IThe “quantum chandelier” housed in a glass case in the chapel on the Rensselaer Polytechnic Institute campus in Troy, New York, is the symbolic centerpiece of an ambitious effort to turn upstate New York into a high-tech hub — what Silicon Valley is to social media or Cambridge, Massachusetts, is to biotechnology.
The silvery sci-fi object — named for the gold grids inside it that suspend, cool and insulate its processor — is the heart of a “quantum computing system” that could herald a new era of computing. It’s the centerpiece of the dream of Curtis Priem, co-founder of Nvidia, the $2.8 trillion artificial intelligence hardware and software company, to turn Rensselaer, or RPI, into a hub for advanced computing and transform this upstate New York area into a new Silicon Valley.
Priem has invested a considerable portion of his fortune in building the Curtis Priem Quantum Constellation, a workshop for RPI students who want to showcase their vision of a quantum computing future. Just as his partners at Nvidia, where he was the company’s first chief technology officer, gave him the freedom to imagine the graphics chip architecture that powers the AI revolution, he hoped his investment would spark a new era of computer-driven innovation in the region.
Priem is betting that the area along the Hudson Valley — from Yorktown Heights, home of IBM’s quantum lab, to Troy, home of RPI and Suny’s NanoTech complex, and west to Syracuse, where Micron is building a $100 billion mega-factory complex — is the future home of American computing technology.
To do this, it is thinking beyond concerns about artificial intelligence and the success of Nvidia’s H100 graphics processing units (GPUs), which underpin up to 90% of generative AI systems.
Wall Street has become more skeptical of the technology. AI has lost billions of dollars as Wall Street became disillusioned with the idea that new technology is about to change the world. But the same thing happened with the overexploitation of the Internet in the 1990s, which led to a boom and a bust, before finally paying off.
In theory, quantum technology will be able to solve problems in seconds that would take today’s supercomputers decades to solve, if they ever do so, unlocking secrets about the behavior of molecules, genetic codes, weather prediction and – the latest concern – breaking the encryption systems that underpin the Internet.
The technical and financial challenges are enormous. The quantum computer built by IBM and Rensselaer is so advanced that it must be cooled to near absolute zero (−273.15 °C, −459.67 °F). To make it work, programming a quantum computer is difficult, to say the least. Traditional computers operate with a binary code of ones and zeros. Quantum mechanics is considered to be one of the most difficult areas of physics because values exist in multiple states at once. In a computer, quantum bits, or qubits, are “entangled,” meaning that the properties of one depend on the properties of the things around it.
For Priem, 66, that means giving the feverish minds of Rensselaer’s science student body the freedom to imagine a new computing world. “It’s exciting to see these kids with no blinders and no restrictions and this technology that no one else has,” he says. “Suddenly, they say, ‘Wow! I can invent the future.’”
The development comes at a time when the United States is racing to advance computer science and repatriate chip manufacturing from Taiwan and China as a matter of national security. The pandemic taught American manufacturers and consumers that supply chains are fragile. In 2022, Congress authorized the $280 billion Chip Act to bring chip innovation back home and protect the United States against future disruption.
“We have to analyze everything,” Priem says. “Everything can fall apart. We thought we were friends with the Chinese, and then it turned out that it all depends on who our leaders are.”
Priem, Martin Schmidt, the former MIT president who is now president of Rensselaer, and Kathy Hochul, the governor of New York, believe that a good portion of that money should go to upstate New York, an area that has four essential qualities: land, water, energy and an intellectual “brain belt.”
AI is notoriously demanding on water and energy consumption: Goldman Sachs predicts that demand on US data centers will grow from 3% of current energy consumption to 8% within six years. And then there’s the demand for water needed to cool the chips.
Since Priem is known for the architecture behind AI, people inevitably ask him what he thinks of it and whether they should be worried. He tells them that whatever comes next with AI, the experiment has already begun with social media.
“Only 1% of our existence is based on reality and 99% on beliefs,” he reasons. “So you have people talking to each other on social media and that is the chaos that AI will represent. Big companies try to detect what is true and what is false, but when you look at social media you can’t tell.”
Artificial reality is fine for a Marvel movie where we accept the fake as better than reality, but outside the cinema, he says, “everyone hates each other.”
Priem sold most of his Nvidia shares when the company was valued at $2 billion. It is now worth $2.5 trillion. Had he not sold, he would now be one of the richest people in the world. His former partner Jensen Huang’s shares were recently valued at $103 billion, but even that is below Nvidia’s June peak.
Priem doesn’t seem too disappointed. The shares he sold, he reasons, may have in turn helped a family buy a house or pay for their children’s college tuition. He remains, he says, “Nvidia’s number one fan. The other founders played the game, and I’m cheering them on from the stands.”
The divorce forced him to sell his Nvidia shares, but he remains in touch with his former Nvidia partners, Huang and Chris Malachowsky. Priem says it has been awkward at times, but he sends them text messages, most recently when the Huang family dogs had diarrhea and about his children.
“The only thing I can do is help from the outside. The last shares I owned were in 2006. It’s incredible what they would be worth, but I couldn’t monetize them without driving down Nvidia’s price,” he says.
Shares could get a further boost this week when Nvidia announces its second-quarter earnings. It is expected to reveal another jump in profits thanks to a doubling of revenue as it benefits from the rapid adoption of generative AI.
If Nvidia hits $210 a share, the three co-founders would be worth $1 trillion. “Everyone thinks that’s how we value success because they have no other way to achieve it,” he says, “but my $200 billion is what I’ve contributed to the American economy.”
“I have to be one of the luckiest people on the planet. What those guys did for me was trust me to design this architecture (now called Cuda) for the GPU. I’ve never seen that level of trust in any other engineer in our industry,” Priem says. “I gave them the framework, the rules, and I’m so happy.” When he met Huang recently at a fundraiser at the Monterey Bay Aquarium, Huang turned to him, shaking his head and saying, “I can’t believe we’re still using the same architecture.”
His time at Nvidia also provided him with unrivaled contacts. As in most businesses, technology is about relationships. Priem and Schmidt persuaded IBM’s head of research, Dario Gil, to send them the quantum computer that might otherwise have been destroyed. Gil and John Kelly, a former IBM director, sit on Rensselaer’s board of directors. Among the big tech bosses, Priem says, “there’s a certain level of trust about how to do these things.”
Schmidt says it takes 40 years for a region to become a full-fledged technology hub. This was the case at MIT after James Watson and Francis Crick discovered DNA in the 1950s, opening up the field of molecular biology around Kendall Square. “The creation of ecosystems creates a virtuous cycle,” Schmidt says.
“The assets in this region are focused on prototyping, packaging and manufacturing chips using extreme ultraviolet (EUV) light. What makes this technology different is that everyone has to share a factory, a manufacturing plant, because no one can afford to have their own facility.”
The goal of turning upstate New York into a tech hub, he says, requires looking at what other areas are doing wrong. Silicon Valley, which began around Stanford University, has run out of space and is now largely focused on social media in any case. San Francisco has quality-of-life issues. Taiwan’s TSMC is considering Arizona, but the state lacks water. Another contender is central Ohio, but it lacks workforce training.
“My goal is to show New York politicians what they have here,” he says. “They don’t even know they have a nanotechnology center. I don’t think there’s another state that can do this. But time is of the essence.” Priem says he was a little disappointed when Hochul recently said she was interested in AI. “We thought, be careful because that’s almost old news.”