Tesla to install $500M supercomputer in Buffalo

Gov. Kathy Hochul announced Friday that two giant computing projects totaling more than $900 million in state and private-sector investment are coming to the Buffalo area, as advanced technology and especially artificial intelligence take on an increasingly prominent role in business and life in general.

Tesla Inc., the electric vehicle maker, will spend more than $500 million over five years to install the Dojo supercomputer it is developing at its RiverBend gigafactory in South Buffalo, Hochul said.

And the governor confirmed that the University at Buffalo will be the home of the state’s $400 million Empire AI computing center and consortium that will focus on encouraging responsible research and development into the use, risks and opportunities for artificial intelligence.

“Whoever dominates the AI industry dominates the next era of human history,” Hochul told a packed audience at the UB Center for the Arts in Amherst on Friday morning. “It is going to be the single-most consequential technological and commercial advancement since the invention of the internet.”

Dojo is a supercomputer that Tesla is developing to bolster its work on developing its autonomous driving system, which has become a big source of work at the Buffalo factory.

Morgan Stanley analysts in September estimated that Dojo could add as much as $500 billion to Tesla’s market value, sparking a jump in the electric vehicle maker’s stock at the time.

“Investors have long debated whether Tesla is an auto company or a tech company. We believe it’s both, but see the biggest value driver from here being software and services revenue,” the report said.

Tesla has been working on Dojo, an in-house supercomputer project, for at least five years. The idea is to develop artificial intelligence systems that can handle the complex tasks that are required by Tesla’s Autopilot driver assistance system, as well as its efforts to develop “full self-driving” capabilities.

Tesla CEO Elon Musk, in comments to investors and analysts Wednesday evening, portrayed Dojo as a high-risk, high-reward project.

“I would think of Dojo as a long shot,” Musk said. “It’s a long shot worth taking because the payoff is potentially very high. But it’s not something that is a high probability. It’s not like a sure thing at all. It’s a high-risk, high-payoff program.”

He added that “a lot of our progress in self-driving is training limited.”

“If you remember when you first started to drive, how much of your mental capacity it took to drive. It was – you had to be focused completely on driving,” Musk said. “Then after you’ve been driving for many years, it only takes a little bit of your mind to drive and you can think about other things and still drive safely. So the more training you do, the more efficient it is at the inference level. So we do need a lot of training.”

“It’s a very interesting program. It has the potential for something special,” Musk said.

While other Tesla factories competed for the site, state officials said Tesla based its decision on the availability of reliable and low-cost power, the “strong talent pipeline” and the availability of usable space in Buffalo.

The Dojo project would give a boost to the $950 million factory on South Park Avenue, which has largely failed to become the transformative economic development project that was envisioned by the Cuomo administration when they made it the centerpiece of the Buffalo Billion initiative.

While initially designed to produce solar panels – and later Tesla’s sophisticated solar roof – the plant instead has become increasingly devoted to producing electric vehicle superchargers and doing data analytical work on Tesla’s self-driving software – creating hundreds of desk jobs that require a high school diploma, rather than the skilled advanced manufacturing positions state officials initially targeted.

The UB project was first unveiled in Hochul’s State of the State address but without specifying a location. The new initiative will be funded by $275 million in state money over 10 years – including $25 million from the State University of New York – and $125 million from private partners.

“This is the most transformational announcement our university has received in many many years, probably decades,” said Venu Govindaraju, UB’s vice president for research and economic development.

“Whoever takes the lead in AI will have the Silicon Valley effect,” he said. “Traditionally, academia is where the brilliant ideas came out. We are taking back that mantle from the profit-making companies. This is going to put New York state and the University at Buffalo at the center of this revolution.”

The consortium will include seven founding institutions – Columbia University, Cornell University, New York University, Rensselaer Polytechnic Institute, the State University of New York, the City University of New York and the Simons Foundation. Hochul said she already has the partners committed, but now has to get the state component through the Legislature.

“We will be the first in the nation to bring together top academic institutions, private sector and government, as well as philanthropy, and make New York State the home of the technology of the future,” Hochul said. “AI can help us diagnose and cure cancer, advance warnings of dangerous storms, and help us in our everyday lives.”

Details on exactly where the new computing center will be located are still being worked on, UB officials said, including which of the two campuses – North or South – as well as whether UB will have to construct a specialized new building or retrofit an existing one.

“This is a special building,” said UB President Satish Tripathi, a computer scientist by training. “All the AI applications require so much computing power, which mean they require a lot of cooling power and other things. It’s a big investment but the machines cost quite a bit.”

UB already has about 200 faculty members doing research or work on AI, Tripathi said. For example, a $20 million project, funded by the National Science Foundation, uses AI to identify speech and learning deficiencies in children.

“Those are the issues that we are looking at here,” he said. “We have experts in those areas.”

Hochul said the two investments will help to make New York into a hub for AI, which is already a $150 billion industry, and which is projected to hit $1.3 trillion in a few years.

“Other states want to be the home of the next generation of supercomputers needed to power AI’s progress. Everyone wants to be that,” she said. “We want to make sure that New York state is the capital of AI development.”

But she also acknowledged the concerns about both the failings of AI, including the potential of “hallucinations” or made-up information, as well as the risk of deliberate misuse of the technology to harm others.

“There’s going to be jobs created. We’re going to be using this in ways that solves problems,” Hochul said. “But that does not mean that I do not want research done and the federal government dealing with the ways people are doing deep fakes. We’re seeing more examples, whether Taylor Swift or whether it’s local elected officials. This is a deep concern of ours, that people’s images can be captured and misrepresented.”