Apple Intelligence: Bridging Hardware and Software in the AI Era

Written by Christopher De Sousa, CIM® | Associate Portfolio Manager | www.marnoa.ca

Apple finally revealed its long-awaited strategy for artificial intelligence (AI) at its Worldwide Developers Conference (WWDC). Apple has been relatively quiet about its AI strategy. Apple has been notably less vocal about AI, particularly generative AI, in their earnings calls compared to tech leaders like Microsoft, Facebook, Amazon, and Google, who have frequently discussed their AI strategies in recent years. This quickly resulted in increasing impatience among Apple’s investors. Experts started to doubt if Apple had a solid AI roadmap. How could a market leader in consumer technology not have clear AI initiatives? This question was repeatedly asked on CNBC every week.

Over the years, we’ve become well-acquainted with Apple’s story. Similar to Microsoft, Apple is among our longest-held holdings—a so-called legacy investment. Despite widespread skepticism about Apple’s seeming absence of an AI strategy, we maintained our belief in the company’s ability to execute an AI plan that could fuel a multi-year product upgrade cycle. Our confidence stemmed from Apple's extensive experience in integrating AI and machine learning (ML) capabilities into its hardware through proprietary silicon chips, which feature a built-in neural engine. This neural engine—named the Apple Neural Engine—is a specialized processor or neural processing unit embedded in Apple silicon chips that are specifically engineered to enhance the performance of AI and ML tasks.

The Apple Neural Engine was first introduced with the A11 Bionic chip in the iPhone X in 2017. The A11 chip was capable of performing up to 600 billion operations per second, enabling real-time processing of ML algorithms for innovative features like Face ID. Fast forward to today and we’ve seen an impressive progression in Apple’s chip technology. The latest A17 Pro chip housed in the latest iPhone 15 Pro and Pro Max contains a neural engine capable of performing 35 trillion operations per second. Yes, that’s trillion with a ‘T’! This represents a nearly 60-fold increase in processing power in just a few years. Just two months ago, in May, Apple introduced the new M4 chip for the iPad Pro that includes a neural engine capable of 38 trillion operations per second—the fastest neural processing unit of any AI PC today.

Apple has been quietly laying the groundwork for AI for several years. Apple’s silicon chips and Neural Engine technology will be the backbone of Apple Intelligence, the newly announced AI system with text and image generating tools. Apple Intelligence will be deeply integrated into iOS 18, iPadOS 18 and macOS Sequoia—the respective operating systems for the iPhone, iPad, and Mac.

Apple Intelligence will use both on-device generative AI models and larger server-based models. The on-device model allows the Apple user to perform many AI tasks locally on their iPhones, iPads, and Macs without sending data off the device. Apple Intelligence will draw on Apple silicon servers in the cloud using Private Cloud Compute to handle more complex tasks requiring additional processing power. Private Cloud Compute is Apple’s new platform designed to process complex AI requests in the cloud while protecting user privacy.

Privacy is a core part of Apple’s business model and brand identity. We believe that as AI becomes more prevalent in consumer products, privacy will become a critical factor in consumer choice. We think Apple’s emphasis on on-device processing and data protection will give the company an edge in developing privacy-preserving AI features. Apple is setting itself apart from other tech firms whose business models revolve around the collection and monetization of user data on their platforms. This differentiation is evident in Apple’s development of proprietary cloud security architectures like Private Cloud Compute which ensures that personal user data transmitted to the cloud is kept confidential, encrypted, and inaccessible to anyone other than the user—not even to Apple.

We think Apple Intelligence could drive the next iPhone upgrade cycle—the first major upgrade since the release of iPhone X in 2017. Apple Intelligence will be available exclusively on iPhone 15 Pro and Pro Max models, iPads and Macs equipped with M-series chips, and the upcoming iPhone 16 slated for release this fall. Consumers who want to access Apple Intelligence, including the new generative AI and enhanced Siri features, will need to upgrade to the latest Apple devices.

Investment firm Wedbush Securities estimates that as many as 270 million of the 1.5 billion active iPhones have not been upgraded in over four years.[1] This large pool of users with older devices presents a significant growth opportunity for Apple. For context, in the year after the iPhone X was released, iPhone revenue grew 18%.[2]

We also have the view that replacement cycles could be shortened going forward. The average iPhone upgrade cycle is approximately four years. In other words, the average iPhone user replaces their iPhone every four years. The next iPhone upgrade cycle is unique because it’s expected to be driven by both software and hardware upgrades, with a strong focus on AI-powered software features like Apple Intelligence. This is a shift from historical upgrade cycles that primarily focused on hardware improvements, such as faster processors, enhanced cameras, and design modifications.

We think software-centric upgrades like Apple Intelligence could create a compelling incentive for consumers to upgrade their devices more frequently as the new AI-powered features and capabilities offer significant utility and convenience. In contrast, we think that incremental hardware improvements, which have historically been Apple's approach to product upgrades, are less likely to motivate widespread upgrades across Apple's ecosystem.

Apple has an installed base of more than 2.2 billion devices, with the iPhone representing a significant portion of this figure. Srini Pajjuri, an analyst at Raymond James, estimates that for every 1% of active iPhones upgraded to newer models, Apple could generate additional revenue of $12 billion to $18 billion and increase earnings per share by 15 to 20 cents (see chart below). All in all, we think that Apple Intelligence is the potential upside catalyst to lower the replacement cycle for iPhones and drive material iPhone revenue upside.

We anticipate a future where AI capabilities will become increasingly advanced, supported by improved software and hardware in both processing power and memory capacity. As these technologies evolve, there's a strong possibility that newer, more sophisticated AI features will be exclusive to the latest iPhone models. This exclusivity could potentially encourage faster upgrade cycles among consumers, as they may be motivated to purchase newer devices to access these enhanced AI capabilities, which might not be available on older iPhone models.

[1] Bloomberg. Apple Is at Start of AI Supercycle. June 14, 2024. [LINK]

[2] Apple Inc. Form 10-K for the fiscal year ended September 29, 2018. [LINK]