Sunday, May 19, 2024
- Advertisement -
More

    Latest Posts

    It will take a multi-year investment cycle before Meta’s AI offerings become profitable: Insights from Meta’s Earnings Call

    Meta indicates that scaling its artificial intelligence (AI) offerings into profitable businesses may require multi-year investment cycles. The company said that building and scaling experiences have historically been a long-term investment process for it and its investors. “..and the initial signs are quite positive here [with AI] too. But building the leading AI will also be a larger undertaking than the other experiences we’ve added to our apps and this is likely going to take several years,” Mark Zuckerberg, the CEO of Meta said during the company’s earnings call.

    Zuckerberg added that, based on Meta’s history of successfully monetizing scaled technologies, the company is confident in its future ability to monetize these AI offerings. “There are several ways to build a massive business here, including scaling business messaging, introducing ads or paid content into AI interactions and enabling people to pay to use bigger AI models and access more compute,” he said, explaining how AI could add value.

    He mentioned that the company would be focused on improving the efficiency of Meta AI and other AI services as it continues to scale them. “Some of this will come from improving how we train and run models. Some improvements will come from the open source community and we’re improving cost efficiency as one of the main areas that I expect that open sourcing will help us improve, similar to what we saw with open compute,” Zuckerberg shared.

    AI’s impact on Meta’s capital expenditure:

    The company anticipates that in 2024, capital expenditures will be in the range of $35 billion to $40 billion increased from the range which it previously estimated ($30 billion-$37 billion). This increased expenditure is a byproduct of infrastructure investments to support Meta’s AI roadmap. It expects that capital expenditure will continue to increase as it invests in AI.

    “In terms of measuring the ROI [return on investment] on our CapEx [capital expenditure] investments, we’ve broadly categorized our AI investments into two buckets. I think of them as sort of core AI work and then strategic bets, which would include Gen AI and the advanced research efforts to support that,” Meta’s CFO Susan Li said. She explained that for core AI work, Meta continues to have an ROI-driven approach and is seeing strong returns as well. For strategic bets, which include potential businesses that the company could build, it will have to invest in developing more advanced models before they drive meaningful revenue.

    Key AI-related developments mentioned during the call:

    AI recommendation system: As of now, about 30% of the posts on the Facebook feed are delivered by Meta’s AI recommendation system. “That’s up 2x over the last couple of years and for the first time ever, more than 50% of the content that people see on Instagram is now AI recommended,” Zuckerberg mentioned.

    Historically, each of Meta’s functionalities (like reels, in-feed recommendations etc) has had their own AI models, but more recently, the company has begun shifting to a new model architecture with the aim for it to power multiple recommendation products. This model began being used to power Facebook reels last year and led to an 8-10% improvement in their performance. This model has now been extended to Facebook’s video tab.

    Launch of Meta AI: Meta AI, Meta’s AI assistant was launched in April this year. The AI assistant was also incorporated into Meta’s collaboration with Ray-Ban. “Glasses are the ideal device for an AI assistant because you can let them see what you see and hear what you hear, so they have full context on what’s going on around you as they help you with whatever you’re trying to do,” Zuckerberg said explaining the rationale behind incorporating Meta AI with vision.

    Meta’s chip-building aspirations: The company intends to continue working on building its own silicon for AI models. “Our Meta training and inference accelerator chip has successfully enabled us to run some of our recommendations related workloads on this less expensive stack,” Zuckerberg said, adding that as the program matures, it will be able to support other workloads as well.


    STAY ON TOP OF TECH NEWS: Our daily newsletter with the top story of the day from MediaNama, delivered to your inbox before 9 AM. Click here to sign up today!


    Also read:

     

    The post It will take a multi-year investment cycle before Meta’s AI offerings become profitable: Insights from Meta’s Earnings Call appeared first on MediaNama.

    Latest Posts

    - Advertisement -

    Don't Miss

    Stay in touch

    To be updated with all the latest news, offers and special announcements.