SOFTBANK Group Corp founder Masayoshi Son is nothing if not willing to make the kind of bold bets that other investors might balk at. But his latest pitch – to create a US$1 trillion artificial intelligence hub that manufactures robots in the Arizona desert – may be his most audacious yet.
It comes after an impassioned speech last year when Son said that he was mourning the death of his father and grappling with what his own legacy would be. Then, he had an epiphany: He was “born to” dedicate himself to developing what he calls “artificial super intelligence”, or AI that is thousands of times smarter than humans.
The newfound purpose might be a motivating factor in bringing this new venture, dubbed “Project Crystal Land”, into fruition. But it’s hard to know how much of it is just magical thinking on the part of the Japanese billionaire better known as Masa.
At this stage, there are more questions than answers. SoftBank, which has yet to publicly unveil the proposal, is reportedly seeking to partner with Taiwan Semiconductor Manufacturing Co (TSMC). But an unnamed source familiar with the matter told the Financial Times that TSMC has not yet been formally approached about the plans. SoftBank is also vying for tax breaks with US federal and state officials, and is trying to lure interest from multinational tech firms including Samsung Electronics. But turning a swath of the American Southwest into a successful high-tech manufacturing base that can compete with China will hinge on also attracting an ecosystem of suppliers and secondary-materials and components makers to join in the efforts.
The inspiration for the new Arizona industrial complex is Shenzhen, the tech hub across the river from Hong Kong and home base of giants such as Huawei Technologies, Tencent Holdings and electric-vehicle maker BYD. Swarms of startups, suppliers, toolmakers and other stakeholders work in close proximity, meaning firms can quickly mobilise to solve manufacturing problems as they arise.
There’s the over-arching question of how Son can possibly raise the money to bring this vision to life. His earlier Stargate venture faces a funding reality check. The billionaire announced the massive data centre buildout plans at a White House press conference in January alongside President Donald Trump, OpenAI’s Sam Altman and Oracle’s Larry Ellison. Even without the obstacles from a trade war, it’s hard to envision how the full US$500 billion Stargate investment will ever be financed. And now he hopes to double that for a fresh project.
BT in your inbox
Start and end each day with the latest news stories and analyses delivered straight to your inbox.
Onshoring tech manufacturing, meanwhile, has been a long-stated policy goal of Trump’s. A spate of tech companies from Apple to TSMC have announced multi-billion-dollar US investment plans as the president has threatened to use tariffs to force them to move production to America. But tariffs alone don’t solve the economic math that uprooting long-established supply chains in Asia requires, or address the infeasibility issues of trying to recreate them in the US from scratch. It risks creating an environment where bold investment announcements are celebrated as political wins, while turning into manufacturing failures.
The success of this project would also rely on upskilling and training thousands of robotics technicians and artificial intelligence (AI) engineers in Arizona – something that can’t happen overnight. Apple, for example, relies on legions of skilled workers in China, a key reason why it still produces the majority of its devices there even as labour costs have ticked up and geopolitical tensions mount.
And when it comes to robotics, China has the upper hand in many areas. In addition to an extensive engineering talent pool, its advanced tech supply chains already give it easier access to components required for bot production. The industry has also recently been flooded with government support and birthed a rash of startups creating humanoids that can do gongfu or run marathons.
The race to lead the world in AI-enabled robots is still in the early stages, but many signs point to China coming out on top. Perhaps that is where Son’s unbridled optimism and willingness to risk it all comes in. His ambition for this project seems to rely on the thinking that if you build it, they will come.
It would be easy to dismiss as impossible, but Son has built his career on taking long-shot gambles. It hasn’t always worked out. But such a mentality has propelled his rise from Japanese slum to the single-largest foreign investor in the US. Some magical thinking along the way must have helped. BLOOMBERG
