Vibe coding platform Base44 is launching a model of its own as AI startups seek to defend themselves

Rule44a dynamic coding platform that Wix was acquired for $80 million Just one year ago — when the company was barely six months old and had a team of eight — it began rolling out its AI model to support its users in creating natural language apps.

The move comes as debate intensifies in AI circles about whether parametric models are the best fit for all use cases. A related question is whether companies built on someone else’s models are truly defensible in the long term. The latest movement by Base44, based in the Bay Area, speaks to both.

Although its custom LLM software has only just been launched, Base44 hopes it will eventually outperform the leading models. According to its founder, Maor Shlomo, “Training and having the model as part of our full stack allows us to further improve latency, cost and efficiency.”

At first glance, this may be a way to stay ahead of competitors such as The Swedish startup is belovedwho arrived Rhino case In the first round of the series last summer and that Depends on external LLMs. However, Shlomo expects others to train their own models — “at least players who have enough volume and speed to get enough data.”

According to Jonathan Osierovichi, general partner at VC firm Headline — whose portfolio includes AI companies like Mistral AI, but not Base44 — data is one of three key components of defensibility for AI startups, along with distribution and the technology stack.

The result is that players with strong brands now rely on their data and infrastructure to increase their defensibility, and Base44 fits this pattern. The first version of the LLM software, Base1, was developed and trained on a dataset generated from “tens of millions of real user interactions on the platform,” the company says.

This data set will continue to grow with the company; But its competitors will, too. The biggest competition may not be from biocoding startups at all, but instead comes from frontier AI labs closer to home for Base44 — now both Cursor and Grok are xAI’s parent company Subsidiary to SpaceXand Cloud Code has become a popular player in the programming space in its own right.

This gives Anthropic and other enterprise AI providers access to data and feedback loops they can use to improve their app creation models, but Shlomo believes specialization gives Base44 an advantage. “Models are progressing, but they will remain very general in what they can do,” he predicted.

For his part, Oserovici warned against underestimating leading models, citing the example of legal technology startup Harvey, which abandoned plans to train its own model. He doesn’t expect applied AI companies to become leading laboratories en masse, but he places Base44’s move in a broader context, one in which inference costs have become a meaningful part of the equation.

Oserovici says cost pressure has led to the change that enterprise customers are now demanding. “They don’t necessarily see a[return on investment]when using the latest models for all use cases, so a whole infrastructure is set up to do orchestration and optimization to choose the right models for them so that costs don’t skyrocket while maintaining the same or similar performance across the majority of use cases.”

Enterprise companies are still a minority among the platform’s enthusiast population, but they represent a growing share of platform revenue, and users of all sizes are beginning to express concerns about the cost of using AI. Base44’s decision to develop its own MBA stems from multiple factors, but cost reduction is likely among the benefits.

“We want to have a model that will be more aligned with what we think is the right thing, will be more optimized for what we see users like in terms of the results we get, and will ultimately be faster and cheaper for customers than using frontier models like Opus,” Shlomo said.

As for Base44 itself, the cost reduction is not so obvious. In a press release, the company explained that “ownership of the model gives Base44 direct control over computational and inference spend, and is expected to result in a structurally stronger margin profile over time.”

Even with the delayed payment, improved margins would be good news for Base44’s parent company, which recently announced it would do so. Laying off 20% of its workforce. In contrast, Base44’s headcount has increased since the acquisition – and announced this Exceeded $100 million in annual recurring revenue A few months ago.

This is still less than likable, who said that It hit $500 million in ARR earlier this month. But Shlomo is betting that the “massive engineering effort” to develop Base1 will cement Base44’s position as “the only vertically integrated implementation of vibrato coding — meaning, in Userovici’s terms, a player that owns its distribution, data, and infrastructure simultaneously.”