Key Takeaway:
- Apple used Google’s TPUs to train its AI models, moving away from Nvidia’s GPUs.
- This strategic shift aims to diversify hardware resources and reduce Nvidia’s dependency.
- Google’s TPUs provide a cost-effective, advanced alternative, influencing industry trends.
- Future Apple AI features include enhanced natural language processing and generative AI.
- This shift highlights evolving competitive dynamics and strategic partnerships in tech.
In a significant revelation, Apple disclosed that it utilized Google’s custom Tensor Processing Units (TPUs) to train its artificial intelligence (AI) models.
This strategic decision marks a notable shift in the competitive landscape of AI technology and hardware. Apple has moved away from Nvidia’s graphics processing units (GPUs), traditionally a staple in AI model training.
Apple’s AI Training Strategy
In a recent technical paper, Apple outlined its AI training process, revealing that its AI system, Apple Intelligence, relies heavily on Google’s TPUs.
Wow, Apple Foundational Models, both on device and server, were fully trained on Google TPU clusters.
They didn’t use Nvidia in any training for it pic.twitter.com/Kn4G9AVfJn
— Max Weinbach (@MaxWinebach) July 29, 2024
Specifically, Apple’s AI models for iPhones and other devices were trained using 2,048 of Google’s TPUv5p chips, while its server models utilized 8,192 TPUv4 processors. Google’s Tensor processing units were initially created for internal workloads but are now being adopted more broadly.
According to their website, Google’s latest TPUs are priced at under $2 per hour of usage when reserved for three years in advance. Google launched its TPUs in 2015 for internal use and made them publicly available in 2017.
Today, they are some of the most advanced custom chips designed specifically for artificial intelligence applications. This decision underscores Apple’s strategic move to diversify its hardware resources and reduce dependency on Nvidia’s GPUs, which currently dominate the AI chip market.
Nvidia’s GPUs have been the cornerstone of AI training for many leading tech companies, including Microsoft, OpenAI, and Meta. However, the high demand and cost of Nvidia GPUs have prompted companies to seek alternative solutions.
Apple doesn’t hate NVIDIA, TPUs are just faster
— The Indomitable American Soul (@laplacesdust) July 29, 2024
Google’s TPUs, explicitly designed for AI workloads, offer a competitive alternative with advanced capabilities and potentially lower costs.
Apple’s adoption of Google’s TPUs highlights a growing trend among tech giants to develop or adopt custom AI hardware to enhance performance and manage costs effectively.
Apple’s decision to utilize Google’s TPUs has significant implications for the AI industry. This move suggests a shift towards more diverse and potentially cost-effective AI training solutions.
I know Apple and Nvidia have beef, but I didn’t expect them to have less beef than Google.
— Eshu Marneedi (@EshuMarneedi) July 29, 2024
It also reflects leading tech companies’ complex strategic partnerships and competitive dynamics. By leveraging Google’s advanced TPUs, Apple aims to boost the efficiency and scalability of its AI models, which could lead to more innovative AI-driven features and applications.
Looking forward, Apple plans to introduce several new AI features under the Apple Intelligence system. These advancements include improved natural language processing, AI-generated summaries, and enhanced functionalities for Siri.
Additionally, Apple is expected to integrate generative AI capabilities, such as image and emoji generation, across its devices. This integration will showcase the potential of AI technology in enhancing user experiences and expanding the functionalities of consumer electronics.
I heard it’s because googles infra is much more reliable compared to apple spinning up their own cluster and learning the knowledge in house. Apple is known to be Google clouds largest cloud customer for sure
— Jaeyoon Kim (@vibe_engineer) July 30, 2024
Apple’s strategic decision to train its artificial intelligence models on Google’s custom TPUs instead of Nvidia’s GPUs represents a significant development in the AI hardware landscape.
This choice underscores Apple’s innovative approach to AI technology and highlights the evolving competitive dynamics among major tech companies.
As Apple advances its AI capabilities, the industry can anticipate more groundbreaking developments and a broader adoption of diverse AI hardware solutions.
Other Apple News:
Apple’s use of Google’s advanced chips for AI training signifies a major shift in tech alliances, enhancing AI capabilities and reducing reliance on Nvidia.
29-07-24 – Apple’s AI Features Will be Missing from Upcoming iPhone Updates
17-07-24 – Apple and Others Secretly Trained AI on YouTube Creators’ Videos!
16-07-24 – AI iPhone Hype Pushes Apple Stock to New Heights, Analysts Predict $5 Trillion Valuation!
10-07-24 – OpenAI Board Update: Microsoft and Apple Withdraw Observer Roles
04-07-24 – Apple Reportedly Secures Observer Seat on OpenAI Board!
01-07-24 – Apple’s Device Longevity Makes AI Indispensable!
For more news and trends, visit AI News on our website.
