It was 2003. The dot-com bubble had burst, leaving the tech industry bruised but not beaten. At Qualcomm’s San Diego headquarters, David Julian, fresh from completing his PhD in information theory at Stanford, joined a team led by Avneesh Agrawal, a fellow Stanford alumnus. Their mission: explore whether OFDMA (Orthogonal Frequency-Division Multiple Access) technology could be applied to cellular networks—a move that would challenge the company’s own dominant technology.
Until then, Qualcomm’s business heavily relied on royalties paid for patents on CDMA (Code Division Multiple Access) —the technology that allowed multiple cell phones to share the same frequency channel by assigning each a unique code. OFDMA represented a fundamentally different approach that divided radio spectrum into subchannels, offering greater spectral efficiency and data speeds.
“At that point, Qualcomm was deeply invested in CDMA, which they held most of the patents for,” Julian recalls. “Outwardly, everyone believed CDMA was the only way, but OFDMA was this emerging tech with a lot of potential.” The team developed a system that eventually became LTE, the foundation of 4G networks worldwide, with OFDMA as its downlink access method. The experience left both with a crucial insight: technological inflection points create enormous opportunities for those willing to challenge orthodoxy. It was a lesson they would apply more than a decade later when founding Netradyne.
Their paths later diverged after this project. Julian went on to explore brain-inspired computing and neural networks, chasing yet another technology before it went mainstream. Agrawal, however, continued rising through Qualcomm’s ranks. From leading Qualcomm’s 4G LTE research, standardisation, and chipset development from 2002 through 2009, to serving as president of Qualcomm India and South Asia, between 2011-2015.
In 2015, when Julian and Agrawal reconnected over coffee in San Diego, their discussion led to a realisation that three technologies had simultaneously reached maturity: cameras had become ubiquitous and inexpensive, edge computing was powerful enough to run complex algorithms locally on-device, and deep learning had advanced to handle real-world visual processing, — creating a commercial opportunity in its application. This was enough to compel Agrawal to quit Qualcomm.
The same year, they incorporated Netradyne, initially uncertain about which application to pursue. Autonomous driving was tempting, but crowded with well-funded competitors. Agricultural drones and retail analytics
also had potential. But it was driver safety—monitoring and coaching commercial drivers in real-time—that emerged as their focus.
Their initial system ingeniously combined an Nvidia Shield TV box with a Google Nexus 5X phone, creating a makeshift AI platform that could process video in real-time. To gather training data, they secured access to neighbourhood carpet cleaning trucks in San Diego. This expanded to several small fleets, providing invaluable real-world training data for their models without burning through capital.
After initial seed funding from Agrawal himself, they secured their Series A from Reliance Industries in 2016 for $16 million. Microsoft’s venture fund M12 and Point72 Ventures led their Series B of $21 million in September 2018, followed by additional B rounds bringing in Hyundai Cradle.
A breakthrough came in 2020 when Amazon deployed Netradyne’s cameras across its massive delivery fleet. In 2021, their Series C in 2021 brought in a massive $150 million from SoftBank Vision Fund, and their Series D in January 2025 added $90 million from investors including Point72 Ventures, Qualcomm Ventures, and Pavilion Capital Partners, valuing the firm at $2.25 billion.
Today, the company serves over 3,000 customers and has deployed 450,000 vehicles across eight countries, including marquee names such as Amazon, Shell, Indian Oil, and Greenline Mobility. It has collected over 29 billion km of road data that includes driver behaviour insights, traffic patterns, and road-hazard details. Their newest models incorporate specialised cameras that even detect pupil dilation patterns indicating driver fatigue.
Fresh from achieving unicorn status, the company is now targeting a new market: India’s ride-hailing sector, where accidents are often caused due to driver fatigue, and negligence. Its financial performance has been impressive. In FY23, Netradyne generated `1,000 crore ($116 million) in revenue — reaching this milestone in just nine years of operations. Beyond US, and India, they have expanded to Europe, Australia, New Zealand, with plans to enter markets in Malaysia, Canada, Germany, France, Spain, Japan, and West Asia.
Netradyne’s hardware is manufactured in Gurugram, while software development spans both Bengaluru and San Diego. “From the beginning, I have been in San Diego and Avneesh in Bengaluru, working mostly out of Zoom meetings. We were very clear that we want to utilise the talent, network, and training data available in both countries,” Julian says.
In an era of increasing technological nationalism, Netradyne’s cross-border success is a compelling reminder the most powerful innovations might still emerge from bridges rather than walls — especially when built by leaders who can recognise when technologies are mature enough to leap from theory into practice.
What began as a collaboration between two technologists at Qualcomm has evolved into a unicorn that demonstrates how innovation can thrive when it connects different technologies, problems, and parts of the world.