Bank of America Securities is framing Apple’s 2026 Worldwide Developers Conference AI announcements as a meaningful redesign of the company’s AI stack that could strengthen ecosystem ties and underpin subsequent hardware upgrade cycles. The brokerage reiterated its Buy rating on the iPhone maker and kept a $380 price objective, arguing that investors may be underestimating the implications of Apple’s new Siri architecture.
At the core of BofA’s view is the repositioning of Siri from a standalone assistant into a context-aware, multimodal intelligence layer. The firm describes the revamped assistant as capable of drawing on personal data, screen content and app activity across Apple’s product lineup - including iPhones, iPads, Macs and Watches - which could increase engagement with Apple’s services and deepen user retention inside the ecosystem.
Apple unveiled a hybrid AI approach that blends on-device computation, a Private Cloud Compute (PCC) tier and cloud infrastructure powered by Google Cloud and Nvidia GPUs for more demanding workloads. As described by BofA, the architecture aims to route simpler requests locally on devices while delegating heavier reasoning and image-generation tasks to cloud-based resources when required.
A central technical announcement was the introduction of a five-model Apple Foundation Models family, dubbed AFM 3. That suite includes two models intended to run on devices and three cloud-based models designed to tackle a range of needs - from routine queries to advanced reasoning and image creation. Among these, BofA singled out AFM 3 Core Advanced as particularly notable.
AFM 3 Core Advanced is a 20-billion-parameter sparse model that activates only a subset of parameters for each request. According to the brokerage, this sparsity lets Apple deliver stronger on-device capabilities while keeping memory and compute requirements lower than full-dense models. The expected benefits, BofA says, include improved response quality, enhanced privacy and reduced latency, as well as less dependence on continuously expensive cloud compute.
Despite the emphasis on local processing, BofA flagged that cloud usage and its cost profile remain material to Apple’s margins. The analysts estimate that by 2030 roughly half of AI requests could be handled on-device. They also project that the most advanced Cloud Pro model might represent only about 5% of requests yet could account for as much as 67% of weighted cloud-computing costs because of its higher processing intensity.
That math, BofA warns, makes efficient routing of AI workloads critical to preserving profitability as adoption of Apple Intelligence scales. The brokerage also noted Apple’s collaboration with Google for cloud-based models as a pragmatic step to accelerate generative AI capabilities, while observing that how often users call on complex cloud reasoning and image generation will influence future margin sensitivity.
Overall, BofA maintained that Apple is well positioned given the potential for AI-driven iPhone upgrade cycles, growth in services revenue and continued shareholder returns. Still, the firm emphasized that successful execution of the new Siri AI platform will be a central determinant of the company’s long-term growth trajectory.
Key points
- BofA holds a Buy rating on Apple with a $380 price objective, citing the strategic importance of the new Siri architecture.
- Apple’s hybrid AI stack - combining on-device processing, Private Cloud Compute and cloud services via Google Cloud and Nvidia GPUs - aims to balance privacy, latency and compute cost.
- The AFM 3 family, particularly the 20-billion-parameter sparse AFM 3 Core Advanced model, is expected to enhance on-device AI capability while limiting memory and compute demands.
Sectors impacted: Consumer technology hardware, cloud infrastructure and semiconductor/cloud GPU providers.
Risks and uncertainties
- Margin sensitivity to cloud usage - if users rely heavily on cloud-based reasoning and image-generation functions, cloud costs could rise and compress margins, affecting profitability in technology and cloud services segments.
- Execution risk for the new Siri platform - the long-term benefits BofA cites depend on Apple successfully deploying and scaling the revamped AI architecture across devices and services, which is critical for consumer hardware and services revenue outcomes.
- Workload routing efficiency - preserving profitability hinges on routing workloads effectively between on-device, PCC and cloud layers; inefficiencies could increase compute expenses for Apple and influence related cloud and GPU markets.
Note: The analysis above reflects Bank of America Securities’ assessment as presented regarding Apple’s 2026 AI announcements and does not introduce additional facts beyond that assessment.