Skylark Labs in talks to raise around $150 million; eyes unicorn status

The fundraising comes as investors increasingly bet on physical AI, the convergence of artificial intelligence and robotics that enables autonomous machines to perceive, learn and perform real-world tasks. In the first quarter of 2026, investors poured a record $16.3 billion into robotics and physical AI startups across 492 deals, according to data platform PitchBook data.

“The next phase of AI is moving from software into machines that can perceive, reason and act in the physical world. Physical AI will emerge as the next major AI investment theme after generative AI,” said Siddharth Sawant, managing partner, Varyon Partners, a global strategy and transaction advisory firm focused on Robotics, Physical AI and Industrial technology. “The simultaneous build-out of enabling infrastructure, including data centres, edge compute and satellite communications, creates a favourable backdrop for defence and autonomy companies, strengthening long-term demand and investor interest.”

"During my research, I developed AI models that learn from limited data and run on edge devices, capabilities well suited for defence. That led me to start Skylark Labs," Singh told Mint.

“Our speciality is designing humanoids for mission-critical defence tasks, from handling to firing weapons. Unlike conventional systems that take months to train, our AI models can learn these skills in about 15 minutes. We are the only company in the world which can teach these skills with maybe 15 minutes of training, as opposed to other companies with 10 months of training,” Singh added.

Started in 2021, the company serves a mix of defence, government, and commercial customers across India and overseas. In India, it has worked with the Indian Army, Indian Navy, Indian Air Force, Border Security Force, Coast Guard, and companies such as ideaForge and Internationally, it counts the US Air Force, US Navy, Pepsi, Mercedes Benz and others.

In defence, the company claims, sovereignty is critical because the data used to train these systems comes from Indian defence and law enforcement agencies, along with their operational protocols.

“By developing the technology in-house, we ensure that classified data and operational processes remain within the country. Once deployed, the AI can continue learning at the user's end, allowing agencies to teach the system new tasks without sending that data back to us. That way, even the most sensitive operational knowledge never leaves their control,” added Singh.

Defence has become a structural investment theme following the Russia-Ukraine war, the recent West Asia conflicts and changing alliances. Governments are increasingly prioritising sovereign defence capabilities, said Sawant. "AI-enabled autonomy and greater control over critical technologies. Companies that can demonstrate operational value in defence are well positioned to secure long-duration, sticky contracts, although procurement cycles remain long,” he added.

Capital use

The fresh capital will primarily be deployed to expand Skylark Labs' workforce, strengthen its global footprint, and accelerate development of autonomous defence technologies, founder and CEO Singh said.

"We have around 50 people in Pune today and will hire another 100 with the latest funding. These hires will work across the entire AI pipeline, from collecting field data and annotating it to training machine learning models, engineering products and scaling sales," Singh said.

Skylark Labs will also expand its physical presence, opening another office in Pune, a new office in Navi Mumbai and scaling up its Delhi operations, expanding its existing offices in Menlo Park, Indiana and Texas, and plans to establish another office in Abu Dhabi.

The new funding will help broaden the robots' capabilities beyond handling firearms and metal detectors to tasks such as mine detection and operating a wider range of defence equipment. "Right now, our robots can fire two types of guns and handle metal detectors. We will expand them to mine detection and other defence applications," said Singh.

The company claims to have generated revenue of around $3.3 million in FY25 and is targeting approximately $11 million in FY26.

Skylark Labs competes with Physical Intelligence, a San Francisco-based startup developing foundation models that enable robots to perform a wide range of real-world tasks, and Skild AI, which is building a general-purpose AI model for robots to help them operate across different environments.

In the adjacent AI space, competitors include Ineffable Intelligence, a London-based startup that is developing reinforcement learning-based AI systems capable of continual learning, and Adaption Labs, a San Francisco-based startup that is building AI models that can learn and adapt in real time for large language models.