Meta hired Alexandr Wang to build AI. It’s Zuckerberg’s job to sell it

A year after spending more than $14 billion to hire Alexandr Wang and a group of top Scale AI engineers to revamp its AI efforts, Meta is back on the map, at least in AI, but there’s still OpenAI, Anthropic and Google in the market.

Wang’s big hit was the delivery of the Muse Spark AI model in April; This marks Meta’s first foray into proprietary core models and a move away from strict adherence to open source, or open-heavy, more commonly referred to as AI. The Meta Superintelligence Labs group led by Wang was formed to give the company some momentum in the hottest corner of the tech industry.

Now that CEO Mark Zuckerberg has his new model, it’s up to him to turn it into financial success. This means showing that the company can attract paying users for its AI tools, rather than using the technology only to enhance and support its core advertising business.

“Meta needs to provide more evidence of both adoption and commercialization,” said William Blair analyst Ralph Schackart, who recommends the stock purchase. “Beyond the significant positive impact AI has had in improving advertising models, investors are looking to Meta to monetize a new AI-first product.”

Wall Street isn’t particularly impressed, at least so far. Meta’s shares lost 18% in the last 12 months, the worst performance of the megacap group. MicrosoftSomething that has its own challenges in AI. This is true even after Meta reported 33% revenue growth in the first quarter, the fastest growth rate of any period since 2021.

According to Meta, the problem started with what some industry experts call, at least in hindsight, a strategic mistake. The company has moved into AI with its Llama model family, offering an open-source approach that allows developers to tinker freely while other major model makers charge for access.

In April last year, Meta’s Llama 4 release flopped, failing to attract developers and prompting Zuckerberg to rethink his company’s approach to AI development. Two months later, Zuckerberg shocked the tech world by announcing that his company had invested $14.3 billion for about half of Scale AI and, more importantly, hired Wang and his top brass.

Wang’s development and launch of Muse Spark in April this year has made great progress. IThomas Randall, an analyst at Info-Tech Research Group, said that instead of focusing on third-party developers, the new model is designed to easily connect to Meta’s applications such as Facebook and Instagram, as well as AI-powered devices such as Ray-Ban Meta glasses. This is on top of the standalone Meta AI app and site.

“There will be a lot of these leading model providers that will fundamentally change in many different ways, and Meta needs a consistent, reliable, proprietary model that it has itself,” Randall said. He added that Meta would be “lost” if Zuckerberg didn’t open his wallet for Wang and other famous AI insiders last year, in what Randall called a “strategic restructuring” for the company.

Randall said Meta didn’t take the “most optimized route,” but at least “I can now see a vision of what they were trying to accomplish and what Wang was trying to accomplish.”

Since the launch of Muse Spark, Meta has announced new AI and business-related subscription plans as part of an effort to expand its business beyond online ads. Historically it hasn’t worked. Meta still gets 98% of its revenue from advertising.

Schackart said he wants to see “concrete evidence of the growing list of new, AI-first products created by Muse Spark, even if monetization is delayed.” “That’s what investors are looking for,” he said.

No matter how good Wang’s model is, Zuckerberg has a tall hill to climb as developers recover from the Llama debacle.

“I think the AI ​​community has largely ignored Meta at this point,” said Rob May, CEO of Neurometric, a token engineering startup.

May said it’s difficult to gauge how much success Wang has had at MSL because the company has only released one AI model so far, which Wang called a “yawn” among the AI ​​community because the technology isn’t widely available.

Although Meta has relied heavily on third-party developers with Llama, May said the company’s efforts under Wang appear to be geared toward internal uses. May said she had been in regular contact with Meta on Lama-related matters but was now “unable to get them to respond to messages”.

May agrees that Meta’s focus on AI for its core ad products makes sense because the company’s $200 billion-a-year business needs to be protected.

“That company made the machine,” he said.

Andrew Moore, CEO of enterprise startup Lovelace and former Google Cloud AI chief, said it’s not too late for Meta to find a way forward.

Meta focused on making its models more efficient with training techniques. This could make a significant difference among developers worried about the rising costs of base models, Moore said.

“If they make proprietary, computationally efficient models, it will be very different from what happened in this death match between the big guys,” Moore said. “They could really benefit from it.”

Moore added that Meta must show an advantage somewhere in terms of cost, latency, or other technical nuances important to developers.

Krish Subramanian, CEO and former head of product at consulting firm KOI AI IBM’s Developers are more excited about Google’s AI models than those offered by Meta, Consulting said. Llama’s appeal was that it was specifically aimed at developers who wanted exploit-heavy alternative models; He said that Meta has made little effort in this direction with Muse Spark.

In addition to difficulties with developers, there is also a drop in morale. It became a commodity There were layoffs throughout the year, with nearly 8,000 workers laid off in May. The outages have spread across departments, including teams working in trust and safety-related roles, raising concerns about potential problems in AI development, according to people familiar with the matter who spoke on condition of anonymity to speak candidly about the matter.

Meta declined to comment on the layoffs. Regarding security-related issues, the spokesperson noted Wang’s comments on the matter. he said Core Memory podcast “One of the things that is very important to me is the safety of these models,” he said last month.

There is also tension at the top of the AI ​​organization. While the Muse Spark version has received high marks internally, there is pressure from Wang and former GitHub CEO Nat Friedman, who joined last summer as part of an AI spending spree, to deliver meaningful revenue growth from the model and future releases, people with knowledge of the matter said.

Meta technology chief Andrew Bosworth, a 20-year company veteran, is a close confidant of Zuckerberg and someone the CEO can turn to for a larger role on AI if the newcomers are perceived as unsuccessful, sources said. In a May podcast, Wang denied reported infighting.

Wang called the Muse Spark an “appetizer” for the future and said there would be more powerful, “larger models.”

But the AI ​​community is used to a constant stream of updates and new features. This is what they got from OpenAI, Anthropic and Google.

“What I care about is the frequency of launches and the tempo,” said Howard Yu, a business professor at the International Management Development Institute in Switzerland. “When you start something, can you build on that momentum?”

Randall of Info-Tech Research Group said it’s ultimately up to Zuckerberg to figure out that strategy and show “what a superpower they are right now with all their products.”

Yu agreed.

“This is really about leadership, right?” he said, noting that, especially in technology companies, the CEO defines and expresses the vision, especially when it comes to spending billions of dollars.

Yu said that Zuckerberg’s metadata and virtual reality targets have caused a total loss of more than $80 billion since late 2020, making the sale of artificial intelligence more challenging.

“It’s running out of space for its continued reliability,” Yu said. “I think the VR foray may have burned through a lot of its goodwill with investors.”

WRISTWATCH: Meta ‘is tone deaf’