Microagi's $55M Seed, Google Cloud & NVIDIA Partnership, and CEO's 18-Month Warning on Europe's AI Infrastructure Gap

The round was led by Hummingbird Ventures, with participation from Northzone, LocalGlobe, Village Global, and redalpine . The funding is being used to expand Atlas, microagi's data and deployment platform for industrial robotics, and to collect real-world physical data — including footage of ordinary people doing chores — to train robots for factory tasks . As Kilic told Sifted: "It's one-billionth of what Europe needs."

Partnership with Google Cloud and NVIDIA

Just one week after the seed round, on July 22, 2026, microagi announced a strategic collaboration with Google Cloud, using NVIDIA's Blackwell platform to accelerate embodied AI model training and deployment .

Under the partnership, Atlas now trains and serves models on NVIDIA GB300 NVL72 rack-scale systems and RTX PRO 6000 Blackwell Server Edition GPUs running on Google Cloud's AI-optimized infrastructure . The software stack also integrates Google Cloud's Gemini Enterprise Agent platform alongside NVIDIA's hardware-level optimizations . Engineering teams from all three companies are directly involved in the collaboration .

Kilic framed the deal: "Robots learn from data, and data is useless without the compute to train on it. This partnership gives us scale and engineering depth we could not build alone at this stage: NVIDIA optimises down to the card, Google Cloud across the whole machine."

Notably, microagi stated that Google Cloud is not receiving its training data or customer data in return for compute access .

Atlas Platform: Task-Specific Robotics AI

Atlas is microagi's core product — a data and deployment platform that helps industrial companies put robots into production for specific factory tasks . Rather than building its own robots, microagi develops hardware- and model-agnostic software that fine-tunes AI models using each customer's operational data .

The platform collects real-world movement data from human demonstrations (e.g., assembly, sorting, handling) and uses it to train task-specific embodied AI models for humanoid and commercial robotic platforms . Since the Google Cloud / NVIDIA partnership, Atlas runs its entire training and inference pipeline on NVIDIA Blackwell infrastructure through Google Cloud .

CEO Bercan Kilic's 18-Month Warning on Europe's Sovereign AI Gap

Bercan Kilic has publicly warned that Europe has a narrow 12 to 18-month window to build sovereign embodied AI compute infrastructure, or risk an insurmountable technology gap with the US and China .

The core concern: Europe lacks the large-scale GPU compute clusters needed to train embodied AI models at the pace of US and Chinese competitors. Without urgent investment, Kilic argues, European robotics companies will become permanently dependent on non-European cloud infrastructure .

Kilic framed microagi's own need to partner with US cloud providers (Google Cloud / NVIDIA) as a symptom of Europe's infrastructure gap — even a record seed round cannot buy what doesn't exist locally . He has urged European policymakers and investors to treat compute infrastructure as strategic sovereign infrastructure, similar to energy grids or transportation networks .

As Kilic put it: "Industrial Europe cannot reindustrialise on borrowed compute." "If you run factories in Europe, the maths is already on your desk. Your most experienced people retire this decade, and there is no one behind them."