Latest / MBA Training AI / Lovable's $600M ARR: what AI-built apps cost at month six

Transcript

Lovable's $600M ARR: what AI-built apps cost at month six · MBA Training AI

• 0:00
  Host: Welcome back to Leaders Insights.
• 0:07
  Host: $600 million in annualized revenue on code nobody's engineering team wrote, and why it
• 0:12
  Host: matters this week.
• 0:13
  Host: Most of the code running, lovable's $600 million business is garbage.
• 0:19
  Host: And that's exactly why it works.
• 0:20
  Expert: You're half right, which is the most dangerous place to be.
• 0:24
  Expert: The code is often mediocre.
• 0:26
  Expert: That mediocre code that ships beats beautiful code that sits in a branch nobody merged.
• 0:31
  Host: Let's take that $600 million apart, because it's an annualized figure and those are slippery.
• 0:36
  Host: What's actually inside it?
• 0:39
  Expert: Annualized means they took a good month and multiplied by 12.
• 0:42
  Expert: So treat it as a snapshot, not a bank statement.
• 0:45
  Expert: The real tell is the near-billion monthly views on apps built with the platform.
• 0:51
  Expert: Views mean the output isn't dying in a demo.
• 0:53
  Expert: Someone's using these things.
• 0:55
  Expert: Or someone's spun up a thousand throwaway apps and the view count is vanity.
• 0:59
  Expert: That's the crack I'd pry open first.
• 1:01
  Expert: A billion views across how many apps?
• 1:04
  Expert: If it's 10,000 apps averaging 100,000 views, that's a healthy ecosystem.
• 1:09
  Expert: If it's 50 apps carrying the whole number, you've got a lottery, not a business.
• 1:13
  Expert: Lovable hasn't broken that down publicly.
• 1:15
  Expert: And until they do, I'd hold the champagne.
• 1:18
  Host: So the metric that everyone's quoting is the one you trust least.
• 1:22
  Expert: The revenue's fine.
• 1:23
  Expert: That's the story people wrap around it that's junk.
• 1:26
  Host: Everyone's saying, look, AI writes software now, engineers are obsolete.
• 1:31
  Host: That's the consensus.
• 1:33
  Host: And it's lazy.
• 1:34
  Host: Make the case that it's wrong.
• 1:36
  Host: Because $600 million is a loud argument.
• 1:39
  Expert: Here's the distinction nobody's making.
• 1:41
  Expert: Lovable isn't replacing engineering.
• 1:43
  Expert: It's replacing the blank page.
• 1:45
  Expert: The hard part of software was never typing the code.
• 1:49
  Expert: It was knowing what to build and getting version one in front of a human.
• 1:53
  Expert: That's the part the model does well.
• 1:54
  Expert: The part it does badly is the part that shows up at month six.
• 1:58
  Host: Define month six.
• 1:59
  Expert: The moment your app has real users, real data, and something breaks at 2 AM, the model that
• 2:05
  Expert: generated your login flow doesn't know why sessions are dropping under load.
• 2:10
  Expert: Someone who actually understands the stack, the layers of software from the database
• 2:14
  Expert: up, has to walk in and read code they didn't write.
• 2:19
  Expert: It's a worse job than writing it fresh.
• 2:21
  Expert: And it's the bill coming due on all of this.
• 2:23
  Host: You're describing technical debt with a smile on its face.
• 2:27
  Expert: Debt.
• 2:28
  Expert: The shortcuts you take now that you pay for later with interest.
• 2:32
  Expert: And AI-generated apps take on debt faster than any junior developer I've ever managed.
• 2:37
  Expert: Because the machine has no shame and never gets tired of making the same mistake.
• 2:42
  Host: OpenAI and Google both publish numbers on how much code their models now generate.
• 2:47
  Host: How much of that do you buy?
• 2:48
  Expert: OpenAI has said a large share of code in some companies is now model-assisted.
• 2:54
  Expert: And Google's made similar claims about its own engineers.
• 2:57
  Expert: But both sell the tools doing the generating, so those figures are marketing dressed as
• 3:02
  Expert: research.
• 3:03
  Expert: There's no independent audit I'd stand on today.
• 3:05
  Expert: Cross-check before you quote them.
• 3:07
  Host: Then what's the honest read on lovable success?
• 3:10
  Expert: They found the customer everyone ignored.
• 3:13
  Expert: Not the engineering team, the founder, the marketer, the person with an idea and no
• 3:17
  Expert: CS degree.
• 3:19
  Expert: That buyer never had access to software creation before.
• 3:23
  Expert: Lovable didn't beat the engineers.
• 3:25
  Expert: It went around them to a market that was starving.
• 3:28
  Host: And when those apps hit month six?
• 3:30
  Expert: Some die.
• 3:31
  Expert: Some hire a real developer to clean up the mess, which, funny enough, means the tool
• 3:36
  Expert: that was supposed to kill engineering jobs is generating them, just downstream and grumpier.
• 3:42
  Host: So the contrarian take is that lovable proves engineers matter more, not less.
• 3:47
  Expert: It proves the value moved.
• 3:49
  Expert: Away from writing the first version toward maintaining and scaling the thing once it's
• 3:54
  Expert: alive.
• 3:55
  Expert: If you're a developer panicking about your job, that's where you run.
• 3:59
  Host: Give me the one thing a listener does with this tomorrow.
• 4:02
  Expert: Build your prototype with lovable or anything like it.
• 4:05
  Expert: Get it in front of users this week.
• 4:07
  Expert: But before you write a line of real business logic, ask one question.
• 4:12
  Expert: Who reads this code in six months when it breaks?
• 4:15
  Expert: If the answer is nobody, you don't have a product.
• 4:17
  Expert: You have a countdown.
• 4:19
  Host: Sources for today's episode.
• 4:21
  Host: TechCrunch AI, the decoder.
• 4:23
  Host: Ours Technica AI.
• 4:25
  Host: KD Nuggets, Google AI, vendor, AI Lab.
• 4:28
  Host: Open AI, vendor, AI Lab.
• 4:31
  Host: That's your five.
• 4:32
  Host: The full AI track, structured and free of fluff, is at MBA-training.com.