Day One for Niantic Spatial
Helping AI Leap From Screen to the Real World
Today marks day one for Niantic Spatial Inc., a new company with one goal: to build the world’s leading AI for understanding and interacting with the real world.
The challenge: Today’s AI understands text, code, and images, but it doesn’t understand the world we live in – the places we explore, the objects we interact with, or the context that shapes real-world decisions.
Existing digital maps aren’t enough. We need a new kind of map, one that allows machines, from smart glasses to humanoid robots, to comprehend the physical world.
Our mission: Niantic Spatial was founded to solve this challenge. Our mission is to build a Large Geospatial Model to fuse the physical and digital worlds - transforming how people and machines understand and interact with the real world to improve how we live and work.
Why now: We’re starting Niantic Spatial with $250 million in funding, following the sale of Niantic’s games business to Scopely, a transaction completed this week. We’re launching with technology and products built on a smart, living 3D map of the world:
Our near-term goal is serving enterprise customers in specific geographic areas, like warehouses and entertainment sites. Our long term goal is to build and train a Large Geospatial Model that will be the foundation for spatially intelligent AI that can power everything from AR glasses to autonomous systems on a global scale.
-
The Niantic Spatial Platform – Our spatial services include Niantic Spatial’s industry leading Visual Positioning System, advanced 3D scanning and visualization with Gaussian Splatting; and a suite of developer SDKs. Deployments and proof-of-concepts are already underway across logistics, real estate, manufacturing, entertainment, and federal customers.
-
Large Geospatial Model – Built on over 30 billion posed images from millions of locations around the world, this model will enhance Large Language Models with spatial reasoning capabilities that mean they can understand not just where things are, but also what they are and how they relate to the surrounding environment.
Who we are: Our CEO John Hanke and CTO Brian McClendon are the pioneers behind Keyhole, acquired by Google and later evolving into Google Earth and Google Maps. Our team and research groups in London and Oxford hold more than 240 patents and 238 patents pending, with dozens of papers published at top computer vision conferences.
We’re just getting started: While LLMs are revolutionary in their own right, the majority of human life and economic activity still happens in the physical world. We have been fortunate through the evolution of Niantic to build a world class R&D team, to create unique geospatial models, and to deploy early versions of our tech at scale to millions of users. Now, we’re ready for the next phase in the AI revolution: the leap from screen to the real world.
Sound like something you want to work with us on? Dive into our SDKs and tools at nianticspatial.com or reach out to us here.
スクリーンから現実世界へ、AIを次のフェーズへ
本日より、Niantic Spatial Inc. は始動します。
この新しい会社の目標はただひとつ、現実世界を「理解する」世界最高のAI を構築し、デジタルと現実世界がインタラクトできるようにすることです。
短期的な目標は、倉庫やエンターテインメント施設のような特定の空間で、企業向けにサービスを提供することです。
長期的な目標は、ARグラスから自律システムに至るまで、あらゆるものを支える空間知能を持ったAIの基盤となる、「Large Geospatial Model（大規模地理空間モデル）」を構築・学習させることです。
-
Niantic Spatial プラットフォーム - わたしたちの地理空間サービスには、業界をリードするVisual Positioning System(VPS）、Gaussian Splattingによる高度な3Dスキャン技術とそのスキャンデータの可視化、そして開発者向けのSDK群が含まれています。
-
現在すでに、物流、不動産、製造、エンターテインメント、政府機関などの分野で、導入および実証実験が進行中です。
-
大規模地理空間モデル（Large Geospatial Model）- 世界中の何百万もの場所から集めた300億枚以上の空間情報付き画像 (posed image) をもとに構築されたこのモデルは、Large Language Models（大規模言語モデル）に空間的な推論能力を付加します。
-
これにより、単に「どこにあるか」だけでなく、「それが何であるか」や「周囲の環境との関係性」まで理解できるようになります
一緒に未来を作りませんか？
SDKやツールは nianticspatial.com で公開中です。ご連絡をご希望の方は、こちらからご連絡ください。