Last Money in is Powered by Sydecar

Sydecar empowers syndicate leads to manage their investments more effectively. Organize, manage, and engage your investor network effortlessly with Sydecar’s management and communication tools. Their platform also automates banking, compliance, contracts, tax, and reporting, freeing up syndicate leads to focus on securing deals and strengthening investor relations. Elevate your syndicate operations with Sydecar.

Is Cursor Going to Zero? The Bird Scooter of Today's AI Boom

In the rapidly evolving landscape of AI-powered development tools, a familiar story is potentially beginning to emerge. Cursor, the AI coding assistant that rocketed from $1 million to $500 million in annual recurring revenue (ARR) in just two years, may be following a trajectory eerily similar to Bird, the electric scooter company that became the fastest startup ever to reach $1 billion in valuation—only to crash spectacularly when unit economics caught up with hypergrowth.

While the potential for AI remains genuinely exciting, certain areas of the AI boom are reminiscent of 2019-2021, specifically the lack of scrutiny around unit economics that ultimately proved fatal for many high-flying startups.

The Bird Parallel: High-Flying Growth, No Gross Margins

Bird's meteoric rise and fall offers a cautionary tale for today's AI boom. The scooter company became the fastest startup ever to reach a $1 billion valuation, eventually raising $700 million in total funding at a peak valuation exceeding $2 billion. Yet beneath this impressive growth lay disastrous unit economics. In 2018, Bird generated revenues of $58.5 million, but saw gross margins of -$212.2 million and net loss of $367.4 million. In 2019, Bird's revenues reached $150.5 million, but gross margins remained deeply negative at -$135.7 million. When investor patience finally waned and market sentiment shifted, Bird filed for Chapter 11 bankruptcy protection having accumulated over $1.5 billion in losses.

While it's common for venture capitalists to subsidize early losses (and often rightfully so), the concern is that Cursor may lack the unit economics, competitive moat and/or long-term pricing power to avoid reaching the same fate as Bird. The AI coding tool has achieved impressive growth metrics—360,000 paying users and $500 million ARR—but mounting evidence suggests the company is operating with razor-thin gross margins, a limited moat, and increasing competition from the underlying model companies, specifically Anthropic's Claude.

The Margin Crisis: When AI Costs More Than Revenue

The core pricing issue plaguing Cursor mirrors Bird's fundamental problem: the cost of delivering the service often exceeds the revenue generated from users.  

• Unsustainable Unit Economics: Posts from users estimate that heavy Cursor users cost the company $60 per session and more while paying only $20/month for the Pro plan. This suggests that power users—likely Cursor's most engaged customers—are operating at massively negative margins.

• Rising Input Costs: Cursor faces increasing pressure from AI model providers. Token prices from Anthropic, OpenAI, and other LLM providers have changed for Cursor, with Claude Opus 4 now costing est. $15 per million input tokens and $75 per million output tokens. Since Cursor doesn't control these models, it has minimal to no leverage over its primary cost driver.

• Loss Leader Strategy: Cursor's approach is a "loss leader strategy," prioritizing market share over profitability. This works for many, many companies, but without a strong moat, I’m skeptical. Subsiding pricing for market share is generally viewed as acceptable. Doordash did with success. Uber did it with success. As have many, many other VC backed companies. But they both had extremely strong moats - the power of network effects and with that pricing power and operating leverage. Right now it’s not clear Cursor has those same benefits unfortunately, and in my view, is looking more like Bird than Doordash.

There’s quite a bit of evidence for this.

Cursor's pricing overhaul in June 2025 revealed the strain on its business model. The company shifted its $20/month Pro plan to a token-based system and introduced a $200/month Ultra plan—an attempt to align revenue with soaring AI costs. The backlash appeared pretty intense.

Users reported unexpected charges as high as $100 per day, with unclear usage limits leading to billing surprises. The CEO's public apology and promise to cover excess API fees demonstrated that Cursor was willing to absorb costs rather than lose users—a clear sign of margin distress and lack of pricing power…

Perhaps most troubling for Cursor's long-term prospects is the emergence of direct competition from the very companies it depends on. Anthropic's Claude Code, built on the same Claude 3.5 Sonnet and 3.7 models that power Cursor, offers similar functionality without the middleware costs.

Early adopters are already making the switch. Users report migrating to Claude Code, citing better performance and cost-effectiveness. 

This competitive pressure raises a critical question: what is Cursor's sustainable competitive advantage? 

The company operates primarily as a sophisticated middleware layer, routing queries across multiple AI providers (Anthropic, OpenAI, Google, xAI) while optimizing the user experience. 

Cursor doesn't control its core technology, making it vulnerable to direct competition from model providers. AI coding capabilities are rapidly becoming table stakes, with multiple providers offering similar functionality. Unlike traditional social products, marketplaces and some other SaaS products, Cursor's value doesn't increase enough with more users. 

The concern extends beyond individual users to the broader business model. If Cursor's margins don't improve as it scales and the company begins to legitimately lose market share to Claude and others, the path forward becomes dire extremely fast. A bloated cap table could make it challenging for another party to provide any kind of rescue financing. 

The Verdict?

While Cursor isn't destined for immediate collapse, the parallels with others that have struggled with similar problems are too striking to ignore. Rapid growth through unsustainable unit economics, real questions around customer lock-in, and increasing competitive pressure from established players all raise significant concerns. Cursor needs its moats to kick in so it can increase pricing or build leverage, of which today I don't see either.

However, the AI and SaaS landscape offers several instructive examples of companies that navigated similar challenges successfully. Datadog succeeded despite competing with AWS CloudWatch and other native monitoring solutions by offering better cross-platform integration and user experience. More recently, OpenAI's ChatGPT Enterprise demonstrated that AI middleware can command premium pricing when it provides genuine workflow integration and enterprise features. Even Zapier, which essentially routes data between APIs, built a billion-dollar business by making integrations seamless and reliable. The key differentiator may be whether Cursor can evolve beyond a simple wrapper to become deeply embedded in their users' workflows—a transition that requires significant product development but has historically proven achievable.

I don't know what's going to happen here, and for context, I HOPE Cursor thrives! But it's hard to look past the flashing warning lights.

If you enjoyed this article, feel free to view our prior posts on adjacent topics 

• Why Syndicate Leads Don’t Survive 

• The Syndicate LP Churn Problem 

• What Really Drives Investment Interest in SPVs

Last Money in is Powered by Sydecar

Sydecar empowers syndicate leads to manage their investments more effectively. Organize, manage, and engage your investor network effortlessly with Sydecar’s management and communication tools. Their platform also automates banking, compliance, contracts, tax, and reporting, freeing up syndicate leads to focus on securing deals and strengthening investor relations. Elevate your syndicate operations with Sydecar.

If you enjoyed this post, please share on LinkedIn, X (fka Twitter), Meta and elsewhere. It goes a long way to support us!

Follow the Last Money In authors on LinkedIn

✍️ Written by Zachary and Alex