Indian Chip Startup Bharat Semi Secures $24 Million Seed Round for Edge AI Sensors
The New Delhi-based fabless designer is developing programmable image sensors that process artificial intelligence directly on the chip. The all-equity round will fund the expansion of its engineering teams ahead of a planned 2027 commercial launch.
Share this story
New Delhi-based semiconductor startup Bharat Semi has closed a $24 million all-equity seed round to fund the development of image sensors that process artificial intelligence directly on the chip. The nearly five-year-old company is attempting to collapse the standard multi-component camera architecture into a single piece of silicon.[1][3]
While the startup remains pre-revenue and its hardware is still in the prototyping phase, it has already secured early design wins. United States defense technology contractor General Atomics and Indian contract manufacturer Kaynes Semicon have signed on to integrate the unreleased sensors into their hardware.[1]
The financing represents one of the largest early-stage bets in India’s emerging hardware sector. Info Edge Group, South Park Commons, and Indus Bridge Ventures led the round, with participation from Micromax, Valour Capital, and New York-based Gaingels.[3][4]
Bharat Semi will use the capital to expand its engineering footprint and push its first commercial products to market by the end of 2027. Co-founder Vrinda Kapoor stated the funding provides a two-and-a-half-year runway, by which point the company aims to be revenue-positive with active deployments.[3]
Collapsing the camera stack
Conventional machine vision systems rely on a fragmented architecture. A standard setup requires a discrete image sensor to capture light, an image signal processor to clean the visual data, and a separate system-on-chip to run the actual artificial intelligence inference.[1][5]
That separation forces data to travel between components, consuming power and introducing latency before a device can react to what it sees. Bharat Semi is designing programmable sensors that integrate capture, processing, and inference into a single package, moving the computing workload to the absolute edge.[4][5]
The company claims this consolidated architecture can reduce data movement between components by up to 99 percent. Because the sensor extracts feature points and makes decisions locally rather than streaming full video files to a remote processor, the overall power draw drops significantly.[4][5]
Kapoor estimates the integrated approach could cut total system costs by at least 70 percent compared to conventional multi-chip modules. However, those efficiency and cost metrics remain theoretical targets until the silicon is manufactured at scale and tested in live customer environments.[1]
"Our aim here is to move from sensors to sense-making solutions," Kapoor said in an interview following the funding announcement. The startup is betting that future growth in the sensor market will be driven entirely by machines capturing data for other machines, rather than for human viewing.[1]
The shift toward edge inference is becoming critical as the sheer volume of visual data overwhelms traditional networks. Morgan Stanley analysts project that approximately one billion humanoid robots will be operational by 2050, all requiring sensor suites that can perceive and act without waiting for cloud-based instructions.[4]
Currently, no widely adopted intelligent sensor suite exists that can meet those future power and latency requirements. Bharat Semi is attempting to define that missing category, positioning its silicon as the foundational sensor layer for the next generation of autonomous machines and vehicles.[4]
Target markets and timelines
The immediate applications for on-chip inference lie in devices constrained by battery life or network connectivity. Bharat Semi is targeting aerospace, industrial automation, surveillance, and smart home appliances, where sending continuous video feeds to a cloud server is either too slow or too expensive.[1]
Kapoor pointed to autonomous drones that need to process visual information without carrying heavy computing hardware, or air conditioners that adjust automatically when a room empties. In both cases, the device only needs the final decision, not the raw high-resolution image data.[1]
Despite the $24 million injection, Bharat Semi is still years away from volume production. The startup’s first-generation sensors are currently in the prototyping stage, while a more advanced second generation featuring heavier artificial intelligence capabilities is scheduled to reach prototype status by mid-2027.[1]
If that engineering timeline holds, the company expects to deliver the second-generation sensors to its initial customers by the end of 2027. The current funding is strictly allocated to completing this product development cycle and scaling the software stack that supports the hardware.[1]
The technical execution is being led by chief technology officer Boyd Fowler, who previously held the same title at major image-sensor manufacturer OmniVision. His involvement provides a layer of established industry credibility to a startup attempting to redefine standard sensor architecture.[1]
Expanding the geographic footprint
Bharat Semi currently employs roughly 75 people, split between its headquarters in India and a United States outpost. The American operations, based in Santa Clara, California, house 10 to 12 employees focused on core engineering and global business development.[1]
The new capital will finance a significant expansion of that California team. The startup plans to double its United States headcount to approximately 20 employees by mid-2027, while also opening recruitment for specialized engineering roles in Europe to support its international customer base.[1]
"It's a story of can we build a global technology company out of India," Kapoor said. The company was founded in late 2021 by Kapoor, a former healthcare technology executive and South Park Commons fellow, alongside managing director Vinayak Dalmia, a former McKinsey consultant.[1]
The startup operates strictly as a fabless designer, meaning it develops the intellectual property and software but outsources the actual physical manufacturing of the silicon to third-party foundries. This allows the company to avoid the billions of dollars in capital expenditure required to build a chip plant.[1]
In 2024, Bharat Semi did announce a separate initiative involving a technology partnership with the United States Space Force to explore a domestic semiconductor fabrication project. Kapoor clarified that the Space Force initiative remains in early development and will not draw any capital from this $24 million seed round.[1][5]
India’s shifting semiconductor landscape
The massive seed round reflects a broader shift in how private capital views the Indian semiconductor ecosystem. Historically, the country served as a back-office hub for global chip giants, providing outsourced design and verification services without owning the underlying intellectual property.[1]
That dynamic is changing rapidly as domestic startups move toward commercializing their own silicon. According to industry data, Indian semiconductor startups raised $92 million in the first five months of 2026 alone, surpassing the total capital raised across the entire sector in 2025.[2]
The surge is driven by early government design incentives maturing into actual products, alongside growing global investor interest in diversifying chip supply chains away from China. Venture capital firms are increasingly willing to fund hardware startups that can demonstrate a clear path to tape-out and customer adoption.[1]
The Indian government has committed billions of dollars to establish a domestic manufacturing base, though the country has yet to prove it can operate commercial-scale foundries. By remaining fabless, Bharat Semi isolates itself from those domestic manufacturing hurdles, relying instead on established global supply chains to print its designs.[1]
"We're all waiting for smarter AI that will improve our lives in the physical world," said Prateek Mehta, general partner at South Park Commons. "That world is perceived through sensors that work on fixed logic and will not scale in the future. They need to decide and react in the moment."[4]
For Bharat Semi, the challenge now shifts from raising capital and marketing architectural concepts to delivering working silicon that meets its aggressive efficiency claims. The semiconductor industry is littered with startups that promised revolutionary architectures but failed to achieve acceptable manufacturing yields.[1][4]
The semiconductor industry is littered with startups that promised revolutionary architectures but failed to achieve acceptable manufacturing yields.
The true test of the $24 million bet will arrive in late 2027. If the company can transition its prototypes into commercial volume production for General Atomics and Kaynes Semicon, it will prove that high-value semiconductor intellectual property can successfully emerge from the New Delhi ecosystem.[1]
Key points
- Bharat Semi raised $24 million in an all-equity seed round led by Info Edge Group, South Park Commons, and Indus Bridge Ventures.
- The startup is developing programmable image sensors that integrate capture, processing, and AI inference into a single chip.
- The company claims its architecture can reduce data movement by 99 percent and cut total system costs by 70 percent.
- While still in the prototyping phase, the sensors have secured early design wins with US defense contractor General Atomics and India's Kaynes Semicon.
Open questions
- Whether the theoretical 70 percent cost reduction will hold true once the sensors reach commercial volume production.
- Which specific third-party foundry Bharat Semi will use to manufacture the highly integrated silicon.
- Deeptech Investors
- Believe that physical AI applications require entirely new hardware architectures built at the sensor level.
- Fabless Chip Designers
- Argue that integrating processing directly on-chip drastically reduces latency and system costs for edge devices.
- Defense & Industrial Customers
- Seek autonomous perception capabilities that do not rely on continuous, vulnerable cloud connectivity.
Perspectives this story doesn't cover
- Traditional sensor manufacturers whose multi-chip architectures are being challenged.
- Third-party foundries tasked with manufacturing highly integrated edge AI silicon at commercial scale.
Sources
[1]TechCrunchFabless Chip DesignersIndian chip startup Bharat Semi lands $24M seed round as it eyes global AI sensor market
Read on TechCrunch →
[2]The Economic TimesFabless Chip DesignersBharat Semi raises $24 million seed round to build AI-powered smart sensors
Read on The Economic Times →
[3]MintDeeptech InvestorsBharat Semi raises $24 million in seed round led by Info Edge, South Park Common
Read on Mint →
[4]Business StandardDeeptech InvestorsBharat Semi raises $24 million to develop intelligent sensors for AI
Read on Business Standard →
[5]Indian Startup NewsDefense & Industrial CustomersBharat Semi raises $24 million in funding to build AI sensors for robots, cameras and drones
Read on Indian Startup News →
More in Technology
See all →AI Agents
LangChain Raises $300 Million at $3 Billion Valuation to Build Enterprise AI Agent Infrastructure
7 sources
AI Infrastructure
Mercor Raises $750 Million at $20 Billion Valuation to Expand Enterprise AI Training
3 sources
Defense Tech
Defense Startup Onodrim Raises $300 Million to Acquire and Consolidate European Military Suppliers
5 sources
Drone Logistics
Zipline Reaches $21.5 Billion Valuation to Roll Out Drone Delivery at 300 Walmart Sites
8 sources
Comments
Every angle. Every day.
Get Technology stories with full source coverage and perspective breakdowns, free every day.




