Reportedly Meta invests almost $ 15 billion On the Labeling Scale AI scale and get a 49% share of the start, while bringing CEO Alexandr Wang to help drive a new “Superintelligence” workshop to the company.
The deal is reminiscent Obtaining 19 billion dollars of Whatsapp and they are Purchase of $ 1 billion Instagram. At a time when the mergers were closed, many people suggested that Meta had been overly paid over the platforms – and today’s reason is no different. There is no shortage of investors and founders who were left to scratch their heads this weekend over Meta’s last connection.
In the end, WhatsApp and Instagram became an integral part of the CEO Mark Zuckerberg empire. The question is whether the AI scale agreement will work similarly in favor of Meta, proving once again Zuckerberg’s prophetic strategy or if the company catches the straws in an incorrect attempt to cover up opponents such as Openai, Google and Anthropic.
In this case, Meta does not bet on an emerging application of social media, but to the data used to educate top AI models. For the last few years, which leads AI laboratories, such as Openai, they have been based on the AI scale for the production and labeling of data used for model training. In recent months, Scale AI and its data commentary competitors have begun to hire people with high specialization, such as doctoral and higher engineering scientists, to create high quality data for Frontier AI Labs.
It can benefit Meta to have a close relationship with a data provider such as scale. Meta leaders have complained about the lack of innovation around the data in the company’s top AI groups, according to a person familiar with the issue.
Earlier this year, Meta’s AI genetic unit launched Llama 4, a family of AI models that failed to match the capabilities of the Ai Lab Deepseek Chinese models, and this was widely regarded as frustration. It does not help the issues, Meta is trying to combat a friction problem. According to the data compiled by Signalfire, Meta lost 4.3% of its top talent in AI laboratories in 2024.
Meta is not only betting on the AI scale to rejuvenate AI’s efforts, but also to Wang to drive the aforementioned new Superintelligence team. The 28-year-old CEO has proven to be a powerful starting founder-it is well known around Silicon Valley to be ambitious, a good seller and very well connected. In recent months, Wang was Meeting world leaders to discuss AI’s impact on society.
However, Wang did not drive a AI laboratory of its kind before, however, and does not have the same AI research background that many other AI Lab leaders, such as the Safe Superintelligence or Arthur Mensch of Mistral’s Arthur Mensch. This is perhaps because Meta is also said to hire talent high profile Like Deepmind’s Jack Rae complete AI’s new research team.
Fate after acquiring the AI scale the company is a bit dark. The role of real world data in AI models are changing-designated AI laboratories have brought home data collection efforts, while others have increased their dependence on synthetic data (ie AI produced). In April, information reported that AI scale had lost some of its revenue goals.
According to co -founder Anyscale Robert Nishihara, many AI Frontier laboratories are exploring new ways to exploit and optimize data, many of which are quite computational.
“The data is a moving goal,” Nishihara told TechCrunch in an interview. “It’s not just a finite attempt to cover – you have to innovate.”
It is likely that the Meta and Wang relationship could scare other AI laboratories that have traditionally worked with the AI scale. If so, this agreement could be a benefit for AI scale competitors, such as Turing, Surge AI, and even non -conventional data providers, as recently launched LM Arena.
Turing CEO Jonathan Siddhaarth told TechCrunch by email that it has received increased interest from customers in the light of rumors of Meta’s deal with Scale AI.
“I think there will be some customers who prefer to work with a partner who is more neutral,” he said.
Only time will say how Meta’s investment will overcome its efforts, but the company has an important reason to offset. In the meantime, competition does not slow down. Openai promotes the release of the next flagship model, GPT-5, as well as its first openly available model from a chronic model that will compete with Meta’s current and future releases.
