This may be the most private and secure cloud-based AI solution around

What if you could build a business-specific generative AI (genAI) solution that was hosted in the cloud? What if that cloud used the most energy-efficient systems in the world? And what if those services were also all running on the world’s most secure computing platform, the Mac?

Wonder no more, because MacStadium and webAI just made this real.

We’ve talked about both companies before. MacStadium is a force to be reckoned with and offered the first major deployment of hosted Apple-as-a-service cloud solutions; webAI, meanwhile, has created quite a stir with its offering, which runs quite happily on a MacBook Air – and runs even better when hosted in the cloud. The two companies together now offer a service that lets you install and use webAI models on MacStadium-hosted racks of Macs. This delivers quite a lot to business users: it means they can run their own, self-trained genAI models to support their own unique business goals on Macs in the cloud, making those models available across their company as a hosted solution.

This beats other hosted approaches, as the service remains business-unique, protected by Mac security, and boosted by the lower energy costs Macs require. In other words, the combined solution should help businesses maintain their own data privacy and security policies by being hosted on a highly secure set of platforms. This isn’t for every company, of course — many will want to create their own Mac-based clusters. But the service does mean that companies of any size can now reasonably deploy and manage a bank of Macs to support their business with AI.

Et tu, Apple?

To my mind it also hints at a highly plausible future for Apple’s own server-based solution, Private Cloud Compute. It’s a reasonable next step to imagine Apple’s protected solutions offering up AI-as-a-service to enterprise customers, all protected by a low-energy, high-security promise. Even if Apple never does commence competing in the hosted cloud infra space, you now have the alliance between MacStadium and webAI.

“This collaboration with webAI represents a significant milestone in the evolution of AI for the enterprise,” MacStadium CEO Ken Tacelli said in a statement. “By combining our macOS cloud expertise with webAI’s interconnected model approach, we’re creating a platform that will fundamentally change how organizations develop and deploy advanced AI systems on Apple silicon.”

There are some significant advantages to what’s on offer — for example, the system supports up to 20,000 concurrent API requests per minute. In addition, the way webAI has been architected means you get an up to 30% reduction in model size with minimal loss of accuracy. It also makes it possible for enterprise customers to host and operate the world’s largest AI models natively on Apple Silicon.

That’s really important when you consider just how much more energy efficient Macs are at this level and should help cut a big chunk out of the environmental running costs of AI. That’s a good thing, not just in terms of the dollar per kWh, but also in terms of the wider impact of energy consumption. The service also makes use of AI inferencing, which helps reduce model size without undermining accuracy.

More affordable than Nvidia?

Cost is a positive force in these deployments. webAI co-founder and CEO David Stout told server-focused site, The New Stack, that on a cost-per-token basis, these Mac clusters are more affordable to run than Nvidia GPUs. “It’s more private than any other solution,” he said.

Working with MacStadium, Stout claims that the system should be seen as one of “the most secure systems for private processing off-site. And it’s going to be an AI-native solution, not something that we retrofitted to fit AI into the story.”

What this means to most businesses, of course, is that they can deploy powerful, cloud-hosted AI models on the world’s most secure platform at an affordable cost, which should enable even smaller business to securely deploy, and provision hosted AI solutions in order to compete. And all with a Mac.

You can follow me on social media! Join me on BlueSky,  LinkedIn, and Mastodon.