99% of all AI today is black-boxed — you don’t know how results are generated. Meanwhile, a startup called Cognifiber is turning optical fibers into processors that could accelerate AI inference by 1,000x while cutting power consumption by 8,000x. This panel brings together an AI prediction company, a photonic computing startup, a data interoperability firm, and Microsoft to debate how AI is rewiring financial services from risk assessment to fraud detection.

WHAT THIS PANEL COVERS

• How AxonJ uses hybrid reinforcement learning and real-time signals — like website changes or stopped TikTok campaigns — to predict company risk 40% more accurately than balance-sheet models alone
• Why Cognifiber’s fiber-optic processors could shrink city-sized data centers to a single building while democratizing AI access
• How MMC Convert helps banks find “invisible creditworthy” people through utility bill patterns and psychometric AI models
• Microsoft’s perspective on multi-agent orchestration, low-code platforms, and why the dollar-per-token cost is actually falling
• The centralized vs. decentralized debate: why on-premises AI uses up to 80% less electricity and drives democratization

PANELISTS

• Robin — Moderator
• Jean-Philippe (JP) — Founder & CEO, AxonJ
• Eyal — Founder, Cognifiber
• Shanu — Founder, MMC Convert
• Ujjwal Kumar — Principal Architect, Office of the CTO, Microsoft

unDavos is a community-driven summit running during WEF week in Davos, democratizing the conversation around global challenges.

🌐 undavos.com

Tags: AI financial services, fintech, agentic AI, photonic computing, data interoperability, risk prediction, fraud detection, Microsoft, decentralized AI, energy consumption, machine learning, reinforcement learning, banking, insurance, data centers, AI governance, enterprise AI, unDavos, Davos 2026, WEF