Meta has launched its first standalone artificial intelligence app in a direct challenge to OpenAI’s ChatGPT, just one day before releasing its first earnings report of the year.
The new app, announced at Meta’s developer conference in Menlo Park, California, is built on the company’s Llama models and includes a text-based AI assistant, image generation tools, and a curated ‘discover’ feed that highlights user prompts. This mirrors functionality seen in other AI offerings such as Google’s Gemini and Elon Musk’s xAI-powered Grok.
You can view Meta’s official announcement here.
The launch underscores Meta’s ambitions in the highly competitive AI landscape, as it races to catch up with rivals who have moved quickly to integrate generative AI into products and services. Despite marketing its Llama models as open-source and widely accessible to developers, the latest version, Llama 4, has faced criticism over inconsistent performance and transparency issues.
In particular, developers have raised concerns about benchmark scores used to promote the model. “They should have been more explicit that the Maverick model was not the same one released,” said Ion Stoica, co-founder of AI research hub LM Arena, referencing dissatisfaction over internal test versions that differed from the final release.
Still, Meta remains bullish. Its leadership has declared 2025 “the year a highly intelligent and personalised AI assistant reaches a billion people”, but early signs suggest developer enthusiasm may be waning. Meta’s open-source approach, once seen as a key competitive advantage, now faces scrutiny in a market growing more cautious about generative AI’s real-world utility.
Investor attention is also closely focused on the financial implications of Meta’s AI investments. Ahead of Wednesday’s earnings release, analysts are projecting earnings per share (EPS) of $5.21 (£3.89) on revenues of $41.2bn (£30.75bn), representing a 13 per cent year-on-year increase. However, in recent weeks, consensus estimates have been revised downward, reflecting increased caution about the near-term benefit of Meta’s AI push.
Capital expenditure has ballooned as Meta ramps up data centre capacity and computational infrastructure to support its generative AI ambitions. Free cash flow is forecast to drop 31 per cent to $8.6bn (£6.42bn), due in large part to the projected $60–65bn (£45–49bn) in AI-related infrastructure spending this year.
External pressures are also mounting. The pending removal of the US ‘de minimis’ tariff exemption on 2 May – which currently allows imports under $800 (£597) to enter duty-free – will affect large Chinese advertisers such as Shein and Temu. This raises the risk of reduced digital advertising budgets later in the year, which could dent Meta’s revenue.
Analysts from CFRA remain cautious. “Second-half visibility is murky, and the tariff shift may only amplify uncertainty,” said CFRA’s Angelo Zino, adding that Meta’s financial guidance is likely to cover a wider-than-usual range of outcomes.
Despite this, some believe Meta’s sheer scale makes it better placed than rivals to absorb the shocks. “Meta is more resilient than peers,” said Ted Mortonson from Baird, noting the platform’s 3.4 billion monthly users across its apps – which include Facebook, Instagram, and WhatsApp – offer a massive base for AI product integration.
Investors will be watching closely to see whether Meta has made meaningful progress in monetising its AI investments. “Any concrete progress on AI monetisation could lift the multiple,” said Malik Ahmed Khan, equity analyst at Morningstar.
Wall Street remains optimistic. According to FactSet, 88 per cent of analysts rate Meta a ‘buy’, with an average price target of $707 per share. That optimism, however, hinges increasingly on Meta’s ability to turn its bold AI ambitions into tangible financial returns in the face of rising competition and operational costs.
You can follow Meta’s earnings release live on its investor relations site.
