Meta Cuts Hiring Costs to Focus on AI Development

Whether due to service price increases, weak consumer demand, continuously shifting trends, or still experiencing the economic effects caused by the pandemic, layoffs have become and continue to be common amongst tech, especially Big Tech, companies.

With the many company expenses that often become larger and larger, the first cost-cutting measures that come to mind in Big Tech is cutting costs on hiring and wages, especially when resources need to be reallocated for them to focus on big projects.

This is the case for social media company Meta, previously Facebook, run and built by CEO Mark Zuckerberg. Actively pursuing further development in AI to compete with other Big Tech companies such as Amazon, Alphabet, and Microsoft, Meta plans to keep hiring costs to a minimum, dedicating its resources to AI model training and infrastructure.

Financial Measures

Layoffs aren’t a new thing to Meta. During an earnings call made by Meta execs, Zuckerberg referred to 2023 as a “Year of Efficiency,” filled with company policies, decisions, and actions that proved successful, prompting the company to adopt the same, or at the very least a similar, approach for the upcoming years.

The action includes a couple of massive layoff rounds that started in the third fiscal quarter of 2022. Up until the end of 2023, Meta witnessed a 22% decrease in its total employee headcount, from over 86K the previous year to about 67K.

Meta believes that the layoffs have played a monumental role, resulting in its positive 2023 fourth-quarter stats. The company reported a 25% growth in sales, reaching over $40B. The company’s stock had also almost tripled, which led to it being rated 2nd in the S&P 500, following Nvidia. Meta’s net income also rose, extending as far as 201%, valued at $14B.

While hiring stopped being a priority for Meta, it does not mean the company has stopped taking in new hires altogether, with Zuckerberg stating that the employee headcount increase will just be “relatively minimal,” especially compared to Meta’s previous years. Instead of focusing on quantity, Meta is choosing to prioritize quality, only accepting those with higher experience levels and great technical expertise.

The company has made statements insinuating that mass layoffs are completely out of the picture in 2024, and they will try to “keep things lean” just by being more fastidious in hiring. However, still trying to run on peak efficiency, Zuckerberg plans to eliminate roles the company finds unessential, as he once did with technical program manager roles at Instagram sometime last year. According to Business Insider, this prompted a policy where affected employees are given the opportunity to reapply and reinterview for available roles, in which, if they choose not to, they will be laid off.

All of these efforts in cutting costs were made with two goals in mind, which, according to Zuckerberg, include focusing on training Meta’s AI models and further developing Meta’s Reality Labs, the unit behind the Metaverse.

Focusing on AI and the Metaverse

With the massive AI market still continuing to grow, Meta aims to reallocate the resources they once used in wages and hiring processes into massively building up and advancing its AI model by further developing its AI infrastructure, turning it into a “world-classs compute infrastructure.”

One of the main ways they aim to do this is by spending billions on Nvidia’s H100 AI chips, which are the favored, and so presumably superior, AI chip for most, if not all, companies focusing on the field of AI at the moment.

It is reported that Meta had already purchased about 150K of Nvidia’s H100 chips in 2023, which, when summed up with Nvidia’s A100s and all the other AI chips Meta already has, could potentially lead to almost 600K GPUs by the end of 2024. Even without adding the amount of GPUs that Meta will most definitely get more of this year, the company has most probably tripled the number of GPUs other companies have, only tied with Microsoft.

The work done to create Llama 2, Meta’s latest open-source AI model released last July, cultivated an amplified ambition in Zuckerberg, to not only start working on and training Llama 3, which is what all the new GPUs and strong infrastructure are intended for, but to also start reaching for AGI soon, if feasible.

As for the Metaverse, though Meta’s Reality Labs unit has actually bled out costs and the hype for Metaverse has died down, Zuckerberg still has faith in the Metaverse remaining significant, especially as Apple Vision Pro’s release presented an existing interest in the niche field of AR/VR.

Zuckerberg hopes that all the decisions made in changing how Meta operates can give the company “the ability to go through what is a somewhat unpredictable and volatile period,” in the tech industry, which Zuckerberg predicts will continue to last as far as the next five to 10 years.

Meme & AI-Generated Picture

Job Posting

• Meta - Display Systems Engineer - Sunnyvale, CA+ (In-Office/Hybrid)

• Meta - FPGA and Emulation Engineer - Sunnyvale, CA (In-Office/Hybrid)

• Meta - Reality Labs UX Research Lead, Youth & Integrity - Burlingame, CA (Remote/Hybrid)

• Meta - Research Scientist Intern, AI Systems Machine Learning - Menlo Park, CA+ (In-Office/Hybrid)

Promote your product/service to Digger Insights’ Community

Advertise with Digger Insights. Digger Insights’ Miners are professionals and business owners with diverse Industry backgrounds who are looking for interesting and helpful tools, products, services, jobs, events, apps, and books. Email us [email protected]

Your feedback would be greatly appreciated. Send it to [email protected]