Andreessen Horowitz general partner Anu Midha is launching a new venture focused on building foundational computing infrastructure, according to sources familiar with the matter. The move marks a significant step in Midha’s career and underscores growing investor interest in the next wave of compute technologies needed to power advanced artificial intelligence systems. Midha, who has been a key figure at the influential venture capital firm since joining in 2015, has long been focused on the intersection of AI and infrastructure. Her new company will aim to develop high-performance computing solutions tailored for the demands of large-scale AI training and inference workloads. This includes innovations in hardware architecture, software stacks, and system-level optimization to improve efficiency and reduce costs. The venture is expected to target the growing need for specialized compute capacity as AI models become more complex and data-intensive. With major tech companies and startups alike racing to scale their AI capabilities, demand for faster, more efficient computing platforms has surged. Midha’s new firm is positioned to address this gap by creating infrastructure that can support the next generation of AI applications. While details about funding and specific technology plans remain under wraps, sources say the company will likely focus on building systems that go beyond traditional cloud computing. This could involve custom silicon, novel memory architectures, or distributed computing frameworks designed to handle AI workloads more effectively than current solutions. Midha’s departure from Andreessen Horowitz signals a shift in her professional journey from investor to founder. She has played a pivotal role in the firm’s AI and enterprise technology investments, backing companies like Anthropic, Databricks, and Stripe. Her new venture reflects a broader trend of top-tier venture capitalists stepping away from their firms to launch startups in areas they believe are critical to the future of technology. The launch comes at a time when compute is emerging as a central bottleneck in AI development. As models grow larger and more resource-intensive, access to scalable, efficient infrastructure is becoming a competitive advantage. Midha’s new company could play a key role in shaping how AI systems are built and deployed in the years ahead.