3 ex-Apple researchers want to make AI less awkward. Lightspeed, Accel, and Nvidia are giving them $50 million to do it.
THE SO WHAT
A $50 million seed into ex-Apple researchers focused on emotional intelligence and conversation says investors think UX, not just scale, is now a frontier for foundational models. Teams building on top of LLMs should assume user expectations for nuance, tone, and “feel” will rise quickly over the next 12–24 months.
READ THE SOURCE
MORE FROM THE WIRE
Startups & VentureSatellite startup Open Cosmos raises €300m amid space frenzy
€300m into Open Cosmos is a bet that vertically integrated smallsat constellations — build plus operate plus data services — are now a European industrial asset, not just a startup story. If you rely on space-derived data, expect more regional options and pricing leverage, but also more fragmentation in standards and APIs.
Startups & VentureIntrepid Growth Partners closes a $525m AI fund with state money from four countries
$525m of state-backed capital across four countries going into AI is industrial policy via venture term sheets. Founders should expect LP governments to care where jobs, IP, and compute land — and operators should read this as more regionally anchored AI champions, not just another global SaaS wave.
Startups & VentureOpen Cosmos raises €300M to scale satellite manufacturing across four European factories
Scaling to four European satellite factories with €300m behind it turns Open Cosmos into infrastructure, not just a mission integrator. Anyone building on space data should expect shorter lead times and more tailored constellations — but also start thinking about vendor lock-in at the manufacturing and downlink layers.
Startups & VentureIntel-Backed Construction AI Firm Buildots Raises $130 Million
Buildots pulling in $130 million to speed up data centers, fabs, and hospitals is a bet that AI-native construction telemetry becomes standard for mega-capex projects. If you own large builds, treat site data capture and AI progress tracking as a core capability, not a nice-to-have vendor experiment.