PearX, a prestigious 12-week accelerator program capped at 20 startups, revealed its newest cohort at a recent demo day in San Francisco. Venture capital interest centered on five standout companies pushing boundaries in robotics, AI hardware, privacy, financial planning, and industrial design.
- Spatial model startup Speridlabs offers persistent 3D geometry editing
- Saia develops low-power AI chip running directly from flash storage
- Ren delivers fully automated, privacy-focused AI personal assistant
What happened
PearX, an accelerator run by Pear VC focused on pre-seed and seed startups, held its semiannual demo day showcasing a selective group of 16 companies. Distinguished by a smaller cohort size and higher investment ceilings—up to $2 million—PearX maintains secrecy about its startups until the event, differentiating it from larger programs like Y Combinator. At the San Francisco event, five startups drew significant venture capital attention for their innovative technological approaches and market potential.
Among the highlights, Speridlabs presented a spatial foundational model dubbed Mundus, which offers persistent, modifiable 3D world models aimed at sectors such as robotics and gaming. Saia introduced a chip designed to bypass the traditional power-hungry memory bottlenecks found in Nvidia and Google AI hardware, promising enhanced speed and capacity with dramatically lower power consumption. Other notable startups included Ren, focused on a secure and fully automated AI assistant, Veros, delivering AI-driven trust and estate planning services, and Datum, which uses proprietary technology to expedite industrial design by indexing 3D part libraries.
Why it matters
PearX's approach to startup acceleration—with a tight cohort and substantial capital deployment—creates a powerful platform for emerging technologies to attract early-stage funding. The startups featured exemplify a trend toward integrating AI deeply into practical domains such as personalized security, physical product design, and financial services. Innovations like Saia's AI chip address crucial industry challenges in hardware efficiency and supply constraints, signaling potential shifts in how local AI inference might be deployed at scale.
Moreover, the demand for privacy-conscious AI solutions, exemplified by Ren's fully automated assistant, aligns with growing user concerns regarding data handling and human intervention. Veros's use of AI for trust and estate management reflects increasing adoption of automation within traditionally manual, lawyer- and advisor-heavy areas. Through these developments, PearX continues to set a distinct footprint in the startup ecosystem, providing a launchpad for breakthrough technologies with strong investor appeal.
What to watch next
Key developments to follow include the progress of Saia’s chip fabrication plans, especially its collaboration talks with Samsung and the goal to reach mass production by 2028. Success in this hardware segment could disrupt current market leaders and solve pressing GPU and TPU supply challenges for AI on-device processing. Speridlabs’ ability to commercialize its 3D model platform Mundus and compete with established companies like Runway and Google will also be crucial in validating this new paradigm of modifiable spatial AI.
Furthermore, the regulatory progress for Veros in securing a trust charter will impact its ability to fully operate as a regulated trust company, potentially transforming AI’s role in wealth management. For Ren, tracking user adoption and how its security-focused automation compares against human-operated services will reveal market appetite for privacy-first AI assistants. Overall, PearX’s upcoming cohort performances and follow-on funding rounds will shed more light on the viability and impact of these innovations in shifting tech landscapes.