Latest / International Business with Fexingo: Global Trade, Cross-Border Deals, and Multinational Operations / How New Yorks AI Data Center Ban Affects Global Tech Trade
Transcript
- Lucas: So New York just became the first U.S. state to ban new AI data centers. Signed into law this morning by Governor Hochul — a moratorium on any new facility larger than fifty megawatts, citing energy grid strain and water usage. The bill's language is pretty sweeping: it applies to both new builds and major expansions of existing sites. Luna: That's a big deal. I mean, New York has been aggressively courting tech investment — remember the Micron semiconductor plant in Syracuse? This feels like a reversal. Lucas: It does, but it's also a sign of how quickly the conversation around AI infrastructure is shifting. Data centers are power-hungry — a single large facility can use as much electricity as a small city. And in places like upstate New York, where the grid is already constrained, you're seeing local pushback. The bill says the moratorium lasts two years, during which the state will study the environmental impact. Luna: Let me ask the obvious question: does this actually matter for global trade? New York is one state out of fifty. Chips get made in Taiwan and shipped everywhere. Lucas: It matters because of the signal. New York is the financial capital of the world, and it's effectively saying: we don't want your AI compute here. That sends a message to hyperscalers — Amazon, Microsoft, Google — who have been racing to build data centers globally. If a state like New York slams the brakes, other states might follow. And that accelerates a trend we're already seeing: AI infrastructure getting built in places with cheap energy and loose regulation. Luna: Right, because the real bottleneck for AI isn't just chips anymore — it's power. So where does that infrastructure go instead? Lucas: Exactly. Look at Saudi Arabia — they're spending billions to become an AI data center hub, using their natural gas and solar potential. Malaysia is another one. The sovereign wealth funds are pouring capital into those regions. And if the U.S. starts restricting data center construction, more of that capital flows overseas. That's a trade story — it shifts the geography of compute, and with it, the flow of technology and services. Luna: And it's not just New York. There's also the federal angle — the Biden and now Warsh administrations have been tightening chip exports to China. This week a trade official said 'very few' NVIDIA H200 chips have actually shipped there. So you've got export controls on the hardware side, and now a state-level moratorium on the infrastructure side. The whole AI supply chain is getting squeezed. Lucas: Yeah, and that's the through-line. The H200 is NVIDIA's most advanced chip that's still technically allowed for export to China under current rules — but in practice, licenses are being denied. So Chinese firms can't buy the chips, and now they can't build data centers in New York either. The net effect is that AI development is being geographically reorganized. And trade data is starting to reflect it. Luna: Let's look at the numbers we have. The U.S. trade deficit in goods and services widened to $77.6 billion in May — that's up from $54.6 billion in April. A pretty big jump. Imports are up, exports are up too, but imports grew faster. Lucas: Right, and the dollar has been weakening — the trade-weighted index is at 120.5, down from 120.75 last week. That makes U.S. assets cheaper for foreign buyers, which could actually attract more sovereign wealth money into U.S. infrastructure. But if states start banning data centers, that capital goes elsewhere. So there's this tension: the U.S. wants to lead in AI, but local politics are pushing back against the physical footprint. Luna: And the dollar weakness also makes U.S. exports more competitive. So maybe the widening trade deficit isn't all bad — but it does signal that demand for foreign goods is strong. Lucas: If these conversations have moved your work forward in some small way, listener support is what keeps this show ad-free and independent. We put a lot of research into each episode, and it's listener-funded. If that's valuable to you, you can support at buy me a coffee dot com slash fexingo. Luna: Yeah, it really makes a difference. A lot of episodes start from questions you all send in, too. Lucas: Okay, back to New York. The ban also has implications for the job market. JPMorgan just announced ai related job cuts this week — so there's this double effect: AI is displacing some jobs, and now the infrastructure to train those models is being restricted. It's a complex picture. Luna: And it's not just JPMorgan. Goldman Sachs made headlines for the SpaceX IPO fees — that's a different sector, but it shows how finance is adapting. Banks are looking for new revenue streams as traditional M&A slows. Luna: And on the trade finance side, the yuan's rise has been a big story — but today I want to focus on how U.S. policy is affecting the flow of AI hardware and capital. Because that's where the action is. Lucas: So let's zoom in on one specific trade flow: semiconductors. The H200 is a good example. NVIDIA designs it in the U.S., it's manufactured in Taiwan by TSMC, then shipped to customers worldwide. But with export controls, a lot of those chips are being diverted to other markets — Southeast Asia, the Middle East. And now, those markets are building data centers to host them. Luna: So the chip export controls are actually creating demand for data center construction in places like Malaysia and Indonesia. That's a supply chain shift with real money behind it. Lucas: Exactly. And sovereign wealth funds are stepping in. Saudi Arabia's Public Investment Fund, for example, has committed $40 billion to AI infrastructure. They're building data centers powered by their own natural gas. That's a direct competitor to U.S.-based facilities. If New York's ban spreads, you'll see more of that. Luna: What about the water angle? The New York law specifically mentioned water usage. Data centers need massive amounts of water for cooling. In upstate New York, that's a real concern for communities near the Great Lakes. Lucas: Yeah, a single large data center can use millions of gallons of water per day. So environmental groups are pushing back. And that's another reason why arid regions like Saudi Arabia or the U.S. Southwest are less ideal — they have water scarcity. So you see a push toward liquid cooling and other technologies, but that's expensive. Luna: So the trade story here is multifaceted: chips, capital, cooling technology, and even water rights are all part of the cross-border flow. Lucas: Absolutely. And let's not forget the financial flows. The dollar weakening means foreign investors get more for their money when buying U.S. assets. But if U.S. states ban data centers, the assets they want to buy aren't being built. So they invest elsewhere. That's a capital account story as much as a trade story. Luna: What about the response from the tech companies? Has anyone announced they're pulling out of New York? Lucas: Not yet — it just happened today. But you can bet that Microsoft and Amazon are re-evaluating their pipeline. They have billions committed to data center construction in the state. The moratorium doesn't affect existing facilities, but any expansion plans are on hold. That's a lot of uncertainty. Luna: And that uncertainty ripples through the supply chain. If you're a manufacturer of cooling equipment or server racks, you need to know where demand will be. This could shift orders to other states or countries. Lucas: Right. And it could also accelerate the trend of onshoring — but not to New York. To states like Texas, Ohio, or Virginia, which have more permissive policies. But those states have their own grid constraints. So the bottleneck just moves. Luna: So ultimately, the AI data center ban is a trade policy tool — even though it's a state-level environmental law. It reshapes where compute happens, and therefore where jobs and capital flow. Lucas: Exactly. And it's a reminder that trade isn't just about tariffs and shipping containers. It's about energy policy, environmental regulation, and state-level decisions that have global ripple effects. I think we'll see more of these local restrictions as AI infrastructure grows. Luna: One last thought: the Fed. Warsh today talked about a 'regime change' to fight inflation. If rates stay higher for longer, that could strengthen the dollar again, making U.S. assets more expensive for foreign buyers. That might slow the capital flow into AI infrastructure overseas. Lucas: Good point. So you've got monetary policy, trade policy, and state-level environmental policy all intersecting. That's the reality of international business in 2026. Luna: All right, let's leave it there. Next episode, we'll look at how the Strait of Hormuz toll reversal is affecting oil trade — that's a big one. Lucas: Looking forward to it. Thanks for listening.