• by Laurie Sullivan , 6 hours ago

On Thursday night, Sundar Pichai, Sam Altman, Tim Cook, Mark Zuckerberg, and other technology leaders attended a White House dinner with President Donald Trump, but Elon Musk was notably absent.

Reports suggest that Musk received an invitation and that a representative had intended to take one of the more than 30 chairs around the table, but was not identified.

Trump has secured billions in corporate commitments to drive development of an AI infrastructure, Bloomberg reports.

Meta CEO Mark Zuckerberg pledged to spend about $600 billion on data centers and infrastructures in the United States through 2028, all earmarked for innovation. Much of that will go toward supporting advertising.

Microsoft co-founder Bill Gates at the event focused on gaining Trump's support for global health programs and reconsidering cuts to USAID. 

Earlier this year, Apple CEO Tim Cook pledged to invest $600 billion in U.S. manufacturing.

Google committed $150 million to AI education at the AI Education task force meeting held earlier in the day. It will be chaired by First Lady Melania Trump. OpenAI CEO Sam Altman also attended.

During the AI education meeting earlier in the day, Melania Trump said: "The robots are here, and our future is no longer science fiction," adding: "I won't be surprised if AI becomes known as the greatest engine of progress in the history of the United States of America."

On Thursday, Hitachi Energy in the U.S. announced plans to invest more than $1 billion in an electric grid infrastructure that could support AI's growing power demands -- all needed for media buying and ad serving.

The performance of GPUs, as well as CPUs, is critical to the digital advertising ecosystem -- influencing everything from the real-time ad placements in a fraction of seconds, to complex media modeling and viewing.

This speed enables a rapid cycle of computations that directly impacts business outcomes, while the energy required to power this infrastructure poses a growing challenge. 

New Street Research, in a note published Friday, answered questions about Graphics Processing Unit (GPU) spend by Amazon, Google, and Meta Platforms. Analysts took a closer look at chip purchases and whether they were or will be installed, and the share of GPU-related spend for Google, Amazon, and Meta.

Knowledge and understanding of GPUs for marketers will help them make decisions for their company that best support campaign performance, despite the view that GPUs are highly technical and may be a challenge for advertisers and marketers to understand, and may not be necessary for them to perform their jobs.

GPUs -- the specialized electronic circuit used to rapidly process and render images and videos -- will support all types of innovative features built on generative artificial intelligence (GAI) that render those moving images in videos and ads.

New Street Research stated in a note that it expects Meta's share of GPUs purchased to pull ahead of Google and Amazon, but there is limited visibility into the timeline for production and deployment of Meta's MTIA custom chips.

“We currently expect custom silicon-related spend to remain below 10% of AI capex through 2027,” New Street Research stated in a research note. “Said another way, 90% of META AI capex is GPU-related.”

Custom silicon is the bigger driver at Google and Amazon, and the analyst firm raised estimates for the companies' GPU-related spending by ~29% and ~25% for 2026 and 2027, respectively, while custom silicon spending at both is expected to increase by ~25% and ~18% for 2026 and 2027.