New Delhi chip startup Bharat Semi has raised a $24 million seed round to build image sensors with AI built in, TechCrunch AI reports. The company is almost five years old and has no revenue yet. It says it has design wins with U.S. defense tech company General Atomics and India’s Kaynes Semicon. Commercial production is planned for 2027.
⚡ Key Points
- The raise: $24M seed round, all equity. Info Edge Group, South Park Commons and Indus Bridge Ventures led it. Valour Capital, Micromax, Gaingels and individual investors also joined.
- The product: Programmable image sensors that capture images, process them and run AI on a single chip.
- The pitch: The company says it can cut system costs by at least 70% compared with standard setups.
- The customers: Design wins with General Atomics and Kaynes Semicon. Neither has turned into revenue yet.
- The timeline: First-generation sensors are being prototyped now. Second-generation prototypes are expected by mid-2027, with customer deliveries by the end of 2027.
🔍 What Bharat Semi Is Building
Most camera-based systems today use three separate chips: an image sensor, an image signal processor (ISP), and a system-on-chip (SoC) that does the actual computing. Raw visual data has to move between all three. That adds cost, power use and delay.
Bharat Semi wants to put all three jobs on one piece of silicon. “Our aim here is to move from sensors to sense-making solutions,” co-founder Vrinda Kapoor told TechCrunch.
Kapoor gave two examples. One is an air conditioner that can tell a room is empty and adjust itself. The other is a drone that processes what it sees without expensive computing hardware elsewhere on board. The company is targeting smart homes, industrial automation, surveillance and aerospace.
🧠 Why This Matters
What stands out to me is the direction. A lot of AI hardware money goes to huge data center accelerators. Bharat Semi is going the other way, toward cheap and power-efficient intelligence at the edge, right where the data gets captured.
This matters because AI inference is moving into everyday devices. If an air conditioner or an industrial camera can make decisions on a sensor that costs a few dollars, you don’t need a cloud connection or a strong processor for basic perception. That’s good for privacy, delay and cost all at once.
The team has relevant experience too. CTO Boyd Fowler held the same job at OmniVision, one of the biggest names in image sensors. That gives the technical roadmap real credibility.
🌏 The India Angle
This deal also fits a bigger story. India has committed billions of dollars to building a domestic chip industry. It’s betting that U.S. companies will want supply chains outside China. So far, the country hasn’t become a major manufacturing hub.
Bharat Semi shows a different route: design in India, sell to the world, and leave the manufacturing to others. The company plans to stay fabless, meaning it outsources fabrication and focuses on chip design and software.
“It’s a story of can we build a global technology company out of India,” Kapoor said.
Here’s how the company’s footprint looks:
- About 75 employees across India and the U.S.
- 10 to 12 people in Santa Clara working on engineering and business development
- Plans to grow U.S. headcount to around 20 by mid-2027
- Hiring engineers in Europe has already started
A side note: in 2024, Bharat Semi announced a separate fab project in India with a technology partnership involving the U.S. Space Force. Kapoor says it’s still in development and the company is looking for domestic partners to build and run it. None of the new money goes toward it.
⚠️ What to Watch
Treat this as an early bet, not a done deal. Some realities to keep in mind:
- Pre-revenue risk: The sensors are still in development, and design wins aren’t the same as purchase orders.
- Tough incumbents: Sony, Samsung and OmniVision dominate image sensors and are adding more on-chip processing of their own.
- 2027 is a long way off: Hardware schedules slip often, and second-generation AI features won’t reach customers for at least a year.
If you build edge AI products in robotics, drones, smart appliances or industrial vision, this is a supplier worth tracking. Single-chip sensors that “see and think” could make your bill of materials much simpler.
The real test comes in 2027, when Bharat Semi has to show these chips can ship at scale. Full details are in the original TechCrunch AI report.