Oxide AI launches new Hybrid AI generation for finance

Oxide AI says it has launched a new generation of its Polychaos Hybrid AI platform to help enterprises make high-stakes decisions, with an initial focus on finance, risk and investing. The move comes as the Malmö-based company expands through partner-led growth and reports traction with banks, a Nordic risk data company and a retail investing platform.

Why it matters: - Oxide AI is targeting decision-making in finance, where the company says standalone large language models have not gained much ground. - The new Polychaos release is designed for regulated, high-stakes use cases where explainability, privacy and regulatory requirements matter. - Oxide AI says the platform is already producing commercial results in credit risk, risk intelligence and retail investing.

What happened: - Oxide AI launched a new generation of Polychaos, its Hybrid AI platform, on Sept. 22, 2026. - The company said the platform is built to solve complex problems and automate high-stakes decisions in competitive business environments. - Chief Product Officer Kateryna Wikström said the release is a major step forward and that the fastest way for customers to evaluate it is to run Oxide AI’s hybrid system head-to-head against existing solutions. - CEO Anders Tylman-Mikiewicz said Oxide AI is investing most of its resources in technology development and scaling through strategic partners with global reach.

The details: - Oxide AI said its credit risk modeling work with banks is helping improve credit decisions and shorten time-to-market for new models that meet regulatory requirements. - The company said that credit risk use case is being prepared to scale with a global partner. - Oxide AI said it has entered a multi-year partnership with a leading Nordic risk data company that acquired Oxide AI’s Redox system to build a risk intelligence offering for know-your-customer and anti-money laundering use cases. - The company said it co-developed a retail investing capability that answers plain-language stock-screening questions with near-real-time analysis across every US listed company. - Oxide AI said that retail investing capability was built with a leading retail investor platform. - Chief AI Officer Lars Hård said the new generation uses large-scale code synthesis to create customer-specific value in fully isolated environments. - Hård said advanced quantitative optimization in finance requires a different stack than an LLM alone can provide. - Oxide AI said its architecture is built to keep AI private and sovereign in high-value decisions. - The company said the new generation extends its lead in use cases where privacy and sovereignty are table stakes. - Oxide AI is headquartered in Malmö, Sweden. - The company was founded in 2020. - Oxide AI said it spent five years developing a proprietary architecture that specializes around each customer’s business, learns continuously after deployment and produces fully explainable output that meets regulatory requirements. - The company said the architecture is patent pending.

Between the lines: - Oxide AI is positioning itself against the limits of language-first AI and against generic enterprise AI products in regulated markets. - The company’s partner-led model suggests it wants to prove a use case directly, then scale through firms that already control customer relationships. - Oxide AI’s emphasis on credit risk and KYC/AML points to a strategy focused on problems with clear financial upside and strong compliance pressure. - The company cited a16z figures showing AI startups earn $3 billion a year from coding versus $50 million from financial and investment analysis, underscoring why finance remains a harder monetization target.

What's next: - Oxide AI said the credit risk use case is being prepared for broader scale with a global partner. - The company expects its partner-based growth model to continue expanding as it repeats the direct-relationship-to-partner rollout pattern. - Oxide AI said demand is accelerating as customers test its hybrid system against current solutions.