The appointment of Adit Singh as an infrastructure partner at Mayfield highlights a growing shift in venture capital toward the physical and technical foundations required to support the next generation of artificial intelligence.

Singh joins the 57-year-old venture firm after building a reputation as an early-stage technology investor with a strong semiconductor background. His investment career includes helping source and co-lead Foundation Capital's first funding round in Cerebras, a then-little-known AI chip startup that has since grown into one of the industry's most valuable semiconductor companies.

Foundation Capital ultimately became a major Cerebras shareholder. By the company's IPO in May, the firm held roughly 7% of the company, making it its third-largest shareholder. Cerebras now carries a valuation approaching $50 billion, illustrating the potential returns available when investors identify critical infrastructure companies at an early stage.

At Mayfield, Singh will focus on four interconnected areas: hardware, infrastructure software, cybersecurity and physical AI. The mandate reflects the increasing overlap between these markets as AI moves beyond cloud-based software into chips, machines, robotics and other physical systems.

Singh's semiconductor expertise is central to the appointment. Trained as a chip designer, he approaches technology from the application layer down to the underlying transistor level. That perspective can give an investor an advantage when evaluating companies whose competitive differentiation depends on architecture, power efficiency, computing performance or specialized hardware.

The importance of this expertise is increasing as the AI industry enters an infrastructure-intensive phase. The rapid expansion of AI applications has created demand for specialized processors, advanced computing systems, networking technologies, data-center infrastructure and software capable of efficiently managing increasingly complex workloads.

This environment is also changing the economics of venture capital. AI infrastructure companies can require substantially more capital and longer development cycles than conventional software startups, particularly when they are building semiconductor products or physical systems. Investors therefore need technical expertise capable of distinguishing genuine architectural advantages from incremental improvements.

Mayfield's existing semiconductor portfolio was reportedly a major factor in Singh's decision to join the firm. Its investments include Upscale AI, which was recently valued at $2 billion, and Lumilens, which raised $700 million at a $5.5 billion valuation.

The appointment also fits within Mayfield's broader AI investment strategy. The firm has committed $3 billion to AI and is increasing its emphasis on the foundational technologies behind AI systems. Its portfolio includes companies such as NUVIA, which was acquired by Qualcomm, alongside AI infrastructure businesses including Lumilens, SambaNova Systems, Frore Systems and Upscale AI.

For Mayfield, adding Singh strengthens its ability to identify technical founders at the earliest stages of company formation. The firm says approximately 70% of its investments begin at inception, positioning it to participate before companies have reached the scale that attracts larger growth investors.

Singh's own career reinforces that early-stage orientation. After his work at Foundation Capital, he co-founded Neotribe Ventures in 2017 and spent four years leading the firm before joining Cota Capital. His experience across multiple venture platforms gives him exposure to different approaches to identifying and supporting technical startups.

The Cerebras investment is particularly relevant to his new role because it illustrates the value of understanding technological shifts before they become obvious to the broader market. A decade ago, specialized AI computing was still a relatively narrow investment thesis. Today, AI accelerators and computing infrastructure have become central to one of the world's largest technology investment cycles.

The next opportunity may lie in identifying what comes after the current generation of AI infrastructure. As model development expands, investors are increasingly looking at technologies that improve efficiency, reduce energy consumption, accelerate inference, enable autonomous machines and make AI systems more practical outside traditional data centers.

Physical AI is especially significant in this context. The combination of AI with robotics, industrial equipment and autonomous systems requires a different infrastructure stack from purely digital applications. It involves sensors, processors, control systems, software and security working together in real-world environments.

Cybersecurity is similarly becoming inseparable from AI infrastructure. As AI systems gain access to sensitive corporate information and increasingly control automated processes, vulnerabilities in the underlying hardware and software can carry consequences beyond conventional data breaches.

For Mayfield, these interconnected markets create an investment thesis that extends beyond funding individual AI applications. The strategy focuses on the companies building the systems that allow AI to operate at greater scale, speed and reliability.

The move also reflects a broader maturation of the AI investment market. During the initial wave of generative AI, attention was heavily concentrated on model developers and consumer-facing applications. Capital is increasingly moving toward the less visible infrastructure that determines how efficiently those applications can operate.

Singh's appointment therefore represents more than a personnel change. It strengthens Mayfield's technical investment capabilities at a time when semiconductor design, AI computing, physical automation and cybersecurity are becoming increasingly interconnected. His track record with Cerebras gives the firm an investor who has previously identified a major infrastructure opportunity before it reached mainstream recognition.

The strategic question for Mayfield will now be whether it can repeat that pattern across the next generation of infrastructure companies. If AI continues moving into physical environments and enterprise systems, the firms capable of identifying foundational technologies early could capture significant value before those markets become crowded.

The appointment ultimately underscores a central development in the technology economy: the future of AI will depend not only on better models, but on the chips, systems, software and physical infrastructure that allow those models to function at scale.