Bay Area-based AI coding platform Base44, which Wix acquired for $80 million just a year ago when the startup was only six months old and employed eight people, has launched the rollout of its proprietary large language model (LLM), Base1. The move marks a significant milestone in the company’s strategy to strengthen its AI-powered application development platform while reducing dependence on third-party frontier models.
Base44 enables users to build software applications using natural language prompts, making app development accessible to both technical and non-technical users. With the introduction of Base1, the company aims to deliver faster, more cost-efficient, and highly optimized AI experiences tailored specifically for app creation.
The rollout comes as the AI industry continues to debate whether general-purpose frontier models remain the best solution for every use case. At the same time, investors and technology leaders increasingly question whether startups built entirely on external AI models can maintain long-term competitive advantages.
Rather than relying solely on third-party AI providers, Base44 has chosen to invest in its own infrastructure by developing a proprietary model trained specifically for its platform.
Explaining the strategy, Maor Shlomo, Founder of Base44, said, “Training and owning the model as part of [our] entire stack allows us a lot more optimizations on latency, cost, and efficiency.”
Initially, Base44 will continue refining Base1 while gradually expanding its deployment across the platform. Over time, the company expects its custom model to outperform general-purpose frontier models for application development tasks by leveraging platform-specific data and workflows.
The launch also positions Base44 more directly against competitors such as Swedish AI startup Lovable, which achieved unicorn status following its Series A funding round and currently relies on external large language models to power its platform.
However, Shlomo believes that other successful AI application companies will eventually follow a similar path.
According to him, “At least the players that have gotten enough scale and velocity to have enough data.”
Industry experts also view proprietary data as a critical competitive advantage in artificial intelligence.
Jonathan Userovici, General Partner at VC firm Headline, whose investment portfolio includes companies such as Mistral AI, explained that successful AI startups build defensibility through three key pillars: distribution, proprietary data, and technology infrastructure.
Reflecting that strategy, Base44 revealed that it trained the first version of Base1 using a proprietary dataset generated from tens of millions of real user interactions on its platform. As user activity grows, the company expects this dataset to become increasingly valuable in improving model performance and personalization.
Nevertheless, competition within the AI coding market continues to intensify. Beyond emerging vibe coding startups, frontier AI companies such as Anthropic, xAI, and Cursor have expanded aggressively into AI-powered software development. These companies also benefit from extensive user feedback and training data that continuously improve their foundation models.
Despite that competitive landscape, Shlomo believes specialized AI models will continue to outperform general-purpose systems for specific enterprise workflows.
“Models are progressing, but they’ll stay very general in what they can do,” he predicted.
Meanwhile, Userovici cautioned against dismissing frontier AI providers too quickly. He pointed to legal technology company Harvey, which ultimately abandoned its plans to develop its own foundation model in favor of existing frontier AI systems.
Instead, Userovici suggested that AI companies increasingly focus on optimizing model selection rather than replacing frontier models altogether.
He explained, “They don’t necessarily see a [return on investment] when using the latest models for all use cases, so an entire infrastructure is being set up to do orchestration and optimization to select the right models for them so that costs don’t skyrocket while maintaining the same or similar performance across the majority of use cases.”
Cost optimization has become especially important as enterprise customers account for a growing share of revenue across AI application platforms. Businesses increasingly seek AI solutions that balance performance with affordability instead of relying exclusively on the most expensive frontier models.
Accordingly, Base44 expects Base1 to improve both customer experience and long-term operating economics.
Highlighting those objectives, Maor Shlomo said, “We want to get a model that is going to be more aligned to what we think is the right thing, is going to be more optimized to what we see users like in terms of the results we’re getting, and is going to be faster and cheaper for customers eventually than using the frontier models like Opus.”
The company also stated that owning its own AI model gives it direct control over compute infrastructure and inference spending, which should strengthen profit margins over time despite the significant engineering investment required to develop Base1.
The launch comes as parent company Wix recently announced plans to reduce approximately 20% of its workforce. In contrast, Base44 has continued expanding its team following the acquisition while maintaining strong business momentum.
Earlier this year, Base44 announced that it had surpassed $100 million in annual recurring revenue (ARR). Although that figure remains below competitor Lovable, which recently disclosed $500 million in ARR, Base44 believes its vertically integrated approach will provide sustainable long-term advantages.
According to Shlomo, the extensive engineering effort behind Base1 supports the company’s ambition to become the only vertically integrated vibe coding platform, controlling its distribution, proprietary data, and AI infrastructure under one ecosystem.
As competition within the AI coding industry accelerates, Base44’s investment in Base1 reflects a broader shift toward vertical integration and proprietary AI development. By owning its model, training data, and platform infrastructure, the company aims to improve performance, reduce costs, strengthen margins, and differentiate itself in one of the fastest-growing segments of artificial intelligence.
