Dive Brief:
- Apple debuted two new chips for its desktop computers, furthering its efforts to draw enterprise attention to the AI capabilities of its devices. The M6 chip in Apple’s new Mac mini and M5 Ultra in Mac Studio deliver advanced compute and power efficiency, the tech giant said in a Tuesday announcement.
- The M6 chip features increased unified memory bandwidth, a larger 12-core CPU complex and 12-core GPU with neural accelerators, a set of enhancements that Apple hopes will draw enterprises as they pursue greater compute power. The M5 Ultra chip also includes larger CPU and GPU cores and 50% more memory bandwidth than its predecessor, the M3 Ultra.
- The new chips have been created to enable future AI workloads, including agents, in the enterprise, said Ranjit Atwal, senior director analyst at Gartner. “This is allowing agentic workloads to operate in a way they need to work,” Atwal told CIO Dive.
Dive Insight:
Apple’s latest chip and PC advancements signal a continued push toward the enterprise.
CEO Tim Cook, who led the company’s growth to a trillion dollar valuation, described the enterprise market as a promising area of growth for the company during a 2015 earnings call. The executive will step down Sept. 1 as John Ternus, SVP of hardware engineering, takes over.
Mac revenue grew 29% year over year, reaching $10.4 billion, during the third quarter of 2026, Cook said during Apple’s earnings call in July. Cook described Mac as the “ultimate AI powerhouse” with capabilities including high throughput and on-device inference.
Creative teams at Disney — one of Apple’s enterprise customers — are using Mac for on-device AI workflows that reduce token costs, according to Kevan Parekh, senior VP and CFO at Apple. Meanwhile, French retail bank Credit Agricole is using on-device AI to streamline regulatory workflows.
“More companies are choosing Mac for on-device AI advantages, including lower costs, better performance and enhanced privacy and security,” Parekh said during the call.
One challenge facing enterprises is that Apple devices are high-end and expensive, Atwal said. However, enterprises are experiencing rising AI costs, as agentic AI in particular consumes large amounts of tokens. The technology can run in continuous loops, querying and reprocessing data, racking up higher and higher bills.
Moving workloads to devices that enable more powerful models could help enterprises mitigate both token costs and complexity as the technology evolves, Atwal said.
“I don’t think that budget necessarily comes from a traditional place in enterprises,” he said. “The budget for this type of high-end, what I call an infrastructure PC, will come from the infrastructure part of an organization. This then becomes part of an organization thinking about cloud to device strategy.”
The enhanced memory capacity of Apple’s latest chips, particularly the M5 Ultra, can enable models with hundreds of billions of parameters to run on a desktop, a level of on-device inference Nvidia and AMD haven’t caught up to yet, Brendan Burke, research director of semiconductors, supply chain and emerging tech at The Futurum Group, told CIO Dive in an email.
“Enterprises want frontier models running on hardware they control and memory capacity is what makes that possible,” he said. “The test is whether local inference at this scale pulls enterprise workloads off rented GPUs and onto the desk.”
