Artificial intelligence (AI) has been used for various purposes for a long time, including determining search results and filtering spam emails. However, the recent AI wave that has gained significant attention is generative AI, which includes image generators and chatbots capable of writing in natural language. This surge in interest is possible due to the improved hardware that can implement developers’ ideas. The hardware enables AI models to be trained with large amounts of data, requiring significant computing power.
AI hardware goes beyond acceleration and focuses on enabling tasks. For example, microcontrollers need to be fast enough to process AI applications. This development has raised questions about the intelligence of AI and its consequences for work, leisure, and society. Heise and The Decoder provide daily updates on the most important AI developments and examine different aspects of the AI revolution every Friday with experts.
Nvidia has been a prominent player in the AI accelerator market, with their hardware widely discussed in professional contexts since 2017. However, the rapid development of AI applications and associated tasks presents a challenge for those entering the billion-euro business of AI processors. AI training needs to be faster and more efficient. Data centers also require substantial space to accommodate the massive systems used for training new applications. Microsoft faces issues in Dublin, where the data centers’ sizes create problems for trucks delivering equipment.
Data centers also operate their own power plants to reduce dependence on general electricity prices. It remains unclear whether AI development will accelerate even further with the increased capacities provided by data centers or if the data centers will struggle to keep up with the development of new AI models. This constitutes a classic “chicken and egg problem” that cannot be fully resolved. However, it is certain that both data centers and AI development will continue to grow.
In conclusion, the advancement of AI accelerators and data centers plays a significant role in driving the AI revolution. The current hype surrounding generative AI has been made possible by improved hardware capabilities. The potential of AI and its impact on various aspects of life make it a topic of great interest and discussion.
