If there’s one thing we can be sure of, it’s this: Gov. Phil Murphy has the state of New Jersey all-in on AI.
From the AI conference (and hub) in Princeton, to the celebration of CoreWeave’s decision to stay in state (it’s moving to the NEST in Kenilworth), to EDA programs offering incentives to AI entrepreneurs (up to $400,000), the state appears to be doing all it can to fulfill Murphy’s vision of making New Jersey the capital of artificial intelligence on the East Coast.
Or is it?
An extraordinarily insightful and entertaining panel on the topic last week at the N.J. Utilities Association conference in Morristown detailed issues with the biggest challenge to creating AI centers: Energy.
Questions were everywhere:
- Can the state’s utilities handle the increased load needed for data centers?
- Are state policies that seemingly are going against great warehouses sending the wrong signal?
- If the state can solve its power generation problems — a big ‘if’ in the short run — should those solutions be used to build super data center as opposed to other needs?
- Will big data centers use up so much energy that other sectors are hurt?
Dan Kennedy, the CEO of the New Jersey chapter of NAIOP, said what many are thinking: There are mixed signals coming from state government.
“You’ve got an administration that has executive order saying, ‘We want AI in our state,’ and then you’ve got the Legislature that, if you look at the legislation being dropped, would basically render AI impossible to have in the state,” he said. “It’s a complete juxtaposition.”
Or is it just thoughtful examination?
Zenon Christodoulou, a commissioner on the N.J. Board of Public Utilities, offered that idea. He said understands the growing impact of AI in society – and economic desire to be first to market.
But he told an overflow crowd that each plus comes with minus. So, while he understands the business case for AI, he asks how increased use of AI will impact all businesses — and employees.
What does it mean for the businesses that can’t afford the best AI models, Christodoulou asked and then attempted to answer.
“Are they put out of business immediately by the more efficient ones that are able to incorporate it?” he said. “How does that transform the industry in the sense that now you’ll have more concentrated power in terms in fewer businesses and maybe limiting competition, which is really something that we should all be interested in.”
As far as its impact on the economy as a whole, Christodoulou noted reports that indicate increased AI could cut as much as 50% of white-collar jobs — and said this is something regulators need to take into consideration.
“How does (AI) make our lives better, or how does it burden us?” he said. “That’s something which, as a regulator, we have to be somewhat concerned with.”
Christodoulou, as is the role of BPU, questioned the impact on ratepayers, too.
Ed Gray, VP of asset management and pricing at PSE&G, said costs associated with potentially big projects should not necessarily be a factor in debating whether to accept them.
“I will offer that the data centers, like every other large industrial customer we’ve ever had, pays a lot more into rates than it typically takes to hook them up,” he said. “So, when you spread out the load over much larger loads, they typically will draw down distribution transmission rates.”
Where do we go from here?
Kennedy suggested the state become more realistic about its goals. Stop chasing unicorns, he said.
“One of the challenges is that some of the projections that are being reported out are based on these massive data centers running these massive models,” he said. “As much as I’m a homer for New Jersey, I don’t really know if that’s going to be New Jersey’s future.”
There is a future, Kennedy said.
“I think there’s a lot of opportunity for ancillary, lesser energy demand type of AI uses, data center tech, all that stuff that’s already here in the state to grow,” he said. “And if we’re in that territory, some of the things that commercial real estate does to partner with both utilities and an emerging renewable energy sector is going to be important.
“We can’t meet the load requirements for these big data processing centers, but with things like rooftop solar, community solar, we can do some of the balancing (that’s being discussed) and meet those demands.”
Of course, the slower you move, the more likely the emerging data center sector will go elsewhere — as it already is.
New Jersey’s hopes and dreams could be lost — and lost without escaping the cost, Gray said.
“The challenge is that if we don’t make it work — don’t build it in New Jersey — they (will) go across the Delaware River and build it in Pennsylvania, and we take all the concerns about power prices and get none of the associated benefits of having it in New Jersey,” he said.
