The Brooklyn skyline just got a new reflection, this one built from silicon and swagger. Reflection AI, the open-source superintelligence lab founded in 2024 by former DeepMind researchers Misha Laskin and Ioannis Antonoglou, just raised a jaw-dropping $2B Series B at an $8B valuation. Led by NVIDIA with heavy hitters B Capital, Citi Ventures, DST Global, Hillspire, GIC, and 1789 Capital in the mix, plus Eric Schmidt and Eric Yuan lending strategic muscle, this isn’t just a funding round, it’s an inflection point. Open-source AI just went mainstage.
Laskin and Antonoglou didn’t come to play. Between them, they’ve touched some of the most advanced systems ever built, Gemini, AlphaGo, MuZero. Now they’re building an intelligence stack that doesn’t hide behind NDAs or gated APIs. Their Brooklyn HQ on Kent Ave is home to a team of ~60 researchers and engineers pushing code comprehension to the frontier. Not generation. Not suggestion. Actual understanding. Their flagship product, Asimov, doesn’t just fill in lines, it reads, interprets, and contextualizes entire codebases. In blind tests with open-source maintainers, Asimov’s responses were preferred over Cursor Ask and Claude Code more often than not. That’s a quiet revolution happening in plain sight.
This $2B round pushes Reflection AI’s total raise to $2.13B, after a $25M seed (led by Sequoia & CRV) and a $105M Series A (co-led by Lightspeed & CRV). Funding like that isn’t about luck, it’s earned by vision and execution. And the mission here is clear: make superintelligence accessible, not locked up behind closed labs. Reflection AI’s architecture blends LLMs, reinforcement learning, and Mixture-of-Experts at scale, training on 10T+ tokens across text, code, image, and audio.
Now comes the expansion. The new capital is earmarked for compute scale-up using NVIDIA DGX SuperPODs, global R&D growth in London & Toronto, and an internal safety lab to ensure frontier AI doesn’t lose the plot. Leadership is equally stacked: Sarah Chen (ex-OpenAI) as VP Eng, David Kim (ex-Anthropic) leading ML Safety, and Aisha Patel (ex-Databricks) steering the finance ship. Board guidance from Eric Schmidt and Eric Yuan keeps this lab wired to both capital and conscience.
Reflection AI’s next moves are set, frontier LLM launch in Q2 2026, multimodal systems by H2 2026, and sovereign AI deployments on deck. This isn’t about chasing OpenAI or Anthropic; it’s about proving that openness can scale. That an AI born in Brooklyn can hold its own on the global stage. The irony? In a space obsessed with synthetic thought, Reflection AI might be the most human move yet, clarity, transparency, and a little Brooklyn grit written right into the code.
