Ad
Short

Google's new TPU v7 Ironwood chips are now general available for training and running large AI models. According to Google, Ironwood delivers ten times the peak performance of the previous TPU v5p and is four times more efficient than the TPU v6e. Early adopters like Anthropic, Lightricks, and Essential AI are already using Ironwood for demanding AI workloads. Amin Vahdat, VP and GM for AI Infrastructure at Google, recently said every TPU Google has built so far is currently in use.

Alongside Ironwood, Google is introducing Axion, a new line of Arm-based virtual machines aimed at everyday computing tasks. Vimeo and ZoomInfo report that Axion N4A instances deliver up to 60 percent better price-performance than comparable x86 systems. With both Ironwood and Axion, Google is challenging competitors like Nvidia and working to offer more flexible AI infrastructure with better cost efficiency, especially for serving its own AI products like Gemini.

Ad
Ad
Short

Snap has signed a $400 million partnership with Perplexity AI to make its AI search engine the default option inside Snapchat starting in 2026. In addition to the existing "My AI" chatbot, Perplexity will be available to users worldwide. Snap CEO Evan Spiegel says the deal could open the door to future partnerships with other AI providers.

The agreement includes a mix of cash and equity and will be recognized as revenue beginning in 2026. Perplexity, which now serves more than 20 million users, is currently facing lawsuits over alleged terms-of-service violations from both Amazon and Reddit.

Ad
Ad
Ad
Ad
Short

Barclays estimates that data centers for AI giants like OpenAI, Meta, and Amazon could need as much as 46 gigawatts of electricity. That's about the same amount of power used by 44 million US households, or roughly a third of all homes in the country. Building out these projects could cost $2.5 trillion.

This massive expansion is already putting pressure on the power grid. Nvidia, Microsoft, and OpenAI are warning about possible grid instability caused by rapid swings in electricity demand. Some plans call for energy sources like solar plants and gas storage to be integrated directly into the data centers. OpenAI has even asked the US government to add 100 gigawatts of new power generation each year. It's still unclear how many of these projects will actually get built. According to Barclays, it's hard to draw the line between projects that are real and those that are still just speculation.

Google News