When the CEOs of OpenAI, SoftBank, and Oracle stood beside Trump and announced the $500 billion, four-year "Stargate" project, "AI Infrastructure" finally arrived.
This is certainly not the first project regarding "AI infrastructure," but it is the first entity composed of giants playing their respective roles:
SoftBank is responsible for funding, and OpenAI is responsible for operations;
Four equal-equity initiators: SoftBank, OpenAI, Oracle, and MGX (the UAE sovereign fund, chaired by His Highness Sheikh Tahnoon Bin Zayed Al Nahyan—of course, a member of the royal family);
Oracle, Nvidia, and OpenAI are closely cooperating as components (it is worth noting that Jensen Huang is currently in China);
Arm, Microsoft, OpenAI, Oracle, and Nvidia are the core technical partners, and the first data center is currently under construction in Texas.
In a series of intensive one-on-one exchanges last week, we discussed HyperScale Data Centers extensively, even touching on Nvidia and Oracle.
Actually, it is quite interesting to put oneself in the shoes of every player at the table and consider their potential choices.
For instance, the previously mentioned issues regarding "Conflict of Interest," or the "independence" of Google, Amazon, Meta, and xAI.
To be honest, whether Stargate actually has that much capital—especially the $100 billion set for immediate deployment—remains a big question mark. However, as long as the project progresses, Nvidia doesn't have to worry about potential "order cancellations," and OpenAI gains more leverage in negotiations with Microsoft.
It’s all theater. In that case, we might as well play along with AI infrastructure.
In fact, these were the focus of my outlook for this year at the end of last year.
Conflicts of Interest; Energy and Data
Dao Ming, WeChat Official Account: Dao Ming Digital Nomad Life
It’s time again for the 2025 AI Outlook: two keywords, several predictions, and one potential risk.
Data and Energy
Dao Ming, WeChat Official Account: Dao Ming Digital Nomad Life
"2025 AI Outlook" Refinement and Additions: Data and Energy; Efficiency Gains and Organizational Change; Products and "User Experience."
It seems Google can continue its self-developed efforts, while the success of Amazon's self-developed chips will depend on Anthropic.