In 2024, the tech industry saw significant advancements in AI hardware, particularly in laptops and smartphones. Leading tech companies such as Qualcomm, Intel, AMD, Apple, and MediaTek introduced cutting-edge chips, enhancing the AI capabilities inside our favourite tech gadgets and setting new performance standards.
Here’s a look back at some of the top highlights of 2024 in terms of AI chips launches and upgrades…
This year can be labelled as the birth of the modern AI laptop, complete with a fully-functional NPU inside. After a lot of teasers and hints, it wasn’t until Computex 2024 in June when chipmakers AMD, Intel and Qualcomm showcased their recent advancements in the AI chip race.
Where Qualcomm unveiled their Snapdragon X Elite chip powered AI laptops, Intel only had prototype units of its AI laptop chip dubbed Lunar Lake, while AMD announced its Ryzen AI 300 chips for laptops at Computex 2024. Subsequently, first AMD (in August) followed by Intel (in September) revealed their AI chip powered laptops – AMD Ryzen AI 300 series and Intel Core Ultra 200V chips, respectively.
These developments signify a pivotal shift towards integrating advanced AI capabilities directly into consumer laptops, enhancing performance, efficiency, and overall end-user experience.
In March 2024, during NVIDIA’s GTC AI developer conference, the company unveiled its Blackwell architecture, introducing the B200 GPU and the Grace Blackwell GB200 superchip, capable of supporting larger AI models and enhancing inference capabilities – all while enhancing efficiency and reducing hardware footprints. Particularly, the Blackwell B200 GPU represents a significant leap in AI processing, offering up to 20 petaflops of compute performance, which is more than quadruple the capabilities of its predecessor, the Hopper H100. This hardware leap will no doubt accelerate AI training and inference tasks across various industries.
Also read: NVIDIA becomes world’s most valuable company: 3 reasons why
Despite initial production challenges and rumours of overheating earlier in the year, NVIDIA has addressed these issues, according to CEO Jensen Huang. He confirmed in November that production is moving “full steam” ahead, with shipments of the Blackwell chips set to begin in the fourth quarter of 2024.
In its response, Intel unveiled its Gaudi 3 AI accelerator chip as a more cost-effective and power-efficient competitor to NVIDIA’s H100 chip, hoping enterprise AI customers will choose it as a viable alternative for their AI training and inference needs.
Apple’s M4 chip, introduced in 2024, powers the latest MacBook Pro and iPad Pro models – along with the M4 Mac mini. According to Apple, the M4 chip features a 16-core Neural Engine, delivering up to 40% faster AI performance compared to its predecessor, the M3 chip. This enhancement accelerates machine learning tasks, including image and speech recognition, and enables more efficient on-device AI processing.
In comparison to competitors, the M4 outperforms Qualcomm’s Snapdragon X Elite, Intel’s Lunar Lake, and AMD’s Ryzen AI 300 series in single-core performance. However, in multi-core tasks, the competition remains robust, with each offering unique strengths in AI processing and power efficiency
Qualcomm’s Snapdragon 8 Elite and MediaTek’s Dimensity 9400 are set to power flagship smartphones in late 2024 and 2025.
The Snapdragon 8 Elite features Qualcomm’s second-generation custom Oryon CPU cores, with two high-performance cores clocked at 4.32 GHz and six efficiency cores at 3.53 GHz. It includes an upgraded Adreno GPU supporting ray tracing and a Hexagon NPU for advanced AI tasks, enabling features like on-device AI processing and enhanced camera capabilities.
MediaTek’s Dimensity 9400 utilizes the Armv9.2 architecture, comprising one prime Cortex-X925 core at 3.63 GHz, three Cortex-X4 cores at 3.3 GHz, and four Cortex-A720 cores at 2.4 GHz. It incorporates the Immortalis-G925 GPU, offering a 40% improvement in ray tracing performance, and the NPU 890 for AI processing. This chip supports advanced AI features, including real-time language translation and AI-enhanced photography.
Both chips are built on TSMC’s 3-nm process, offering significant improvements in AI performance and energy efficiency.
Apple’s iPhone 16 Pro series is powered by the A18 Pro SoC, featuring a 16-core Neural Engine capable of 31.6 trillion operations per second. This advancement enhances AI-driven features such as improved computational photography, real-time language translation, and augmented reality experiences. The A18 Pro’s efficiency also contributes to extended battery life, maintaining high performance without compromising endurance.
In a similar vein, Google’s Pixel 9 series utilizes the latest Tensor G4 chip, which is designed to elevate AI and machine learning capabilities. The Tensor G4 enhances on-device processing for features like real-time language translation, advanced photo editing, and personalized user experiences. Google’s integration of AI at the hardware level allows for more responsive and intuitive interactions, aiming to set a new standard for smart device functionality.
Also read: The ARM processors’ enterprise boom