NEWS
Turin-based Reply has launched Reply
Model Factory, a platform designed to help enterprises build industrial-grade generative AI models using their own operational knowledge, systems and governance frameworks.
Reply launches Model Factory to help enterprises build proprietary generative AI systems
The initiative aims to address one of the biggest enterprise AI challenges: generic models that lack understanding of company-specific processes, terminology and regulatory requirements. Reply said the platform allows organisations to train frontier AI models grounded in internal documentation, software repositories, business data and domain expertise.
The platform creates secure knowledge vaults where enterprise data is prepared and structured for training. Reply Model Factory then manages the full AI lifecycle, including data preparation, training, evaluation, deployment and continuous improvement. The system also supports modular integration with existing technology stacks.
According to Reply, the training layer includes pre-training for domain awareness, supervised fine-tuning for task-specific expertise and reinforcement learning to strengthen agentic behaviour aligned with operational objectives and governance policies. Distillation and speculative decoding are also used to improve deployment efficiency.
Each model and derived dataset is released as a versioned enterprise asset with controlled interfaces, embedded quality controls and alignment with the EU AI Act.
Tatiana Rizzante, CEO, Reply, said proprietary AI models will become a major source of competitive differentiation as organisations increasingly embed generative AI into innovation, decision-making and operational processes.
The platform is designed to help enterprises maintain greater control over intellectual property, compliance and long-term AI development strategies while enabling models to evolve continuously alongside changing business knowledge, operational workflows and industry requirements globally.
SEALSQ targets Orbital AI security market with postquantum semiconductor technologies
SEALSQ has outlined plans to position itself as a cybersecurity and semiconductor provider for the emerging market of orbital AI infrastructure and space-based data centres.
The Geneva-based company said growing demand for Artificial Intelligence processing is pushing aerospace and technology companies to explore orbital computing architectures capable of supporting AI processing, distributed storage and autonomous cloud services directly in space.
SEALSQ believes these next-generation orbital platforms will require quantum-resistant cybersecurity systems built around trusted hardware identity, secure satellite-to-satellite communications and autonomous protection systems capable of operating across distributed constellations.
The company’ s post-quantum semiconductor technologies are designed to support secure hardware authentication, encrypted AI inference and zero-trust communications between satellites, orbital compute nodes and terrestrial infrastructure. SEALSQ added that orbital AI systems create unique cybersecurity challenges because physical intervention is limited in remote space environments. Its secure microcontrollers, cryptographic chips and post-quantum technologies are intended to deliver long-term resilience, low power consumption and protection against evolving cyber threats.
Carlos Moreira, Founder and CEO, SEALSQ, said orbital data centres would become a critical part of the future digital economy as AI infrastructure expands beyond Earth. He said that trust, identity and quantum resilience would need to be embedded directly into semiconductor architecture to protect future autonomous digital interactions and AI processing systems operating in space.
SEALSQ highlighted opportunities around sovereign AI, sovereign cloud infrastructure and blockchain-enabled authentication systems designed to secure machine-to-machine ecosystems across satellite constellations, IoT networks and edge computing environments. The company said capabilities would support European-led space initiatives. www. intelligentcio. com
INTELLIGENT CIO EUROPE
9