AI trades back in the green after that tough week last week comes on the back of the reported

massive $18 billion financing deal for an Oracle Tide data center in New Mexico.

Our dear Dubosa is looking at that and this new divide in AI for today's tech check.

Hey, Dee.

Hey, Carl.

So Oracle, that's just the latest AI mega debt deal.

And it really underscores a growing contrast in the AI race at large.

Now, in the U.S., the AI boom, it is being built on borrowed money. So massive data centers

financed through private credit and bond markets. In China, as a contrast, it is being built on

efficiency. That's cheaper chips and open source models, far leaner infrastructure requiring far less capital.

Now, this is the artificial analysis intelligence index cited by Goldman in a note over the weekend.

Chinese models like Kimmy and Alibaba's Quinn, they're ranking alongside top American models on benchmarks despite that massive difference in investment.

Now, the latest breakthrough, it came last week from a Chinese startup backed by Alibaba called Moonshot.

Its open source model, Kimi K2, outperforming on several standard benchmarks and reportedly costing less than $5 million to train.

Now, if that sounds familiar, it's because it is.

It fits into the story that's really

been emerging from China's AI scene since the beginning of this year. Moonshot, Mini Max, Quen,

and yes, Deep Seek, all continue to close the performance gap with far smaller budgets. Now,

take a look at this chart. U.S. Cloud Giants, they are projected to spend nearly $700 billion

on data centers by 2027.

China's biggest players, Alibaba, Tencent, Bight Dance, and Bidu, they are together expected to

spend about $35 billion. That is a 20 to one gap in capital spending for systems that are

...