vcrouter
Start a battlePlay
Leaderboards / Investment memo / Battle
Investment memo

Lovable — prompt-to-app builder, one of Europe's fastest revenue ramps ever

September 13, 2026 · $0.13 for both answers and the judge

Claude Fable 5 wins

On September 13, 2026, Claude Fable 5 and GPT-5.6 Luna Pro each wrote an investment memo for the same brief, without seeing the other's answer. The judge, Grok 4.6, read both twice, once in each order, and picked Claude Fable 5 both times.

Why, in the judge's words“Sharper thesis/verdict, staged public facts vs unknowns, real deal-killers (churn/NRR), and tighter named-competitor reasoning.”

The two answers

Claude Fable 5Winner
$0.10 · 35s · 2,023 tokens

Investment Memo: Lovable

Company: Lovable (Stockholm, Sweden) Category: AI prompt-to-app builder ("vibe coding") Stage considered: Hypothetical Series A

---

Thesis

Lovable is attempting to do to software creation what Canva did to design: collapse the gap between "I have an idea" and "I have a working product" for the hundreds of millions of people who cannot code. If AI-generated software becomes the default way non-technical founders, operators, and SMBs ship apps, the winner of this category captures a market far larger than the developer-tools TAM — it's a slice of the ~$500B+ spent annually on custom software development and the latent demand priced out of it entirely. Lovable's publicly reported revenue ramp — reportedly ~$17M ARR within ~3 months of launch and past $50M ARR within roughly six months, among the fastest in European startup history — suggests genuine product-market pull, not paid-acquisition vapor. A fund-returner case exists: prosumer land, team/enterprise expand, become the default application layer on top of frontier models. That's a $10B+ outcome if it works.

Product & wedge

Lovable generates full-stack web applications from natural-language prompts: front-end (React/Tailwind), backend and auth via Supabase integration, deployment, and GitHub sync so real developers can eject to code. The wedge is speed-to-first-artifact: a non-engineer gets a live, shareable app in minutes. Critically, Lovable chose the non-developer as its primary user — unlike Cursor (professional engineers) or GitHub Copilot (inside the IDE). This positioning matters: it avoids competing on developer trust and instead competes on accessibility and delight, a Canva-style motion with viral, self-serve distribution. The GitHub eject-hatch reduces lock-in fear while keeping most users inside the product.

Market & competition

The space is brutally crowded and moving fast:

  • Vercel v0 — strongest distribution among developers; backed by Vercel's deployment ecosystem.
  • Bolt.new (StackBlitz) — near-identical positioning; also reported an explosive ARR ramp ($20M+ ARR within months).
  • Replit (Agent) — full IDE + hosting + agents; well capitalized, broader platform.
  • Cursor / Windsurf — adjacent (pro developers) but could move downmarket.
  • Figma Make, Canva Code — incumbents with enormous non-technical distribution entering directly.
  • Frontier labs themselves — Anthropic's Claude Artifacts and OpenAI's Canvas offer "good enough" app generation bundled into subscriptions users already pay for.

There is no obvious durable technical moat: every player rents the same frontier models. Differentiation must come from workflow depth (backend, auth, payments, deploy), community/templates, and brand as "the tool for non-coders."

Traction & business signal (publicly known)

  • Reported ~$17M ARR ~3 months post-launch (early 2025); publicly stated $50M+ ARR by mid-2025 and reports of continued rapid growth beyond that. Among the fastest revenue ramps of any European software company.
  • Raised a ~$15M pre-Series A led by Creandum (Feb 2025); subsequently reported ~$200M Series A led by Accel at a ~$1.8B valuation (July 2025), making it a claimed fastest-to-unicorn in Europe.
  • Team reportedly extremely lean (tens of employees) relative to revenue — exceptional revenue per head.
  • Unknown: net revenue retention, monthly churn, cohort behavior, gross margin after inference costs, mix of hobbyist vs. business users, and what fraction of "apps built" ever see real usage. These are the numbers that determine whether this is a business or a fad-driven top of funnel.

Risks — the three that actually kill the deal

  1. Churn / novelty revenue. The single biggest risk. Much of the ARR may be curious prosumers building toy apps who churn in 2–4 months. Bolt and Lovable both showed identical hockey sticks, which suggests category hype rather than company-specific pull. If NRR is sub-70%, the ARR headline is a mirage and the growth is a leaky bucket funded by TikTok virality.
  2. No moat against the model layer and incumbents. Lovable is a thin(ish) orchestration layer on Anthropic/OpenAI models. If Claude or ChatGPT ship "deployable apps" natively — they are visibly moving this way — or Figma/Canva bundle it free into products with 100M+ users, Lovable's standalone subscription gets squeezed. The eject-to-GitHub feature, a selling point, is also an anti-lock-in feature.
  3. The complexity ceiling. Generated apps degrade sharply beyond prototype scale: state management, security, migrations, edge cases. If users hit a wall at "demo-quality," Lovable becomes a prototyping toy with low willingness-to-pay, not a production platform — capping ACV and enterprise expansion. Security incidents from AI-generated apps (exposed data, bad auth) are an accelerant on this risk.

Verdict

Invest at Series A — with conditions on price and one diligence gate. The category is real, the team has demonstrably out-executed a crowded field, and revenue ramps this steep are rare enough to pay attention even discounting hype. But at a reported ~$1.8B on ~$50–75M of months-old ARR, you're underwriting near-flawless retention.

The one question I need answered first: Show me month-by-month cohort revenue retention. Specifically: of customers who started paying in month X, what percentage of that revenue remains at month 3 and month 6? If cohorts retain 80%+ of revenue at month 6, this is a genuine platform and I invest despite the moat risk. If retention is sub-50%, the growth is churn-masked virality — pass at any price.

GPT-5.6 Luna Pro
$0.007 · 24s · 4,619 tokens

Thesis

Lovable could be a fund-returner because it sits at the intersection of three unusually powerful trends: foundation models making software creation cheaper, millions of non-developers wanting bespoke tools, and software developers moving from “writing code” toward supervising agents. Its promise is not merely faster coding; it is converting a natural-language idea into a deployed, usable application with a database, authentication, payments and integrations.

The upside case is a new software-creation layer: a product used by entrepreneurs, small businesses, agencies and internal teams before they ever hire a developer. If Lovable becomes the default place where applications are specified, generated, iterated and hosted, it could capture subscription revenue, usage-based infrastructure revenue and potentially payments or marketplace economics. A rapid revenue ramp would also indicate unusually strong product-market fit rather than a conventional SaaS sales motion.

The caveat is that this category is strategically exposed to the model providers and developer platforms. A fund-returning outcome requires Lovable to own the workflow and customer relationship—not simply provide an attractive interface over interchangeable models.

Product & wedge

Lovable, formerly associated with the GPT Engineer project, is a prompt-to-app builder. Users describe an application conversationally, and the system generates front-end and back-end code, connects databases and allows iterative changes through additional prompts. The product is designed to produce real web applications rather than static mockups, with integrations including services such as Supabase and GitHub.

Its wedge is accessibility. Traditional development requires technical fluency; no-code tools such as Bubble require learning a visual system; coding copilots such as Cursor assume a developer is already present. Lovable’s pitch is that a founder, marketer, consultant or small-business owner can move directly from a product idea to a functioning application.

That wedge is especially compelling for prototypes, internal tools, client projects and lightweight production applications. The key product question is whether Lovable can maintain quality as applications become complex. Generation is easy to demonstrate; reliable debugging, version control, security, permissions, performance and long-term maintainability are much harder.

Market & competition

The market is broad but crowded. Direct competitors include Bolt.new from StackBlitz, Replit Agent, v0 from Vercel, and Firebase Studio from Google. These products are converging on prompt-driven application generation, with different strengths in browser-based development, deployment, UI generation and cloud infrastructure.

Lovable also competes indirectly with Cursor, Windsurf, GitHub Copilot and other coding agents for developer workflows; with Bubble, Webflow and Framer for no-code and website creation; and with Retool and similar tools for internal applications. Open-source frameworks such as GPT Engineer and rapidly improving model-native coding tools constrain differentiation.

The market can support a large company if “software creation” expands substantially beyond professional developers. However, the TAM should not be justified by adding every developer tool and no-code category together. The investable market depends on how many users repeatedly build and operate applications, not how many people try a prompt once.

Traction & business signal

Public reporting has described Lovable as one of Europe’s fastest-growing software companies. In 2025, the company publicly reported—or was reported as having—reached approximately $100 million in annual recurring revenue within roughly eight months of launch, alongside a $200 million financing at an approximately $1.8 billion valuation. These figures are extraordinary and, if accurate, are among the strongest early revenue ramps in European startup history.

The quality of that revenue is less clear publicly. Unknown: audited revenue, gross retention, net retention, paid conversion, average revenue per account, cohort behavior, infrastructure gross margins, customer concentration and the share of revenue from individual versus business users. Unknown: how much usage is production-critical versus experimentation, and how much code generated by Lovable remains actively maintained in the product.

The financing and reported revenue trajectory are strong business signals: clear demand, unusually efficient distribution and the possibility that the product is becoming a habit rather than a novelty. They are not yet proof of durable defensibility. AI application revenue can grow rapidly while churn, support costs and model-inference costs remain hidden.

Risks

1. The platform-layer risk. OpenAI, Anthropic, Google, Microsoft, Vercel and other infrastructure providers can absorb prompt-to-app functionality into products users already pay for. If Lovable’s differentiation is primarily model access plus a good interface, pricing power and retention may collapse.

2. Reliability and liability. A generated prototype is not the same as a secure production system. Data leaks, broken authentication, vulnerable dependencies, hallucinated business logic or failed deployments could destroy trust, particularly among companies using Lovable for customer-facing applications. Solving this may require expensive engineering, review and support, reducing the apparent software margins.

3. Revenue-quality and distribution risk. The reported ramp may include enthusiastic early adopters and high-intent launch traffic. If users build one application and leave, or if serious teams export code to another platform, the company may have impressive bookings but weak durable retention. A crowded market will also make paid acquisition and enterprise expansion harder over time.

Verdict

Invest at a hypothetical Series A, but only with disciplined pricing and milestone protection. The speed of adoption is sufficiently exceptional to justify underwriting an outlier, and the product addresses a genuinely large behavioral shift. I would not invest on headline ARR alone.

The one question I would need answered first is: what are the gross dollar retention and net revenue retention cohorts for customers who have been paying for at least six months, segmented between individual users and businesses—and how much of their generated software remains actively deployed on Lovable? If those metrics show durable production usage and expanding accounts, Lovable may own a new software-creation category. If not, the company may be monetizing a very impressive demo cycle.

Claude Fable 5 vs GPT-5.6 Luna Pro, overall

Across 30 judged battles, Claude Fable 5 won 21, GPT-5.6 Luna Pro won 1, and 8 were ties. On investment memo alone: 6 to 0, 4 tied.

Compare Claude Fable 5 and GPT-5.6 Luna Pro Judge one yourself, blind

Costs are the provider's list prices at the time of the run.