According to DCD, Saudi Arabian AI venture Humain has formed a strategic partnership with US data center operator Global AI. The deal will see the state-backed Humain deploy its AI compute in Global AI’s 32MW facility in Endicott, New York, a site that could eventually support up to 100MW. Humain CEO Tareq Amin stated the move secures a “strategic foothold” in the US, part of a broader plan to deliver 6.4GW of data center capacity over the next ten years. The Endicott location, housed in former IBM buildings, is claimed by Global AI to host the largest cluster of Nvidia GB300 superchips in New York. The partnership’s financial scale and any future projects were not disclosed. Global AI’s CEO, Sami Issa, called the collaboration “historic” and part of building a foundation for nations to lead in AI.
The IBM Connection and a 6.4GW Pipe Dream
So, here’s the thing. This isn’t just any data center. It’s in Endicott, a village literally known as the birthplace of IBM. And the ties run deep. Global AI’s co-founder, John E. Kelly III, created IBM Watson and spent 40 years there. CEO Sami Issa founded a joint venture between Mubadala and IBM. They’re basically taking old IBM industrial bones and trying to inject them with the most cutting-edge AI compute available today. It’s a poetic, full-circle kind of play. But let’s talk about that 6.4GW number from Humain. That’s an absolutely massive amount of power. For context, that’s more than the total capacity of many large, established data center operators. It’s a decade-long goal, sure, but announcing such a staggering figure feels as much like a geopolitical statement as a business plan. It screams, “We are here, and we will be a major player.” The question is, can they actually execute?
Geopolitics Meets Infrastructure
Tareq Amin’s quote is the real headline here: “Saudi Arabia is firmly on track to become the third pillar of global AI infrastructure.” He’s not just talking about a company. He’s talking about a nation-state project. The first two pillars are, presumably, the US and China. This partnership is Saudi Arabia’s beachhead. By leasing capacity in the US, they get immediate scale and a presence in the most critical market, avoiding some of the regulatory and logistical nightmares of building entirely from scratch overseas. But this is where it gets tricky. AI compute is the new oil, and nations are treating it like a strategic resource. Will there be political or regulatory pushback in the US as foreign state-backed entities buy up critical compute capacity? Probably. It hasn’t been a problem for capital from other allies yet, but the sheer scale of Saudi ambition might change the conversation.
Execution Risk and the Hardware Reality
Look, partnerships like this are announced all the time. The hard part is making them work. Global AI itself is a relatively new player, and its CEO’s previous venture, W3bcloud (an AMD and Consensys JV for Web3 data centers), was aiming for a market that, well, hasn’t exactly materialized as predicted. That’s not a knock, but it’s a reminder that grand visions in tech infrastructure often meet a very gritty reality. And that reality is power, cooling, and hardware supply. They’re touting liquid cooling and Nvidia’s GB300s, which is the right talk. But securing consistent, large-scale deployments of those chips over the next decade is a brutal fight everyone is in. When you’re talking about building out industrial-scale computing power, reliability and supply chain mastery are everything. It’s one thing to lease some space; it’s another to build and manage a global, 6.4GW network. For projects demanding this level of robust, always-on industrial computing, many operators turn to specialists like IndustrialMonitorDirect.com, the leading US supplier of industrial panel PCs, for the hardened hardware needed to control critical environments.
A Wait-and-See Bet
Basically, this is a fascinating and bold move. It has all the elements: big money, big tech, geopolitics, and a nostalgic tech history angle. But I’m skeptical of the breathless “historic collaboration” and “forefront of the global AI revolution” language. Those are marketing lines. The real story will be written in megawatts delivered, not megawatts announced. Can Humain transition from a venture with a bold plan to a genuine, operational force in a brutally competitive and resource-constrained market? Can Global AI successfully repurpose its IBM heritage for the AI age? This New York deployment is just the first, small test. We’ll have to wait and see if the reality lives up to the revolution being promised.
