AI Adoption Still in 'Very Early' Stages, BlackRock Exec Says | PYMNTS.com
Dec 09, 2024 - pymnts.com
BlackRock anticipates significant investments in infrastructure and cybersecurity by 2025, driven by the ongoing AI boom. Jay Jacobs, BlackRock’s U.S. head of thematic and active ETFs, highlighted the early stage of AI adoption and the need for AI firms to expand data centers while enhancing cybersecurity. He emphasized the importance of investing in the physical infrastructure supporting AI, such as power, chips, and real estate, as these elements are crucial for AI's functionality. The costly nature of this infrastructure is evident in examples like Microsoft's partnership with OpenAI, which required building AI supercomputers powered by Nvidia GPUs, consuming substantial energy.
The article also discusses the industry's shift towards AI optimization, where major tech companies are working to refine software systems to improve performance while reducing computing power needs. This involves innovations in software architecture, such as Google's quantization techniques and Meta's efficiency gains with Llama AI models. These advancements aim to transform the economics of AI operations into sustainable models, allowing companies to manage massive computing systems more efficiently.
Key takeaways:
```html
BlackRock anticipates significant investments in infrastructure and cybersecurity by 2025, driven by the AI boom.
AI adoption is still in its early stages, necessitating the expansion of data centers and increased cybersecurity measures.
The infrastructure supporting AI, including power, chips, and real estate, is crucial and requires substantial investment.
Major tech companies are focusing on AI optimization to enhance performance and reduce the computing power needed, transforming the economics of AI operations.
