In a recent update from “Eye on AI,” several significant developments regarding Artificial Intelligence (AI) have been reported. The newsletter highlighted that Big Tech companies such as Microsoft, Meta, and Google are significantly increasing their expenditures on AI development and infrastructure, largely focusing on hardware like Nvidia chips and data centers. Nvidia, however, will provide insights into the AI surge with its upcoming quarterly report.
Meta has announced plans to increase capital expenditures considerably in 2025 due to heightened infrastructure investments. The company adjusted its 2024 capital expenditures guidance to range from $38 billion to $40 billion, a slight increase from the previous estimate. CEO Mark Zuckerberg emphasized the importance of infrastructure for AI investments during the recent earnings call. Meta continues to develop its Llama model and is asserting its identity as an AI-focused company, largely supported by advertising revenue from its social media platforms. Zuckerberg noted improvements in AI have resulted in increased engagement on Facebook and Instagram.
Microsoft has reported positive growth in its AI business, projecting an annual revenue run rate surpassing $10 billion, making it the fastest-growing product category for the company. For the first time, it disclosed its investment in OpenAI, revealing a commitment of $13 billion, excluding an additional $750 million invested in OpenAI’s latest funding round. These investments, however, have impacted Microsoft’s profit margins as losses from its stake in OpenAI contributed to a $683 million expense.
Google highlighted that the AI boom has bolstered its cloud business, with AI now creating over 25% of new code at the company, although engineers still review the output before implementation. There are questions about the specifics of the process and the effectiveness of AI in coding projects, as no further insights were disclosed by Google.
In other AI news, OpenAI has decided to forgo establishing its own semiconductor foundries due to the prohibitive costs and time involved, opting instead to work with Broadcom on AI chip development, secured through Taiwan Semiconductor Manufacturing Company (TSMC). The timeline for this initiative remains flexible, with the first custom-designed chip expected in 2026.
Additionally, investment firms KKR and Energy Capital Partners are committing $50 billion over the next four years to enhance energy access for AI, focusing on natural gas and exploring cleaner energy sources such as nuclear power.
In Portland, Oregon, AI aimed at identifying customers needing assistance with water bills mistakenly offered a discount to a wealthy individual, Columbia Sportswear CEO Tim Boyle. The Portland Water Bureau has allocated $350,000 for this pilot project.
Finally, current AI models frequently deliver incorrect information on voting-related inquiries, with Spanish language queries showing a higher inaccuracy rate compared to English. Testing by Proof News found inaccuracies in up to half of the responses.
The Eye on AI newsletter, providing these insights, invites readers to engage further through its subscription service.
