A Washington, D.C.-based startup says it has built software that drastically reduces the time needed to deploy AI models on hardware, cutting what typically takes around 200 hours down to an automated process. Lola Vision Systems, launched in 2024 by Tayo Adesanya, creates what the company calls a "compiler toolchain" that converts AI models into instructions specific chips can execute. The firm is also developing its own semiconductor chips and has raised just over $1 million in total funding to date.
Manually configuring an AI model on new hardware currently requires roughly 200 hours just to start testing, according to Lola Vision Systems. The company reports that a dozen corporate customers have signed letters of interest to purchase its chips once they're ready, and it has already secured one signed customer. To generate revenue sooner, the startup plans to license its software for use on existing hardware rather than waiting for its own chips to become available. Lola Vision was selected for this year's TechCrunch Startup Battlefield 200, a program featuring 200 chosen startups, and has partnered with SCALE, a microelectronics workforce development program, to collaborate with additional semiconductor labs.
Adesanya told TechCrunch that the software addresses what he describes as a massive bottleneck in AI infrastructure. "Speed is only part of it," he explained, noting that faster setup allows aerospace and "other mission-critical companies" the opportunity to "run more accurate models on their own data, at a lower power." For these clients, he said, "accuracy and reliability aren't nice to have. They determine whether a product passes regulatory review and whether it works reliably in the field." The founder claimed that existing solutions like Nvidia's Jetson computing modules or open source AI models "often break or run poorly out of the box," forcing teams to spend days or weeks just getting them operational, followed by additional weeks of debugging before the models become usable.
Lola Vision positions itself as an alternative to Nvidia's technology for running AI on devices, a market known as edge computing where AI operates directly on hardware like cameras or drones rather than in remote data centers. According to Adesanya's account, even when existing solutions work, power consumption frequently exceeds edge computing budgets or the hardware can't deliver sufficient compute for medium to large models, resulting in recognition models that lag behind targets or misidentify objects. The company's approach allows clients to provide their code and chosen AI model—whether custom-built or open source—and the software translates both into instructions the client's chip can run. Adesanya's nearly 12 years working with microchips and AI processors, helping large manufacturers select chips for their hardware, gave him early visibility into where demand in the AI computing market was heading, he told TechCrunch.
The startup plans to continue developing its own semiconductor chips while licensing its software on existing hardware in the near term, aiming to automate more of the AI deployment process for mission-critical industries where regulatory approval and field reliability are non-negotiable. Adesanya said he applied to Startup Battlefield after about a year of product development and signing the company's first customer, viewing it as the right moment to reach a broader audience. He stated he's looking forward to "making meaningful connections" at TechCrunch Disrupt and, directly, to "investors writing checks." The licensing strategy reflects a pragmatic shift for hardware startups facing long development cycles, while the emphasis on regulated industries suggests the company is targeting customers who value reliability over cutting-edge performance.