Baharat Semi raises $24 million for smart vision sensors: hardware cost equation ahead of regional test
Listen to this article
Read by Anchor
Indian startup Baharat Semi raised $24 million in a seed financing round dedicated to developing AI-enhanced image sensors, a technical step aimed at cutting optical hardware costs by 70 percent, a shift that aligns directly with the Gulf and Egypt’s smart-city and infrastructure plans that allocate large budgets for field monitoring and automation systems.
The company, based in New Delhi, focuses on programmable image sensors that integrate image capture, processing and AI capabilities onto a single chip, eliminating the need to send visual data to separate processors. According to the participating institution’s spokesperson Freeda Kapoor, this integration reduces system cost by at least 70 percent compared with the traditional architecture that separates the image sensor, image-signal processor and the system-on-chip, moving hardware from mere sensing tools to solutions that extract meaning and make decisions at the edge of the chip.
Merging vision processing and AI on one chip removes the need for costly intermediate computing cards.The company’s design targets smart-home, industrial automation, surveillance, and aerospace and aviation technologies. Kapoor cited practical examples such as air-conditioners that can detect empty rooms and adjust operation automatically, and drones that can process visual information on-board without relying on additional expensive computing hardware, thereby reducing weight, power consumption and total operating cost.
Although the firm, founded in late 2021 and not yet generating commercial revenue, aims to start commercial production in 2027, it has already secured design wins with U.S. defence-technology giant General Atomics and Indian company Kines Semicon. It is currently manufacturing prototypes of its first-generation sensors, while planning to bring the second generation, featuring more advanced AI capabilities, to prototype stage by mid-2027, in preparation for delivery to customers before the end of that year.
The financing round consisted entirely of equity led by Info Edge, South Park Commons and IndoBridge Ventures.Valour Capital, Micromax and Genghills also participated in the round alongside individual investors. The company plans to use the liquidity to expand its engineering team, complete product development and launch the sensors commercially. It currently employs about 75 staff across India and the United States, including 10 to 12 employees in Silicon Valley’s Santa Clara who lead engineering and business development, with a plan to increase the U.S. headcount to 20 by mid-2027 and to begin recruiting engineers from Europe.
The firm is led by a team with diverse technical and execution experience: Freeda Kapoor, a South Park Commons alum who previously worked in health-tech, serves as co-founder, alongside chief operating officer Vinayak Dalmiya, who comes from a consulting background at McKinsey and private-equity, while Bob Fuller holds the role of executive technology director after previously occupying the same position at OmniVision, a leading image-sensor maker. Baharat Semi follows a fab-less business model, focusing on chip and software design while outsourcing manufacturing, keeping its separate plans announced in 2024 to build a semiconductor manufacturing complex in India in partnership with the developing U.S. space-force initiative and local partners, without relying on this round’s financing.
The move comes as India pours billions of dollars into building a domestic semiconductor industry, seeking to cement its position as an alternative manufacturing hub amid U.S. companies’ efforts to diversify supply chains away from China, which gives startups a window to design specialised chips that meet global market, defence and automation needs.
What this means for the region
A 70 percent reduction in the cost of computer-vision systems represents a decisive turning point for infrastructure and digital-transformation managers in the Gulf and Egypt. Municipal surveillance projects, crowd-management and security systems in critical facilities currently rely on extensive camera networks linked to expensive edge-processing units or cloud servers that demand costly bandwidth. The maturation of AI-integrated sensors forces procurement and technical teams to recalculate economic viability and shift from centralized video processing to instantaneous inference at the sensor level with the lowest possible hardware cost.
The entry of Indian chip-design firms, backed by U.S. defence contracts, creates a practical path to diversify smart-hardware supply chains for regional organisations. Instead of relying solely on traditional suppliers in East Asia, system-integrators in Riyadh, Abu Dhabi and Cairo can track an emerging Indian supply route that meets international partnership standards and cyber-security expectations.
This shift compels embedded-system developers and robotics engineers in the region to upgrade their skills for handling directly programmable sensors. It is no longer sufficient to connect ready-made camera modules to external controllers or call cloud APIs; there is a urgent need to build talent capable of configuring vision algorithms and tuning them inside the micro-chip itself, ensuring instant response and protecting sensitive data privacy on field sites.