Lovable vs Bolt.new Gap Widens as AI Coding IDEs Consolidate

Two builders, one market, very different trajectories

Throughout 2026, "Lovable or Bolt.new?" has been one of the most repeated questions in AI software tooling, and most of the answers agree on how to frame it. Reviews call Lovable a chat-first product builder for non-technical founders. They call Bolt.new a code-first cloud IDE for developers who want a file tree, a terminal and a choice of models.212628 That framing is accurate, but it now misses the bigger story. The real divide is less about interface design and more about scale, capital and who owns the tools that professional developers sit inside all day.

The broader AI coding market also changed shape this year. Cursor, the biggest AI-native editor, became a subsidiary of SpaceX. Windsurf, its closest rival, was taken into Cognition and given a new name.5540 Read together, these events suggest that prompt-to-app builders and AI code editors are both consolidating, but in different ways.

Lovable's numbers have outrun the comparison

The financial gap between the two app builders is now so wide that side-by-side feature tables look a little quaint. Lovable raised $400 million at a $13.3 billion valuation in August, led by Menlo Ventures alongside the EQT-managed Scaleup Europe Fund. That more than doubled the $6.6 billion valuation it reached in December.14 Its revenue path is just as steep. The company reported $200 million in ARR at the end of 2025 and $400 million by March 2026, then said it passed a $500 million annualized run rate in June and crossed $600 million in September.1415 Lovable also says two-thirds of the Fortune 500 now use the product.15

Bolt.new's public figures look very different. The $40 million ARR number that 2026 comparison pieces keep repeating actually comes from early 2025, which is when Sacra estimated StackBlitz hit that mark.49 The last publicly reported valuation is still about $700 million, from a funding round in January 2025.4349 StackBlitz has not disclosed a 2026 valuation.45

This is the first place where the coverage splits in a meaningful way. Several 2026 comparisons set Lovable's current figures beside Bolt's 2025 figures as if they were measured at the same time.2128 Our reading is that the scale gap is real, but its exact size is unknown because StackBlitz has stopped publishing comparable numbers. Many of the Lovable figures also come from the company itself rather than audited accounts.1115

Where reviewers agree, and where they contradict each other

On the product, reviewers mostly agree on the core trade-offs. Lovable sticks to an opinionated React, TypeScript and Tailwind stack, sets up Supabase for the database and auth, offers two-way GitHub sync, and is widely praised for polished UI.212328 Bolt.new runs Node.js inside the browser using StackBlitz's WebContainers. It supports React, Vue, Svelte, Next.js, Astro and other frameworks, gives users a terminal, and has an open-source version called bolt.diy.3044

Benchmarks of speed mostly point the same way. One hands-on test found Bolt reached a first deployment in about 12–15 minutes, compared with 18–25 for Lovable.29 Another comparison credits Bolt's diff-based edits with making each iteration faster.26

The disagreements are sharper on the backend. A February comparison said Bolt had no database and required manual setup for both auth and deployment.22 Coverage from March onward says Bolt Cloud, which launched with Bolt V2 in September 2025, has largely closed that gap with built-in databases, authentication, storage, server functions and hosting.2144 One October review, checked against StackBlitz's own documentation, confirms that Bolt hosting is now the default path for new projects.41 An app-builder vendor took a contrarian view, arguing that neither tool really includes a database. That claim looks out of date compared with Bolt's current documentation.2741 We conclude that backend support has largely converged, and Lovable's remaining edge is that its Supabase integration is more mature and more automatic.2128

Pricing is also reported inconsistently. Bolt's entry tier is listed variously at $20, $25 or $15 a month.222541 The most recently verified figure is $25 for 10 million tokens, with a free tier of 1 million tokens.41 Lovable's Pro plan is consistently reported at $25 a month for 100 credits plus five per day.2845 So the real difference is the billing model, not the sticker price. Reviewers repeatedly say Bolt's token use is hard to predict as projects grow.4449

Bolt's enterprise pivot

Bolt has answered the scale gap by moving upmarket instead of trying to win on consumer growth. In May, StackBlitz announced a partnership with Microsoft that covers Azure procurement, Microsoft 365 integration and deployment into a customer's own tenant.43 Its Enterprise tier can deploy generated infrastructure into an organization's own AWS or Azure account.45 In September, Bolt made its first acquisition, buying Dokai, which builds autonomous agents for enterprise go-to-market workflows.47 The company also retired its v1 agent in August and has tested an agent built on open-source models called Bolt Forge, along with a waitlist for a $9 Bolt Lite tier.42

User sentiment is still a weak spot. Bolt's Trustpilot rating has hovered around 1.5 out of 5, and most complaints are about token use and unreliable output.49 Lovable has had its own problems. In April it disclosed a bug that left some public projects' chat history and source code visible to other users.45

The editors: Cursor joins SpaceX, Windsurf becomes Devin

Developers who outgrow app builders usually move to AI-native editors, and that part of the market has changed even more. SpaceX agreed in June to buy Anysphere, Cursor's parent company, for $60 billion in stock.51 The deal closed on August 14.55 Coverage put Cursor's annualized revenue at roughly $4 billion and the price at about 15 times revenue. One analysis also noted that, according to Ramp spending data, Cursor's share of corporate spend fell from around 41% to about 26% between June 2025 and May 2026.52 That combination of strong revenue and falling share is worth keeping in mind. It suggests Cursor sold from a position of strength in revenue but facing more competition.

Cursor's product roadmap has been moving toward agents. Cursor 3 arrived in April with an Agents Window for running parallel tasks locally, in the cloud or over SSH, and Cursor 3.1 added a tiled layout.9 Earlier releases added subagents and Skills, and the Composer 2.5 model became the default for new CLI sessions in May.18 Cloud agents now run in sandboxed VMs. One October report says they gained interactive Canvases and an Environments API in beta, and that an Agent SDK has been announced, though these later details are less widely confirmed.1 Commentators are divided on what the acquisition means for users. One analysis warns that third-party models could get lower priority and that Cursor's coding data could be used to train Grok.52

Windsurf's story has a factual inconsistency that should be flagged. Some sources date Cognition's acquisition to December 2025 at about $250 million, while others place it in July 2025.3140 What happened afterwards is clearer. Cognition shipped its own SWE-1.5 model, which it claims runs about 13 times faster than Claude Sonnet 4.5, along with a codebase-mapping feature called Codemaps.3139 In March it replaced credits with daily and weekly quotas and raised Pro pricing from $15 to $20, which matches Cursor's price.3134 Then on June 2, an update renamed Windsurf to Devin Desktop and moved it under one Devin brand with Devin Cloud, CLI and Review.40 Any comparison still calling it "Windsurf" is describing a brand that no longer exists.

What it adds up to

The pattern across all this coverage is that the market is stratifying, not converging. On the consumer side, Lovable has turned an opinionated, low-friction product into revenue that dwarfs anything its closest rival has disclosed. Bolt is betting that developer control, framework flexibility and Microsoft-backed enterprise distribution will hold up better than raw user growth. Among professional editors, independence has mostly disappeared. Cursor is now part of a frontier-model and compute conglomerate, and Windsurf has been folded into an autonomous-agent brand.

For buyers, the usual "beginner or developer" advice still roughly works. Choose Lovable if you want the fastest guided path to a polished, database-backed MVP. Choose Bolt if you want to see and own every file and are willing to watch your token spend.2126 But tool choice now comes with ownership risk as well. Teams standardizing on any of these tools would be wise to keep their code portable through GitHub sync or export, which Lovable, Bolt and the major editors all support in some form.224152