The Venture Codex

Participated · Seed · Jul 2024

Kabata develops AI-powered adjustable dumbbells and an immersive workout platform designed to adapt to users’ progress. The all-in-one dumbbells adjust from 5 to 60 lbs, count reps, analyze form and deliver AI coaching that can change weights as needed. The devices use patent-pending, gaming-inspired haptics that send vibrations, pulses and rumbles to signal form corrections, set completions or personal bests. A velocity-based training feature monitors rep quality to inform users when fatigue may increase injury risk. Kabata raised $5 million in a seed round to further develop its AI-powered hardware and software. The company has attracted both institutional backers and athlete investors, indicating broad early interest in its product development plans. Kabata is a Los Angeles startup developing smart dumbbells that embed proprietary weight-system technology and progression algorithms. The product pairs the hardware with a connected app that creates six-week workouts, offering self-guided and trainer-led sessions. The system collects workout data to customize and continue adapting the user experience, and the company says the hardware can coach correct form and track comprehensive workout metrics. Kabata is currently piloting the product with no specific release date. It is targeting both home users and gyms while it figures out the right pricing model. Co-founder and CEO Kareem Aly is quoted describing the offering as an innovation in a category that has seen little change since 1928.

Participated · Series A · Oct 2023

Founded in 2019 by Swupnil Sahai and Richard Hsu, SwingVision uses patent-pending object-tracking AI to automatically track ball trajectory in 3D from a single device and generate pro-quality stats, highlights, and line calling. Its core product is a video-based analytics and officiating platform for tennis and similar racket sports. The system is used by more than 100 Division I college tennis teams across the U.S., including Harvard, Stanford, TCU, Florida and Yale. SwingVision plans to build tournament-grade officiating and remote coaching capabilities and to expand beyond tennis into pickleball and padel. The company recently raised $6M in a Series A to support those initiatives. SwingVision is a smartphone app that uses computer vision to deliver real-time shot tracking, automated line-calling/umpiring, and video analysis for tennis players. It tracks stroke type, spin, ball speed, shot placement, contact, rally length, footwork and posture, and builds player profiles and highlight reels to help users analyse shot patterns and scout opponents. The product was featured in Apple’s 2021 keynote and leverages recent smartphone ML advances (e.g., A15 Bionic) to enable on-device analytics. Subscriptions are growing 18% month-on-month and 70% of subscribers have been acquired organically, and the company is scaling via partnerships with clubs and federations. Partnerships include Tennis Australia installing 2,500 court mounts, access for 3,000 coaches, use at the Australian Open, and official ball-tracking relationships with the UK Lawn Tennis Association and the US Intercollegiate Tennis Association. Near-term plans focus on refining go-to-market via more partnerships, pre-installing mounts to reduce user friction, and key hires; technical challenges include battery drain and current incompatibility with less powerful phones and Android devices.