Bain Capital’s Ajay Agarwal bet on robots when Silicon Valley wouldn’t. Now he sees a ‘golden era’ for physical AI

Robotics and physical AI startups raised $27.6 billion across 1,009 deals in 2025, more than double the prior year, as Bain Capital Ventures investor Ajay Agarwal sees potential in lower component costs, stronger AI capabilities and experienced founders. Bain was the first investor in construction robotics startup TerraFirma in 2024 and reinvested in its $100 million Series A led by Kleiner Perkins in July. Agarwal’s earlier Kiva Systems investment returned 26.6 times Bain’s initial investment before Amazon acquired the company for $775 million.

Analysis

The investable signal is not the funding total; it is whether physical-AI systems move from impressive pilots to repeatable deployments with attractive payback. Falling component costs and better perception could expand the addressable task set, but they do not eliminate integration, uptime, safety, and site-customization costs. Those frictions may leave value with customers rather than startup hardware vendors: as more capital funds competing products, price competition can compress vendor economics even while adoption grows.

The likely public-market beneficiaries are suppliers of sensing, compute, motion control, and industrial automation; incumbents may also defend share through installed-base service and integration expertise. Conversely, a broad venture funding wave raises the risk of overcapacity and weak returns for later-stage investors. Talent moving from established technology firms is not evidence that those firms’ operating businesses benefit.

Near term, this is weak evidence for changing AMZN or GAP fundamentals; Kiva’s historical acquisition and customer relationships do not establish current incremental exposure. Over 1–3 months, look for disclosed customer deployments and repeat orders, not fundraising announcements. Over 6–18 months, the thesis strengthens only if deployments show measurable labor-hour savings, high uptime, and falling implementation costs. The contrarian risk is that investors extrapolate AI progress into reliable physical-world autonomy too quickly. Failure to convert pilots into scaled contracts, or evidence of persistently high service costs, would falsify the adoption thesis.

Key Decisions for Investors

• No trade in AMZN or GAP on this article alone: neither has a disclosed new deployment or quantified earnings impact here.
• Keep established automation and component suppliers on a watchlist rather than buying a broad robotics theme on venture funding momentum; require evidence of repeat orders and improving unit economics.
• For any private-market exposure, diligence customer concentration, deployment conversion, uptime, implementation expense, and cash burn. A large financing round is not proof of product-market fit.
• Revisit the thesis if companies disclose scaled deployments with quantified labor savings; downgrade it if pilots stall, service costs remain high, or customers defer automation spending.