On August 28, 2026, OpenAI told the AI coding editor Cursor it was ending the contract that supplies Cursor with OpenAI’s models, effective November 12. The trigger wasn’t a technical failure or a policy violation — it was ownership. SpaceX had just completed a $60 billion all-stock acquisition of Cursor’s parent company, Anysphere, and OpenAI decided it couldn’t be confident SpaceX would honor its usage terms, citing past friction with other Musk-owned companies. A contractual change-of-control clause gave OpenAI a window to walk away, and it used it.
Cursor itself will likely be fine. Its own CEO says OpenAI’s models account for only about 5% of its traffic, and both Anthropic’s and Google’s models remain fully available. But step back from the specific companies involved, and the shape of what happened is worth sitting with: a tool millions of people built daily workflows around had a core capability switched off, on a timeline set entirely by a dispute between two other companies, over a contract clause none of those users had ever read or agreed to.
That’s not really a story about AI coding assistants. It’s a case study in vendor lock-in — and the exact same structural risk sits underneath a decision far more small business owners make without a second thought: which platform their own website runs on.
How Cursor Got Here
The scale of what just happened is easy to undersell. Cursor’s revenue went from roughly $100 million annualized in January 2025 to about $4 billion by June 2026, doubling every three months for six straight months, with enterprise customers eventually driving 60% of that revenue. SpaceX’s $60 billion acquisition price — roughly 15 times revenue — made it, by several accounts, the largest acquisition of a venture-backed startup on record. Cursor had reportedly turned down two separate approaches from OpenAI itself before agreeing to the SpaceX deal, having prioritized staying independent.
OpenAI’s public explanation was unusually direct. In its own announcement, the company pointed to Twitter (now under SpaceX/xAI ownership) previously breaching an OpenAI contract, and to Elon Musk acknowledging under oath that xAI had violated OpenAI’s terms of service. It gave Cursor the maximum notice its contract allowed — but the decision itself wasn’t up for negotiation. It was a right OpenAI held all along, triggered the moment ownership changed hands.
What Vendor Lock-In Actually Costs
The Cursor situation is a dramatic, headline-grabbing version of a much more mundane risk that shows up in ordinary business software decisions constantly. A few figures put a number on it:
- The average cost of a full vendor migration for a business runs around $315,000, according to Kong’s 2026 analysis of enterprise vendor lock-in — and that’s before counting the operational disruption of the switch itself.
- Organizations that get trapped in a lock-in situation without having planned an exit path face switching costs roughly 16 times higher than those who thought about portability up front, per GainHQ’s 2026 research on vendor lock-in.
- Buyer behavior has shifted in response: 74% of SaaS buyers now evaluate potential switching costs before they sign up for a platform, up from just 47% in 2018, according to Deloitte data cited in recent SaaS pricing research. People have learned this the hard way.
- Gartner research shows 58% of customers who come to feel “trapped” by a vendor eventually leave anyway — often as vocal detractors — meaning lock-in doesn’t even reliably protect the vendor relationship it’s designed to preserve, per the same analysis.
The mechanism is almost always the same, whether it’s an enterprise buying AI infrastructure or a small business picking a website builder: the platform becomes so woven into daily operations — your data, your customer history, your design, your workflows — that leaving stops being a decision and starts being a rebuild.
The Same Risk Sits Under Your Website
Most business owners never think about this until they try to leave. It’s worth checking before you’re stuck, because the popular all-in-one website builders are explicit, closed ecosystems by design:
- Wix does not let you export your website to another host at all. If you decide to leave, the domain name can move with you — the actual website cannot. You rebuild from nothing, a limitation documented in detail by web design agencies who migrate clients off it.
- Squarespace offers an export, but it’s a content export, not a site export: it covers page text and a single blog, without your design, CSS, or original image files — those stay as links back to Squarespace’s own servers, per the same analysis.
- Even where an export exists, it’s often only partial. Webflow lets you export HTML and CSS but leaves out CMS content, user accounts, e-commerce functionality, and native forms; Shopify lets you export store data but not a working store, so it still has to be rebuilt on whatever platform you land on next, according to a 2026 comparison of website-builder portability.
- Even a straightforward migration off a closed platform adds friction by design — one guide notes that Wix’s lack of direct database export alone adds an average of 4.2 hours to what should be a routine move, before any actual rebuilding starts.
None of this is a defect in these products — they’re built this way deliberately, because a website you can’t easily leave is a customer you don’t have to re-win every renewal. That’s a completely rational business model for the platform. It’s a completely different question whether it’s the right foundation for your business to build years of content, SEO equity, and customer trust on top of.
The Question to Ask Before You Build (or Rebuild)
You don’t need to predict the next $60 billion acquisition to protect yourself from this. You need one honest answer to a simple question, for every platform your business depends on: if you had to leave tomorrow, what would that actually take?
- Can you export everything — content, design, customer data, and functionality — not just the text?
- Can the result be self-hosted anywhere, or does it only run on the vendor’s own servers?
- Do you own the domain, the code, and the content outright, or are you renting access to all three?
- If the vendor raises prices, sunsets a feature you depend on, or gets acquired by someone whose priorities don’t match yours, what actually happens to your site the next morning?
If the honest answer to “what would leaving take” is “export my content and rebuild the rest from scratch,” you’re not choosing a website platform — you’re renting one, and the invoice for that arrangement just isn’t due yet. This is precisely why open, self-hostable foundations like WordPress remain the standard recommendation for a business site meant to last: your files, your database, and your design are actually yours, and moving hosts or agencies later is a technical task, not a demolition.
This is the same due diligence question the Cursor situation just forced thousands of individual developers to ask overnight, except a business’s website is usually far more central to revenue than a code editor. At Craftwebx, this is a foundational part of how we build: on platforms and code you own outright, so a decision made in someone else’s boardroom never becomes an emergency in yours.
FAQ
Why did OpenAI cut off Cursor specifically, and not other AI coding tools?
The trigger was ownership, not the product. Cursor's parent company, Anysphere, was acquired by SpaceX in a $60 billion deal. OpenAI's contract with Cursor included a change-of-control clause allowing it to exit if ownership changed, and OpenAI exercised that right, citing prior contract friction with other Elon Musk-owned companies.
Does this affect everyone who uses Cursor?
No. Cursor's CEO stated OpenAI's models handle about 5% of Cursor's traffic; users on Anthropic's Claude or Google's models are unaffected. Only workflows specifically built around OpenAI's GPT models inside Cursor lose access, and only after November 12, 2026, as things currently stand.
What's the difference between a website builder and a self-hosted platform like WordPress?
A website builder like Wix or Squarespace hosts your site on its own infrastructure using proprietary, closed formats — you're licensing access, not owning a portable asset. A self-hosted platform like WordPress.org runs on hosting you choose and control, using open file formats, so the entire site can be moved, backed up, or rebuilt elsewhere without starting over.
Is switching away from a locked-in platform ever worth the cost?
It depends on how much value you're losing by staying versus what a move costs. The data suggests waiting makes it worse, not better: switching costs compound as more content, integrations, and workflows accumulate on a platform, so the cheapest time to evaluate an exit is before you've built years of content on top of it, not after a price increase or feature change forces the question.
What's the single most important thing to check before choosing a website platform?
Whether you can fully export your content, design, and functionality and self-host it elsewhere. If the honest answer is no, you're accepting a rebuild-from-scratch risk the moment pricing, features, or ownership changes in a way you don't control.
