Meta’s AI division is reportedly going through a big shake-up. According to a Bloomberg report, the company is splitting its AI unit into four different groups, each with a separate focus. On paper, this looks like a way to redefine objectives and avoid confusion. But behind the scenes, AI engineers and researchers seem worried. There are already talks of job cuts and reassignments, even though Meta has spent the last year aggressively hiring AI talent with high salaries.

In other words, this sudden shift in tone from expansion to possible downsizing has left staff unsettled. It has also, as the move comes just days after a rather muted GPT-5 launch from OpenAI, sparked speculation that the AI hype is already slowing down, and suddenly various tech companies might be curtailing their AI ambitions.

Restructuring under new leadership

At Meta, the new structure of the AI division is being led by the company’s recently appointed chief AI officer and former CEO of Scale AI, Alexandr Wang. In an internal note, Wang wrote that Meta needs to be better organised if it wants to reach its long-term goal of building superintelligence. “Superintelligence is coming, and in order to take it seriously, we need to organise around the key areas that will be critical to reach it — research, product and infra,” he wrote.

As part of this plan, the AI division — now renamed Meta Superintelligence Labs (MSL) — will run as four teams. One team will handle large language models such as Llama. Another, under the FAIR banner, will continue long-term AI research. A third, led by former GitHub CEO Nat Friedman, will work on consumer-facing AI products. The fourth will focus on infrastructure, including data centres and computing hardware.

But while this may look tidy in theory, not everyone is convinced. As per a New York Times report, Meta is considering firing some AI engineers or moving staff around within its AI workforce, which now numbers several thousand. Of course, nothing is confirmed yet, but the uncertainty has definitely rattled employees. Some executives have already left, adding to the sense of instability. For a division once described as Meta’s biggest growth driver, even the possibility of layoffs feels like a sharp U-turn.

Is the AI hype over?

All said, the sudden change raises a bigger question — is the AI hype beginning to cool off? This is particularly a relevant question given that Meta’s move comes just days after OpenAI launched a muted ChatGPT-5, which didn’t impress users as much as the company’s earlier AI models did.

Over the last two years, AI has dominated headlines, and companies have rushed to call themselves AI-first. But just like previously hyped technologies — blockchain or the metaverse — is AI also going the same way, or is it real?

Take DeepSeek, for example. It grabbed attention almost overnight with a free, powerful model that could rival OpenAI’s ChatGPT and Google’s Gemini. But since then, it has gone quiet, and the excitement has faded.

Just a day ago, a report highlighted that many organisations that tried to make AI part of their workflow have not been very successful. The Study by MIT — titled State of AI in Business 2025 — found that only about 5 per cent of AI pilot projects actually lead to rapid revenue growth, while the vast majority stall or fail to deliver impact. “Almost everywhere we went, enterprises were trying to build their own tool,” said Aditya Challapally, the lead author of the report.

Even OpenAI CEO Sam Altman is now admitting that the industry is in a kind of bubble, although he qualifies it in terms of short run and long run. According to a Verge report, Altman said, “When bubbles happen, smart people get overexcited about a kernel of truth. Are we in a phase where investors as a whole are overexcited about AI? My opinion is yes. Is AI the most important thing to happen in a very long time? My opinion is also yes.”

But is it really an AI slowdown at Meta?

While there are some signs of the AI train slowing down, the events at Meta AI could be something else entirely — that is just a regular, good old restructuring.

It is possible that the shake-up is indeed more about efficiency. With Meta splitting its AI division into four clear teams could reduce duplication of work and make accountability easier. And, if the job cuts do happen, they may be about trimming overlapping roles rather than pulling back on AI altogether.

– Ends