Mayfield Brings On Cerebras Backer Adit Singh as Infrastructure Partner
Adit Singh has joined Mayfield as an infrastructure partner, the firm announced on August 20, 2026. Singh arrives with a credential that carries real weight in the current funding climate: roughly ten years ago, while a partner at Foundation Capital, he helped source and co-lead the first funding round into Cerebras, a company building specialized chips for AI workloads. Cerebras went public in May 2026 and now trades at a valuation approaching $50 billion. Foundation Capital held approximately 7% at the time of the IPO, making it the third-largest shareholder (TechCrunch).
Singh is a chip designer by training. His electrical engineering background informed his early recognition of Cerebras's wafer-scale approach — building an entire AI processor on a single silicon wafer instead of splitting it into smaller chips — at a time when most investors were not yet focused on silicon bottlenecks. After Foundation Capital, he co-founded Neotribe Ventures in 2017, led the firm for four years, and then joined early-stage firm Cota Capital before his move to Mayfield (TechCrunch).
At Mayfield, Singh will invest across hardware, infrastructure software, cybersecurity, and physical AI — meaning AI systems that interact with the physical world, such as robotics or autonomous systems. That mandate aligns with an existing portfolio that already has meaningful silicon exposure. Mayfield's semiconductor holdings include Upscale AI, recently valued at $2 billion, and Lumilens, which raised $700 million at a $5.5 billion valuation (TechCrunch). The firm also backed SambaNova, which reached an $11 billion valuation after raising $1 billion in financing, and Velaura AI, which raised $110 million to power the physical AI era (Mayfield).
Mayfield managing partner Navin Chaddha has known Singh for 15 years and has sat alongside him on three startup boards, including Upscale AI and Velaura AI. That prior working relationship clearly shortened the diligence path. Mayfield holds $3 billion in assets under management and can write seed checks of up to $20 million, giving Singh the capital flexibility to lead rounds at the earliest stages where technical conviction matters most (TechCrunch).
The firm has been building out its AI and security bench deliberately. In March 2026, Mayfield announced that portfolio company Securiti was acquired for $1.725 billion, and that Exaforce raised a $125 million Series B for agentic security. Beyond direct investments, the firm operates several content series — "Founder Insights," "Humans in AI," and "AI Pathfinders" — and publishes a Saturday newsletter called "Weekend Edition" that distills signals from the AI landscape. NVIDIA CEO Jensen Huang was featured in the "Humans in AI" series discussing his founder journey (Mayfield). Mayfield also raised $525 million across two new funds, including Mayfield XV, a $400 million early-stage fund focused on enterprise and consumer technology companies (Mayfield).
The broader context here is that venture firms with partners who possess genuine hardware depth remain scarce relative to the volume of silicon and infrastructure startups now entering the market. The post-Cerebras IPO environment has made it clear that outsized returns in AI infrastructure can come from the silicon layer, not only from model developers or application companies. Firms that can evaluate wafer-scale architectures, photonics interconnects (which use light instead of electrical signals to move data between chips), and custom accelerator roadmaps at the seed stage hold an informational edge that is difficult to replicate through advisory networks alone.
Singh's move also reflects a pattern the industry has seen in prior compute transitions: the investors who identified the hardware thesis early tend to rotate between firms as their track records mature, and the firms that recruit them gain a concentrated advantage in the next funding cycle. Whether Mayfield's existing silicon portfolio benefits directly from Singh's network and technical judgment, or whether he builds a new cluster of investments from scratch, the firm has now added a partner whose single most notable deal is trading at nearly $50 billion.
For founders building at the infrastructure layer — whether in custom silicon, data center systems, or physical AI applications — the practical signal is straightforward. Mayfield now has a partner whose mandate spans the full infrastructure stack and whose check-writing capacity and firm resources are oriented toward early-stage, technically demanding bets.
