Aashay Sanghvi broke into venture capital directly from college as one of Haystack's earliest full-time investors and has grown up in the craft while helping build the firm from a solo GP fund to a scaled seed platform, giving him an unusual career path compared to VCs who typically come from operating, banking, or consulting backgrounds. He has built a distinct reputation for obsessively trading obscure research materials and surfacing very early technical teams before they appear on mainstream investor radars, while also publishing detailed long-form analysis on infrastructure and financial systems that founders circulate more like analyst notes than typical VC content. 

• Focuses on the convergence of software and the physical world, often referred to as physical AI. He argues that traditional software moats are eroding and that physical nodes like cameras, drones, robots, and satellites now provide the necessary scarce context. He views this as a cyclical return to hardware, noting that the venture capital industry effectively begins and ends with semiconductors.
• Actively seeks investments in cognitive security to combat the exponential rise of low-quality AI-generated content and the mental strain it causes. He has expressed a specific desire to fund more companies in this space, citing his firm's existing investment in Pangram Labs as a foundation for his interest in solving the problem of digital slop.
• Evaluates business durability based on the ownership of unique datasets and proprietary reasoning traces. He suggests that the most valuable companies will be those that integrate distinct data into their processes, comparing this evaluation method to running a long-short public equities fund based on who possesses the right data assets.
• Interested in specific financial infrastructure plays, particularly within prediction markets. He explicitly issued a request for startups building prime brokerage solutions for prediction markets, signaling a desire to fund the backend systems that enable these trading environments to mature.
• Views automation as a polarizing force that is eliminating the middle ground of work. He posits that roles will either become fully automated or remain extremely important, with little room in between. This thesis is reflected in his support for AI productivity tools, such as the calendar network Blockit, which aims to automate time management.