Apple Neural Engine development traced back to failed self-driving car program
Apple's discontinued Project Titan autonomous vehicle program was the primary driver behind the development of the Neural Engine technology now found in Apple silicon. This proprietary NPU architecture provides the foundations for the company's current on-device AI, computational photography, and video processing capabilities.
Key Takeaways
- Neural Engine performance increased from 600 billion operations per second in the 2017 A11 Bionic to 40 trillion in the current M3 Max.
- Project Titan's requirement for real-time sensor fusion forced Apple to solve on-device AI processing years before the generative AI boom.
- Apple reportedly invested an estimated $10 billion into the failed vehicle project over a decade-long lifespan.
- Technological transfer from the car project enabled Apple to avoid cloud-side latency for Siri and video editing tasks.
Why It Matters
The repurposing of automotive silicon for mobile hardware gives Apple a distinct advantage in edge video processing and privacy. By offloading complex AI workloads to local hardware, Apple reduces dependency on server-side compute, an area currently dominated by Nvidia and cloud-centric rivals like Google. This vertical integration allows for sophisticated real-time video manipulation that competitors often struggle to match without network-induced latency. Strategists should monitor future Neural Engine iterations, specifically the rumored M7 Ultra, as a signal for when Apple might transition toward fully local large language models for creative video applications.
Additional Context
Following the formal termination of Project Titan in February 2024, Apple shifted approximately 2,000 employees from the automotive division to its generative AI groups led by John Giannandrea, per Bloomberg. This transition occurred as the company faced increasing pressure to match AI advancements from industry leaders like Microsoft and OpenAI. The internal restructuring served as a catalyst for the 'Apple Intelligence' features unveiled at WWDC in June 2024, which mandate a minimum of 8GB of RAM and specialized NPUs, such as the A17 Pro or M-series chips, for local execution. The hardware legacy of the car project became more evident with the May 2024 launch of the M4 chip in the iPad Pro, which Apple described as its most powerful Neural Engine to date. Capable of 38 trillion operations per second, the M4's NPU represents a 60x speed increase over the original A11 Bionic released in 2017, per official Apple technical documentation. This rapid scaling of hardware performance has allowed Apple to maintain a lead in the 'AI PC' and premium smartphone markets, even as global shipment growth remains modest. Simultaneously, competitors are aggressively pivoting to match this NPU-first architecture. Per Canalys and IDC data from early 2025, the global smartphone market grew roughly 7% in 2024, driven largely by 'AI-ready' devices. While Qualcomm's Snapdragon X Elite and Samsung's latest Exynos processors have reached parity in raw TOPS (trillions of operations per second) metrics, Apple’s early integration of the Neural Engine—inherited from the autonomous vehicle stack—remains the benchmark for sustained, power-efficient on-device video and image processing.
Read full article at techbuzz.ai
Enjoy our coverage?
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source