Managing such a gigantic array of chips to develop Llama 4 will likely pose unique engineering challenges and require enormous amounts of energy. On Wednesday, Meta executives sidestepped a question from analysts about energy restrictions in parts of the United States that are hampering companies’ efforts to develop more powerful artificial intelligence.
According to one estimatea cluster of 100,000 H100 chips would require 150 megawatts of power. Largest domestic laboratory supercomputer in the United States, Captainhowever, it requires 30 megawatts of power. Meta expects to spend as much as $40 billion equipping data centers and other infrastructure this year, an augment of more than 42 percent from 2023. The company expects these expenses to augment even more dramatically next year.
Meta’s total operating costs have increased by approximately 9 percent this year. But overall sales – mostly from advertising – rose more than 22 percent, leaving the company with bigger margins and bigger profits even as it pumps billions of dollars into Lamy’s efforts.
Meanwhile, OpenAI, considered the current leader in cutting-edge AI development, is wasting money even though it charges developers for access to its models. What remains for now a non-profit venture claims to be training GPT-5, a successor to the model that currently supports ChatGPT. OpenAI has said that GPT-5 will be larger than its predecessor, but has not said anything about the computer cluster it is using for training purposes. OpenAI also said that in addition to scale, GPT-5 will feature other innovations, including a recently developed approach to reasoning.
CEO Sam Altman he said that GPT-5 will be a “significant step forward” compared to its predecessor. Last week, Altman responded to news reports stating that OpenAI’s next pioneering model would be released in December writing on X “fake news is out of control.”
On Tuesday, Google CEO Sundar Pichai said the latest version of Gemini’s family of generative artificial intelligence models is in development.
Meta’s open approach to artificial intelligence has proven controversial at times. Some artificial intelligence experts fear that making much more powerful artificial intelligence models available for free could be risky because they could facilitate criminals carry out cyberattacks or automate the design of chemical or biological weapons. Although Llama was enhanced prior to release to reduce inappropriate behavior, removing these restrictions is relatively plain.
Zuckerberg remains bullish on open source strategies even as Google and OpenAI promote proprietary systems. “It seems pretty clear to me that open source will be the most cost-effective, customizable, trustworthy, efficient and easiest-to-use option available to developers,” he said Wednesday. “I’m proud that Lama is a leader in this.”
Zuckerberg added that Llama 4’s recent capabilities should enable support for a wider range of features across Meta services. Currently, Llama’s core model-based offering is a ChatGPT-like chatbot known as Meta AI, available on Facebook, Instagram, WhatsApp and other apps.
Zuckerberg said that more than 500 million people apply Meta AI every month. Over time, Meta expects to generate revenue through advertising on this feature. “People will use more and more queries, and over time, as we get there, monetization opportunities will emerge,” Meta’s chief financial officer, Susan Li, said on Wednesday’s call. With its ad revenue potential, Meta might just be able to subsidize Llama for everyone else.