Of Dave Lee*

After Deepseek’s hit, Meta became the brightest star of Magnificent Seven. Its stock price has rose 20% so far this year, pushing its estimation to a record level of about $ 1.8 trillion, with $ 500 billion to add only in the last 12 months.

At a time when everyone’s nerves are sensitive about the performance of investment in artificial intelligence, Wall Street buys what Meta’s Managing Director Mark Zuckerberg – The share is negotiating with P/E 27 times the mixed expected profits that It is the highest since September 2020.

There are several reasons for this optimism – but also many reasons to question it.

The rally gives the Meta AI strategy validated with the appearance of Deepseek’s open source artificial intelligence model from China. While the initial panic of the market was that China was far superior to Silicon Valley, what it showed – in terms of Meta – was that the cheapest and most open -air models of artificial intelligence can be the future. If this is the case, as the leading US creator in this, Meta is in a strong position.

Meta then released a positive report on its profits that suggests that its advertising activity was terrible. Part of this was due to the fact that it applied AI tools to help advertisers create ads and target users more effectively. A partnership with Nvidia has led to an 8% increase in “quality” ads that appear to users. Four million customers are now using META’s artificial intelligence tools to create better ads, four times more than six months ago.

Zuckerberg reiterated to investors that his company would be the first to have 1 billion users in its chatbot. Meta’s basic force, its numerous huge platforms, testify that there are 3.35 billion people who may encourage artificial intelligence (despite regulatory regulations in some areas). This will be a “lasting long -term advantage,” Zuckerberg said.

He also promised that the original Facebook application, which has long been neglected on the altar of innovation, would be a little refreshed. Zuckerberg did not accurately clarify his plans, but said that the focus next year would be to make the application have a more “cultural influence”.

The fact that he thinks he can bring Facebook back to the forefront again highlights Zuckerberg’s boldness taken into account in investor estimates for the company’s prospects. They see the gold chain, curly hair and his closer relationship with President Donald Trump and the rightists and believe that all of this can be positive for Meta’s prospects. The decision to abolish events and relaxing the policy of moderation has not harmed advertising sales, the company noted. In return for this right turn, he could benefit from this relaxation of regulatory rules inside and acquire an ally in the “battles” he is called upon to give against the European Union. “We now have an American administration that is proud of our top companies,” Zuckerberg said.

But a company can’t just go beyond “male energy”. The question hovering above Meta is how its artificial intelligence raids will also generate revenue in the future. Unlike its competitors, who build artificial intelligence systems and data centers to sell, META’s AI ambitions are clearly more introverted.

Meta’s open source Llama model, however impressive it is, is given to anyone who wants it, in the hope that it will become smarter faster. Then, the plan is to apply this intelligence to its own products in the form of chatbots and with other improvements that have not yet been determined.

But with what purpose? Meta’s artificial intelligence can well win a billion users, but how this achievement will make money remains in the air. Can be achieved through ads. Or, businesses may start paying to integrate it to approach meta users. But this is not a given – once Meta believed that businesses would come together to integrate into the messenger application, but they didn’t.

The evolution of Meta AI to more than a tool capable of distracting or a trick will be difficult. Its utility is less impressive than the “practical” capabilities of an assistant who can use your calendar and mapping software, such as Google, or incorporated into your business daily work tools, such as Microsoft. Nor will it be in the hands of users, such as Google Assistant or Apple’s Siri (assuming Apple can solve the problems its assistant has).

Even in the Code Scripture, one of the most promising first uses of artificial intelligence, Zuckerberg admitted that his “good mid -level engineer” would not be widely available “soon”.

I am a fan of virtual reality, but the argument that artificial intelligence could improve Meta’s “Metavers” is weak. The adoption of headphones and virtual reality glasses is still slow. Despite new models and advertising explosion, Meta has achieved an annual increase in revenue only 1% in the last quarter. Meta burned $ 17.7 billion hunting Metaverse’s vision in 2024.

Zuckerberg is right to defend the profits from the artificial intelligence that Meta has seen so far. But they are mainly related to improving the existing activity instead of creating an exciting new venture in artificial intelligence. Greater ambitions, and possible odds, you can find them elsewhere.

*Dave Lee is a Bloomberg Opinion technician in the US