
In a report released July 9, S&P said a total of 5,509 AI companies will be founded in the U.S. between 2013 and 2023. That’s more than all other countries combined and nearly four times as many as second-place China, which founded 1,446 companies in the same period. Rounding out the top five are the U.K. with 727 companies, Israel with 442 and Canada with 397.
AI startup activity is also reflected in private investment, with a cumulative $335.2 billion invested in the U.S. over the decade, according to S&P Global.
This is more than three times China’s private sector investment of $103.7 billion and far more than the UK’s $22.3 billion, Israel’s $12.8 billion and Canada’s $10.6 billion.
But S&P noted that much of China’s AI spending comes from the public sector and is less transparent, warning that the government plays a large role in AI investment.
Indeed, China has a three-year AI development action plan that includes programs to develop a domestic AI skilled workforce and attract foreign workers to AI projects, the report added.
Globally, S&P projects that private investment in AI startups will reach $800 billion to $900 billion by 2027, representing a compound annual growth rate of at least 70% to 74%.
The data on the booming US private AI industry comes at a time when the Chinese government is censoring AI developers.
Sources told the Financial Times last week that China’s top internet regulator, the Cyberspace Administration of China (CAC), was testing the company’s large-scale language model to ensure it “embodied core socialist values.”
CAC “has a special team that comes to our office, sits in a conference room and does the audit,” an employee at a Hangzhou-based AI company told FT. “We didn’t pass the first time, and we didn’t really know why, so we had to consult with our colleagues. It takes a bit of guesswork and tweaking. We passed the second time, but the whole process took several months.”
Ensuring “core socialist values” also means companies must monitor sensitive questions or keywords that could “incite the subversion of state power” or “undermine national unity”, the FT added.
As a result, queries about the Tiananmen Square incident and most questions about President Xi Jinping were rejected by China’s chatbot.
The Financial Times previously reported that a research centre under the CAC was developing a large-scale language model trained in “Xi Jinping Thought on Socialism with Chinese Characteristics for a New Era”.