Weflayr Secures $2.5M Pre-Seed to Revolutionize AI Inference Routing

Weflayr has secured $2.5 million in pre-seed funding to tackle the growing complexity and cost of AI inference management. By providing an independent intelligence layer for routing workloads, the startup aims to help AI companies move beyond reliance on a few frontier models, optimizing for cost and quality through custom benchmarking and dynamic request matching.

What was announced

The $2.5 million pre-seed round was led by 13books Capital, with participation from a broad group of investors including Founders Future, Kima Ventures, Plug and Play, Station F Ventures, and Roxanne Varza, a General Catalyst scout. The funding will be used to expand the startup’s engineering and AI research teams, accelerate product development, and support a rollout across European and United States markets.

Weflayr’s platform addresses a specific bottleneck in the AI sector: the tendency for companies to stick with a handful of high-profile models because quality is difficult to guarantee across alternatives. The platform allows users to benchmark models using their own proprietary data rather than generic industry benchmarks. This enables companies to define specific routing strategies that dynamically match incoming requests to the best-performing approved models based on their specific workload requirements.

Early performance data from the company's first deployment indicates significant efficiency gains. Weflayr reported that its routing strategy achieved over 55% lower inference costs while maintaining quality levels equivalent to frontier models. In other scenarios, the platform demonstrated a 15 percentage point increase in relative quality when compared to frontier models at an equivalent cost. The company was founded by Romain Mazeries, the former CEO of Mangopay, and Noé Achache, who previously served as Head of AI Engineering at Theodo Data & AI.

“AI inference is rapidly becoming a critical infrastructure layer for the next generation of software companies. As the market evolves and consolidates, we believe there is a massive opportunity for an independent intelligence layer that gives AI companies full control over quality, performance, and economics.”

Mike McFadgen, Partner at 13books Capital

The companies involved

Weflayr enters the market as an infrastructure specialist founded by veterans of the European fintech and data engineering scenes. Romain Mazeries brings experience from his tenure as CEO of Mangopay, a major player in the payments infrastructure space, while Noé Achache provides technical depth from his role at Theodo Data & AI, where he led AI engineering efforts.

The investment is led by 13books Capital, a venture firm with a strong focus on fintech and infrastructure. The firm has a significant presence in the European investment landscape, often backing companies at the intersection of financial technology and enterprise software. Joining them is Kima Ventures, the prolific early-stage fund, and Founders Future, which focuses on supporting ambitious European entrepreneurs.

The round also includes Plug and Play, a global innovation platform and venture fund, and Station F Ventures, the investment arm associated with the world's largest startup campus in Paris. The involvement of Roxanne Varza, acting as a scout for General Catalyst, further highlights the interest from major international venture players in Weflayr’s approach to the AI infrastructure stack.

What FF News has reported before

FF News has followed the activity of Weflayr’s investors across several recent high-profile rounds in the fintech and AI sectors. This includes coverage of Cleavr Secures €8M Funding to Accelerate AI-Powered Accounts Receivable Automation Across Europe, a move that signaled growing appetite for AI-driven financial efficiency.

The publication also reported on Outline Secures $3M to Launch First Agentic Financial Planning Platform for Mid-Market CFOs and Seoul Labs Debuts SeoulLabs Pay to Secure Autonomous AI Agent Payments, highlighting the shift toward agentic AI systems. Additionally, FF News covered Aria Secures €247M in New Funding to Combat Europe’s Late Payments Crisis, reflecting the massive scale of capital moving into infrastructure-heavy fintech solutions.

What this means

The emergence of Weflayr signals a shift in the AI market from model experimentation to operational maturity. As inference costs become a primary line item for software companies, the "black-box" approach to model selection is no longer sustainable. By moving away from generic benchmarks toward workload-specific routing, the industry is acknowledging that "one size fits all" frontier models are often an expensive overkill. This puts pressure on major model providers to justify their premiums while opening the door for specialized, smaller models to gain market share through intelligent routing. The central question for the sector is whether independent layers like Weflayr can maintain their neutrality as the market consolidates.