Startups

April Robotics is pioneering human-to-robot skill transfer for dexterous manipulation. Our foundation models learn dexterity directly from human demonstrations and turn it into deployable robot skill, making intricate manufacturing work both achievable and affordable. We're on a mission to give manufacturers a workforce that scales on demand and takes on the delicate, labor-intensive tasks that have long held production back.

Robots can't handle hand manipulation for complex, dexterous assembly tasks — the bottleneck blocking humanoid deployment in manufacturing.

April Robotics is developing a foundation model for human-to-robot skill transfer for dexterous hand manipulation, using human demonstrations to help different five-fingered robots learn and perform complex tasks.

Modo is the AI success layer for the enterprise, spun out of Stanford University. Modo helps companies measure the ROI of AI from real workflow change, identify what's working across teams, and guide employees toward better AI practices and tools. Modo is backed by Pear VC, One Way Ventures, JFF Ventures, and investors and executives from Google, Meta, and other leading technology companies.

Enterprises are spending billions on AI licenses and tokens with no way to see the ROI, so leaders can't make informed decisions on AI spend or invest in best practices.

Modo gives CIOs, CAIOs, and CFOs visibility into AI ROI and spreads what works across the company. Modo has six-figure ARR committed and is partnering with some of the largest banks and services companies in the world.

Posting Machine helps B2B founders turn their day-to-day work and expertise into LinkedIn content that drives revenue. We capture real context from founders, turn it into authentic content in their voice, and help them build trust and generate sales opportunities with their target customers.

B2B founders know LinkedIn can drive sales, hiring, and partnerships, but consistently creating high quality content takes time they don't have. Our ICP is US based, LinkedIn savvy Seed to Series A founders. They already have a strong product and valuable company context, but lack the marketing bandwidth to turn it into consistent founder led content.

Posting Machine connects to a company’s existing context, including Slack, website content, and meeting notes, and turns it into on brand LinkedIn content and warm outreach. Three months after launch, we’ve grown from 2 to 12 paying customers and reached ~$3K MRR.

Peeps is a #1 app on the App Store that gives close friends and couples a shared virtual space—like Gather Town, but for personal relationships. Text messaging is flat and isolating. We solve this by letting users co-create a virtual home where they can hang out, share everyday moments, and interact playfully in real-time. Instead of just scrolling or texting, Peeps makes online connection feel like actually spending time together.

Traditional social media is optimized for content consumption and attention, not meaningful connection. As Gen Z gets busier and close friends increasingly live apart, messaging and passive feeds are not enough to help them feel present in each other's everyday lives.

Peeps is a gamified social app that helps close friends and couples stay connected through a shared virtual space. Users can complete daily photo quests, decorate a virtual home, raise pets, play together, and turn small everyday interactions into shared experiences. Peeps has grown to more than 200,000 users with zero paid marketing, reached #1 overall on Taiwan's App Store, and currently has approximately 25,000 DAU. Day-30 retention has reached 28%, while MRR grew 51.6% month-over-month. Peeps is backed by 886 Studios, founded by veterans behind Twitch and Guitar Hero.

ProjectX is changing the way humans interact with computers by making computing limitless, accessible and sustainable.

GPU resources are expensive and fragmented — local devices can't run large AI workloads; enterprise deployment costs are prohibitive.

Any device runs GPU-intensive applications and complex agentic tasks, via a single URL with cold start under 3 seconds and 95% GPU utilization. 35,000+ beta users and signups from engineers of more than 138 enterprises. Raised $4.6M seed.

AI adversarial testing platform for mental health.

Adversarial testing is required by regulatory bodies to surface failures in healthcare AI. Two testing methods exist: Human testing and automated persona testing. Human testing has high fidelity, but it is expensive; Automated persona testing is cheap, but has low fidelity.

Thanatos is an evaluation platform that provides high fidelity automated personas for adversarial testing. Our personas are fine-tuned from open and closed source first-person mental health data.

A creative collective that explores love through “understanding.” Love Chara Inc. designs emotions through AI-powered diagnostics, characters, and storytelling.

Brands need a more affordable way to reach Japanese Gen Z women, while building proprietary IP from scratch typically takes years. There is currently no scalable model that combines identity-driven engagement with character-based commerce.

Love diagnosis quiz funnels users into a character IP ecosystem spanning LINE retention, AI chat, merchandise, and B2B brand licensing. LC64 reached 70M activations and 35M unique users in its first month with zero ad spend, including 6M UU in Taiwan and 1.5M UU in Korea. Monthly revenue peaked at JPY 30M and is currently around JPY 18M, with a JPY 1.5M monthly burn rate. Lovetype ranked #1 in Japan’s Teen Buzzword Awards 2025. The company also has active partnerships with Bandai (250 crane machines), ABEMA, Seventeen, and KADOKAWA, while Alfred Lin, co-lead at Sequoia Capital, and SoftBank Singapore have both expressed investment interest.

Alpenglow Technology develops energy-efficient AI processors for the growing edge AI market, helping semiconductor and device companies bring more capable on-device intelligence to power-constrained products. We target applications including smart sensing, wearables, industrial IoT, and consumer electronics through flexible silicon and partnership models.

Small edge AI devices such as wearables, hearables, smart health devices, smart toys, and sensors face strict limitations in power consumption, battery life, heat, and device size. As more AI functions move from the cloud to end devices, traditional computing architectures consume significant energy in moving data between memory and computing units. This makes continuous, always-on AI difficult to achieve in small, power-constrained devices.

ANIX is Alpenglow Tek’s ultra-low-power AI architecture based on Analog Compute-in-Memory (CIM), designed for Micro Edge AI applications. By reducing energy-intensive data movement between memory and computing units, ANIX can achieve up to 100× higher energy efficiency for targeted AI workloads compared with traditional computing architectures. The company has completed its first test-chip tape-out using a 22nm process and is preparing for its next silicon verification in November 2026. Alpenglow Tek currently has three paid semiconductor customer engagements, including one global Tier-1 memory semiconductor company, and has generated approximately 2025 revenue USD 150K. The company is also working with multiple potential customers and partners on technical evaluation, PoC, and future product integration.