Apple is launching upgraded Mac computers designed to run artificial intelligence tasks locally, positioning them as cheaper alternatives to paying cloud providers like OpenAI for computing power. The new Mac minis and Mac Studios, which started shipping this week and can cost nearly $20,000, are aimed at corporate buyers who want to avoid paying per-token fees for cloud-based AI services.

Why Is Apple Suddenly Competing in Enterprise Computing?

Apple has historically avoided the enterprise desktop market, where Windows dominates with 91.3% market share compared to Apple's 4.6%. But the company's decades-long focus on power efficiency in devices like iPhones has unexpectedly positioned it well for a new market: on-device AI computing. When Apple introduced its first Apple Silicon chips in 2020, it combined computing and memory into a unified architecture that improved battery life. That same design choice, which Nvidia and others only recently adopted, makes Macs particularly efficient at running AI models locally.

The shift reflects a broader industry trend toward what Microsoft CEO Satya Nadella calls "unmetered intelligence," where AI runs directly on a user's device rather than in a distant data center. Apple is betting that companies will prefer paying once for a machine and using it repeatedly over paying per token for cloud services.

What Can These New Macs Actually Do?

At its launch event this month, Apple demonstrated four Mac Studios networked together running an AI model with a trillion parameters, a measure of complexity, to find and fix a graphics coding bug. Tasks of this scale typically require data center infrastructure, but the stack of Macs ran off a single wall outlet. The machines include exotic networking features like RDMA over Thunderbolt, a chip-to-chip communication technology that Apple has quietly added over the past two years.

The new Macs are aimed at intense AI tasks such as writing code or handling complex business work without relying on external AI services. Apple's pitch emphasizes that once a company purchases the hardware, there are no ongoing per-token costs, only the electricity to run the machines.

How to Evaluate On-Device AI for Your Organization

• Cost Structure: Compare the upfront hardware cost against your organization's annual cloud AI spending. If you pay significant per-token fees to services like OpenAI or Anthropic, on-device AI may offer faster payback.
• Task Complexity: Assess whether your AI workloads can run locally. Simpler tasks like code generation or business analysis are good candidates; cutting-edge research requiring the latest models may still need cloud access.
• Scalability Needs: Consider whether you need to scale from desktop machines to larger systems. Apple's unified chip architecture allows models developed on Mac minis to scale up to Mac Studios or down to iPhones and iPads without redesign.

Johny Srouji, Apple's chief hardware officer, emphasized the value proposition in stark terms: "Once you have the machine on your desk, you've paid for it. And I believe we provide absolutely great value, not only in terms of performance, but cost. There's no cost per token. You're just using the machine again and again".

Johny Srouji, Apple's chief hardware officer

How Does This Challenge Nvidia's Business Model?

Nvidia's stronghold remains the data center, where companies rent computing power for AI training and inference. At the launch of its new PC chip this summer, CEO Jensen Huang downplayed any intention to compete directly with Apple, saying Nvidia is focused on expanding what Windows PCs can accomplish. However, the emergence of on-device AI represents a fundamental shift in how companies think about AI infrastructure.

Microsoft is pursuing a similar strategy, working with chip partners to streamline AI work through its Windows ML tools and investing in features like RDMA. The competition reflects a broader industry recognition that not all AI tasks require expensive cloud infrastructure, and that efficiency gains from local processing can justify the upfront hardware investment.

Apple's pitch to corporate buyers centers on a simple economic argument: the cost of ownership for a $20,000 machine that runs AI models repeatedly is lower than the cumulative cost of cloud tokens for the same work over time. Whether enterprises embrace this model will determine whether Apple can meaningfully expand beyond its current 4.6% enterprise desktop share.