Austria-based Mostly AI, which simulates synthetic data for AI training, data analytics, and software testing, raises a $25M Series B from Molten Ventures
Austria-based Mostly AI, a startup that simulates synthetic data for AI model training and testing, today announced it has raised $25 million …
Context & Ripple Effects
Mostly AI's $25M Series B from Molten Ventures lands at the front of a funding wave for companies that manufacture training and test data rather than collect it. Two months later, Datagen raised a $50M Series B for synthetic data aimed at computer vision, confirming investors see generation-side tooling as its own category rather than a niche.
The category also has a downstream complement in the corpus: Synthesized's $20M Series A for AI-driven software testing and, later, Vals' $40M Series A for model evaluations — together with Mostly AI they sketch a stack where synthetic data feeds models and independent tooling grades them.
First-order effects
- Molten Ventures gains a position in European synthetic-data infrastructure, and Mostly AI gets the capital to push its platform across its three stated uses: AI training, analytics, and software testing.
Second-order effects
- Datagen, which raised at roughly double Mostly AI's round size for computer-vision-specific synthesis, now faces a well-funded generalist rival spanning training, analytics, and testing — pushing differentiation toward vertical depth versus breadth.
Third-order effects
- If the pattern holds, AI development splits into distinct funded layers — data generation (Mostly AI, Datagen) and model evaluation (Synthesized, Vals) — with enterprises buying assurance about models from vendors who never touched the underlying real-world records.
The trend: Venture capital is consolidating a synthetic-data-and-evaluation layer around AI models, treating generated data and independent testing as core infrastructure rather than afterthoughts.