Apple Inc. (NASDAQ: AAPL) is positioning its latest Mac hardware as a cost-effective alternative to cloud-based AI computing, aiming to undercut the economics of competitors like Nvidia. Apple hardware chief Johny Srouji told Reuters that the company's newest Macs operate with “no cost per token,” suggesting that businesses can pay for the machine upfront and run AI models locally without incurring ongoing usage fees.
To demonstrate the capability, Apple showed four Mac Studios running a trillion-parameter AI model together. Reuters reported that the cluster could operate from a single wall outlet, highlighting the efficiency of the setup.
The company is targeting enterprise customers with this pitch, suggesting that companies can shift inference workloads from rented data-center capacity to local machines. However, the hardware required is expensive. The M5 Ultra Mac Studio starts at $5,499, while a configuration with 256GB of memory and 16TB of storage costs $18,299. A 512GB configuration, capable of handling even larger models, is set to arrive in late October.
Apple raised Mac prices in June, citing rising costs for memory and storage. This move contrasts with the business model of Nvidia, whose CEO Jensen Huang has stated that “compute is revenue.” Nvidia’s data center revenue reached $89 billion in the most recent quarter, a 117% increase from the previous year, driven by spending from cloud providers and AI companies.
Nvidia is also pursuing the local AI market with products like the DGX Spark and RTX Spark, designed to run AI models directly on desktops. Despite Apple's push, prediction markets currently favor Nvidia. Polymarket gives Nvidia a 78% chance of being the world’s largest company by the end of 2026, compared to 15% for Apple. Additionally, IDC data cited by Reuters shows Apple held just 4.6% of the enterprise PC market, while Windows held 91.3%.
