The Vibe-Coding Pivot: How a $100 Million AI Startup Rose from the Ashes of a Profitable Marketplace
The trajectory of the technology sector in 2025 is increasingly defined by “vibe-coding”—a paradigm shift where natural language and intent supersede traditional syntax. This transition is perhaps best illustrated by the journey of Anything, a startup that recently secured a $100 million valuation after its founders made the high-stakes decision to shutter a profitable, multi-million-dollar business.
Dhruv Amin and Marcus Lowe, both 33-year-old former Google employees, found themselves at a crossroads in late 2023. Their original venture, a marketplace titled Create, was generating an annualized revenue run rate of $2.2 million. However, the rapid advancement of large language models (LLMs) signaled a looming obsolescence for their business model. By liquidating their existing success to bet on the future of autonomous software generation, they have emerged as leaders in a new category of innovation that empowers non-technical users to build production-ready applications through conversation alone.
From Human Marketplace to Autonomous Generation
The original iteration of the company, Create, functioned as a bridge between startups and freelance engineers. While successful, the platform relied on human labor to fulfill technical requirements. The debut of OpenAI’s ChatGPT in November 2022 served as a “black swan” event for the founders. They recognized that as machine learning models approached human-level proficiency in writing sophisticated code, the need for a marketplace of human developers would effectively vanish.
The Technical Evolution of “Anything”
The transition from a service-based model to a product-based AI company required a complete re-architecting of their technical stack. The founders moved through three distinct phases of development:
Component Generation: Initially, the team launched a tool designed to generate specific app components, such as authentication forms and UI elements.
Backend Orchestration: By early 2025, they expanded the capability to include automated backend systems, integrating payments and database management without manual configuration.
Full-Stack Vibe-Coding: In April 2025, the company rebranded as Anything, launching a platform that allows users to describe an entire business logic in plain English. The AI then synthesizes the front-end, back-end, and infrastructure layers simultaneously.
This technical leap has enabled individuals without engineering backgrounds to launch commercial products. Notable use cases include a dental hygienist developing a gum-health tracking application and a hair salon owner creating a mobile AI stylist app. These are not merely prototypes but “production products” capable of handling real-world transactions and user data.
Strategic Risk and the AI Plateau
The decision to pivot was not without significant anxiety. Amin and Lowe faced “scary” periods throughout 2024 when LLM capabilities appeared to plateau. During this time, they were pressured by the temptation to settle for “quick wins”—such as building simple design tools or prototyping aids for agencies.
However, they maintained a focus on the “agentic” mission: the belief that AI should not just assist in coding but replace the requirement for coding knowledge entirely. This required a conviction that the underlying generative technology would eventually resume its exponential growth curve. The founders’ decision to lay off half their staff and “go back to zero” was a gamble on the timing of this technological breakthrough.
“I expected it would take about a year to get this technology good enough to build real production systems,” Amin noted regarding the transition. “In reality, it has taken about two, and I still think we’re at the start.”
Market Impact: The Rise of the Non-Technical Founder
The success of Anything reflects a broader shift in the venture capital landscape. The $11 million funding round in September 2025, which valued the company at $100 million, underscores a cooling interest in traditional marketplace models in favor of “pure AI” plays.
In the current market, the value has shifted from the “craft” of coding to the “judgment” of product design. As execution becomes commoditized through automation, the competitive advantage lies in identifying unique user needs and market gaps. This democratization of software creation is expected to lead to a surge in “micro-startups”—highly specialized applications built and managed by single individuals who possess domain expertise but lack formal computer science training.
Comparative Growth and Valuation
Metric Create (Sept 2023) Anything (Sept 2025)
Business Model Freelance Marketplace AI Vibe-Coding Platform
Annualized Revenue Run Rate $2.2 Million $2.0 Million (Initial relaunch)
Valuation Undisclosed $100 Million
Core Technology Human Labor + AI Tools Autonomous Agentic Code-Gen
Export to Sheets
Innovation and Future Implications
The emergence of “vibe-coding” as a viable commercial category suggests that the software development life cycle (SDLC) is being fundamentally compressed. What used to take months of discovery, specification, and engineering can now be achieved in a two-week iteration loop.
Anything’s current success—reaching a $2 million run rate within just two weeks of its relaunch—indicates a massive pent-up demand for tools that lower the barrier to digital entrepreneurship. As AI models continue to evolve beyond the current “plateau,” the sophistication of these autonomously built apps is expected to rival those produced by mid-sized engineering teams.
The takeaway for the broader tech industry is clear: the most successful companies of the AI era may be those willing to cannibalize their own profitable models to align with the shifting capabilities of automation. By painting their founding story in “blood, sweat, and tears,” Amin and Lowe have demonstrated that in the age of artificial intelligence, agility and strategic foresight are the ultimate currencies.
Source: https://www.cnbc.com/2025/12/21/ex-googlers-shuttered-successful-startup-built-ai-company-worth-100m.html
Would you like me to analyze the specific “vibe-coding” frameworks and LLM integrations that allow Anything to generate production-ready backend systems?
