September 25, 2026
TechCrunch reported that Lovable's annualized revenue has crossed $600M, with co-founder Fabian Hedin saying apps built on the platform now get nearly a billion monthly views. That's not a curiosity anymore -- that's a real market. Vibe coding tools let people describe an app in plain language and get working software back, no engineering team required. The number that should catch a business owner's attention isn't the revenue, it's the views: people are actually using what gets built, not just tinkering with demos.
Here's my read: the fight in this category was never going to be won on model quality alone, since most of these tools now sit on similar underlying LLMs. It's won on what happens after the prototype -- can the app handle real users, real data, and real permissions without falling over. That's the gap we've written about before when comparing ViibeStack to Cursor style tools and to v0: a slick UI mockup is not the same thing as a business system with a CRM, billing, and reporting wired together. Lovable's growth proves demand for AI-assisted building is real and mainstream. Whether that demand stays with prototyping tools or migrates toward platforms built for operating a business long-term is the more interesting question, and one I don't think is settled yet.
In a TechCrunch interview, ElevenLabs' CEO -- now reportedly valued at $22 billion -- said businesses should probably disclose when a caller is talking to an AI voice, at least until that becomes the expected default. ElevenLabs' technology already sits behind a large share of AI-powered customer service calls, so this isn't an outside observer's opinion; it's the company profiting from the ambiguity admitting the ambiguity is a problem.
I think this disclosure question is going to become a real compliance issue faster than most companies expect, not a nice-to-have. Voice AI has gotten good enough that customers often can't tell, and 'we didn't think we needed to say' is a weak position once regulators or plaintiffs' lawyers get interested. It rhymes with the debate we flagged around Gemini calling businesses on your behalf -- when an AI is the one placing or answering the call, someone has to be accountable for what it says and whether the other party knew. Businesses running support lines through any helpdesk or voice-enabled workflow should treat disclosure as a design decision now, not a legal patch later. The counterargument -- that disclosure erodes the smooth experience customers actually want -- is fair, but it's a short-term trade against a longer-term trust problem.
Oracle sent a force majeure notice on its New Mexico Stargate data center, TechCrunch reported, a move that lets it delay payments if the facility misses its 2028 target. Force majeure notices are usually reserved for things genuinely outside a company's control, and using one on a flagship AI infrastructure project signals that even well-funded buildouts are hitting real-world friction -- supply chains, permitting, construction timelines, take your pick.
This matters beyond Oracle because so much of the AI narrative assumes compute capacity just shows up on schedule. It doesn't. Every delayed data center is a delayed price drop, a delayed capacity expansion, a delayed everything downstream. For businesses planning around cheaper, more abundant AI compute in the next few years, this is a useful gut check: the frontier model price war is real, but the physical infrastructure behind it is still being built by people pouring concrete and negotiating power contracts, and those things slip.
Which of these trends worries you more as a business owner -- AI tools building your software, or AI voices talking to your customers without saying so?
Sources