Thursday’s announcement of a new version of Meta’s AI model, to be offered at a cost to developers (competing with similar Claude and GPT offerings), was widely reported, and rightly so – it’s a high-growth area for this massive member of the Mag 7. Meta, by all accounts, “overinvested” in AI capacity. It looks like the investment might pay off much sooner than expected.
Plus, Meta just announced a huge AI campus in Canada, replete with ready sources of cheap power.
Initial news reports focused on it as a developer offering, but Meta’s continued heavy investment in this race spills over to its consumer-facing products as well. Similar to competing LLMs, consumers who want extra brainpower may increasingly use “thinking” mode for deep research. Avid tool fans might also begin using Spark’s image creation tool or even create agentic workflows.
There will continue to be plenty of AI chat potential within Facebook and Meta. And coming hard: a huge number of WhatsApp users now ask Meta AI questions directly, rather than popping out of a chat to Google something.
Indirectly, this spills over into better usage numbers for Meta, and more ad revenues.
The picture of how, exactly, Meta will grow ad revenues as its AI usage takes hold among consumers, business partners, and developers, is a bit fuzzy and requires some imagination. In the WhatsApp case, the platform is largely monetized through various B2B revenue streams, charging a toll on businesses for access to users, taking a cut of payments, etc.
As consumers increasingly query Meta AI within this platform, I’d expect no direct attempt to monetize via ads. In other words, the same cash-bleeding “delay monetization via ads strategy” that all major LLM players are pursuing, in a vicious price war intended to put a dent in OpenAI’s share.
Now, given that users will be enjoying this free and integrated LLM usage in an idiosyncratic way inside of Meta’s main properties, it’s time to think outside of the box a bit here. Why would Meta not at some point just release a competing standalone chatbot and app? This has more potential than, say, Threads (though some colleagues may disagree). If Musk fanboys are delightedly @grok-ing their way through the day, is it possible that some branding sleight-of-hand might be lurking here to allow Meta to go more head-to-head with ChatGPT and Google?
Stranger things have happened. And the infrastructure and capacity will be idling in the garage, ready to roar.
With specific regard to ad revenue and the value to advertisers within Facebook and Instagram, Meta’s AI has already led to wins for the advertiser community seeking just the right fit with high-intent and mid-funnel consumers. Constantly improving AI applied to consumer usage patterns, as they intersect with advertiser offerings, will continue to drive more revenue for the latter. That helps Meta’s top line as those ad budgets grow.
Looks like investments in AI are the “spark” Meta needed to stay abreast of their competition. Things are heating up.
Meta’s one-year stock performance: down 8%, against Google’s 97% gain. 🤔