Skip to main content

It's been a long year - and it's only January.

· 2 min read
Jenna Inouye
Jenna Inouye
Lead Editor

From Deepseek to Stargate, if your head isn't spinning you probably aren't paying attention.

Illustrative: A data center, spliced.

In January 2025, significant developments in artificial intelligence (AI) have emerged across various sectors:

Major AI infrastructure investment bums out sci-fi fans​

On January 21, President Trump announced the launch of "Stargate," a $500 billion AI infrastructure project developed in collaboration with OpenAI, SoftBank, and Oracle. This initiative aims to advance AI capabilities and is expected to create hundreds of thousands of jobs in the United States. The project will commence with data centers in Texas, with potential expansion to other regions. (New York Post)

The AI data center is expected to cost $292 million to build, codenamed "Project Ludicrous," and bring in... 57 jobs. (Business Insider)

Exactly what AI needs: even fewer regulations​

On January 23, President Trump signed an executive order titled "Removing Barriers to American Leadership in Artificial Intelligence." This order seeks to enhance U.S. leadership in AI by revoking certain policies and establishing a plan to promote AI development free from ideological bias. (The White House)

Meanwhile, the rest of the world strikes toward sanity​

At the World Economic Forum in Davos, AI leaders expressed differing views on the rapid advancements in AI technology and its potential dangers. Google DeepMind's Sir Demis Hassabis and others warned AI might threaten civilization if not properly controlled. Yann LeCun from Meta critiqued these views, calling them hypocritical given their competitive nature. (Financial Times)

And we don't talk about Deepseek​

DeepSeek, a Chinese AI startup founded in 2023, has rapidly gained attention for its advanced language models. Notably, DeepSeek-R1, released in January 2025, rivals leading models like OpenAI's o1 in performance. This model employs reinforcement learning techniques to enhance reasoning capabilities without supervised data, although it still struggles with a certain T-word. (WSJ)