alk into any of the frontier AI labs, and you'll find wall-to-wall Macs.
Decisions Apple made years ago, long before ChatGPT and the LLM era, are now paying big dividends. In an exclusive conversation with The Deep View ahead of WWDC 2026 in June, Doug Brooks, senior product manager of Apple silicon, discussed how Apple's long-term investments in neural processing, unified memory, power-efficient computing, and hardware-software integration have positioned its devices for the AI era.
He also explained why Macs have become a favorite platform among AI developers, why Mac minis and Mac Studios are emerging as the preferred AI agent machines, and how Apple sees the future of on-device AI across iPhone, iPad, and Mac.
This interview has been edited for brevity and clarity.
Jason Hiner: Tell us where Apple silicon is today when it comes to AI.
Doug Brooks: We're seeing tremendous momentum with how people are using Apple products and Apple silicon, particularly in on-device AI workflows. We're really proud of that fact and of the way Apple silicon provides a strong foundation for AI. It really highlights the tent poles of Apple silicon. It's a very balanced architecture that provides CPU, GPU, unified memory, and the Neural Engine all contributing to performance across the chip. That's particularly important for these evolving agentic workflows. It's not just about the GPU crunching on an LLM anymore. It's about the whole chip contributing to different parts of the task, tool-calling, and the things that are happening around those workflows. It really plays to the strengths of Apple silicon.
At the same time, we're maintaining our core strengths around unified memory and incredibly power-efficient performance. We want to preserve great battery life, particularly on something like an iPhone, while still allowing people to take advantage of AI capabilities throughout everything they do. We think that has positioned us very well to take advantage of this continuing, fast-moving AI economy. On top of that strong silicon foundation, we provide robust software platforms so developers can tap into those capabilities. I like to joke that every transistor we put into a chip gets enabled through software so users and developers can actually take advantage of it. Technologies like Core ML [now Core AI], Metal, Metal Performance Shaders, and MLX allow developers to unlock the full capabilities of Apple silicon. We're seeing people do incredible things across the range of our products, and the momentum around AI capabilities continues to be phenomenal.
Jason Hiner: Apple was working on machine learning long before ChatGPT, and long before transformers and LLMs went mainstream. What were the origins of that journey?
Doug Brooks: I think a lot of it starts with having really good foundational building blocks. When we build Apple silicon, it's incredibly well-architected and scalable. Technology blocks that go into our A-series chips for iPhone can scale up into iPad, and then when we brought Apple silicon to the Mac, we were able to scale all the way up to our biggest and most powerful chips. All of that was done with a relentless focus on power efficiency. That allows you to provide incredible capabilities in a handheld device like an iPhone, and those same foundational principles become even more powerful when you bring them to Mac. From there, we've been able to add technologies that increase the platform's capabilities over time.
When we first introduced the Neural Engine, it was designed to solve very specific compute needs for machine learning. It delivers very high-performance, very power-efficient matrix math to solve those problems in a smart way. We also introduced dedicated accelerators where they made sense. We don't talk about it quite as much, but the CPU has dedicated neural accelerators, originally called ML accelerators, that accelerate AI problems in the CPU domain. The CPU is actually very good at certain kinds of AI workloads, especially low-latency workloads. Speech is a good example. A lot of that happens inside those accelerators. More recently, we brought neural accelerators to the GPU, which provides a huge boost in performance because the GPU is our most scalable engine. It provides scalable compute, scalable memory bandwidth, and taps into the unified memory architecture.
As you move from smaller systems to larger systems with more GPU resources, you're continually expanding the AI capabilities available to tackle bigger and bigger problems. So if you go back to the core of the question, the tent poles of Apple silicon, performance, efficiency, and unified memory, have played extremely well in the AI domain.
Jason Hiner: One of the things that's talked about a lot in AI right now is extreme co-design. How is that similar to the way Apple approaches vertical integration?
Doug Brooks: That's how I think about it as well. I'm glad you picked up on that because it's fundamental to how we design not just our chips but our systems. We're different from silicon vendors that build chips and then have to support lots of different customers, form factors, and use cases. We build a chip knowing exactly what systems it's going to go into. The chip can influence the system design, and the system design can influence the chip requirements. They're married together very tightly. I'd actually go a step further and include software in that equation. As we enable new capabilities in the chip, we want to make sure those capabilities are exposed all the way through the software stack so developers can benefit and customers can benefit. Because if those transistors go unused, what's the point? It's not about adding something and hoping somebody finds a use for it. We want those capabilities available to solve real problems. I think that's what allows us to build better systems and ultimately better products.
Jason Hiner: When I visit frontier AI labs like OpenAI, almost everyone seems to be using Macs. Why do you think that is?
Doug Brooks: Well, I certainly hope it's because we build great products. I think we've always had a strength in the developer community. Developers across not just app development but really every discipline. Whether you're a front-end developer, a back-end developer, or working in another area entirely, the Mac has been a strong developer platform for many years.
The productivity of macOS, the features we provide there, and the rich set of tools available have always been strengths. Apple silicon has been a strength there as well. When we look at things like compile times, automated testing, and development workflows, there's been a real strength on the Mac platform. Now when we look at AI specifically, so many of the tools in the industry are either Mac-only or Mac-first. I think that has really put us at the forefront within that developer community.
Jason Hiner: One of the trends we're seeing is developers wanting to run more AI locally because of privacy concerns, security requirements, and rising inference costs. Are you seeing that trend too?
Doug Brooks: Absolutely. You touched on a lot of the factors contributing to it. Security and privacy are certainly part of it. Why send your data, your code, or your intellectual property somewhere else if you can run it locally? Then there are the economics. The quantity of tokens being consumed is growing dramatically, especially with agentic workloads. I've seen estimates anywhere from three times to ten times higher demand as agents become more common.
People are realizing they have incredibly powerful devices sitting right on their desks that are capable of handling a wide range of workloads. What's really exciting is what modern optimization and quantization techniques have made possible. What you're now packing into a 70-billion-parameter model or even a 120-billion-parameter model is mind-blowing.
People are running these models on laptops. You see posts from developers doing software development at 35,000 feet on an airplane while completely disconnected because they're running powerful models locally.
I'm a big believer in using the right tool for the right job. Sometimes you want access to those very large frontier models in the cloud. But you can also do an enormous amount of work locally. I think the hybrid approach is really interesting because agents can decide what needs to happen locally and what needs to happen in the cloud based on the workload. The foundations of Apple silicon, performance, unified memory, power efficiency, and targeted acceleration, have put all of our systems in a good position to support that. We're seeing that particularly in the agentic space. We've talked publicly about the incredible demand for Mac minis and Mac Studios because they've become the platform for running so much of this work.
Jason Hiner: Let's talk more about Mac minis and Mac Studios. Why have they become such popular platforms for AI agents?
Doug Brooks: First and foremost, they're amazing platforms for AI in general and especially for these emerging agentic tools. They tap into the strengths of Apple silicon and unified memory in a very power-efficient way, and increasingly they're delivering compelling price-performance as well. For agentic workloads, people often want a system that's under their control, isolated from their primary machine, and capable of running 24 hours a day, seven days a week.
A Mac mini is an amazing system for that. It also goes back to something we talked about earlier. Many of these AI tools are either Mac-first or Mac-only. Being able to run cutting-edge AI software on a powerful Apple silicon platform has really enabled customers to adopt and push this technology forward.
Jason Hiner: How do you think about the future of on-device AI for iPhone and iPad?
**Doug Brooks: **One thing that really excites me is the work we've done around our foundation models and Apple Intelligence APIs, and how developers have adopted them. Sometimes I think of it as transparent AI. There are so many AI features embedded throughout the operating system and throughout applications that don't necessarily jump out at you as AI. They're powered by AI under the hood, but they're just helping people get things done. The work we've done with foundation models and exposing those capabilities to developers has allowed AI to become pervasive across applications in both small and large ways. Of course, there are also applications that are more obviously AI. I use on-device LLM apps like Localy AI on iPhone. One of my favorites is Draw Things, which runs across iPhone, iPad, and Mac. It's an AI image-generation application with highly optimized models.
We're also seeing very domain-specific applications emerge. When you think about an iPhone, it's not just an AI device. It also has cameras and sensors that provide context about the world around you. One example is SwingVision for tennis and pickleball. It tracks gameplay, recognizes what's happening, and provides insights. That's incredibly powerful AI happening directly on the device and augmenting the user in real time. It's a great example of the kinds of experiences developers are creating.
Jason Hiner: What are some of the capabilities you think developers still aren't taking full advantage of?
Doug Brooks: What continues to blow me away is how quickly AI capabilities are showing up across creative and entertainment applications. Every time a major software release comes out, so many of the new features are powered by AI. At NAB this year, for example, we saw new AI capabilities in tools like DaVinci Resolve that put a tremendous amount of power into the hands of artists and creators. They can create faster, iterate faster, and try new ideas. I mentioned Draw Things earlier. It has a highly optimized feature called Lightning Drafts where you can change a prompt, change a color, make something red instead of blue, and the image regenerates almost instantly.
It becomes a creative tool for exploration. I'm continually blown away by what developers are doing on Apple silicon and by how they're using AI to help people save time, be more productive, and be more creative. The speed of AI development right now is just crazy. I can't imagine where we're going to be a year from now, three months from now, or even a month from now. We're excited that Apple silicon is at the forefront of that and is helping developers and customers across all of our devices bring those experiences to life.