Hardware
CAPITAL FIGURES ARE MEDIA-EXTRACTED ESTIMATES, NOT VERIFIED FILINGS.
EXTRACTED FROM 25+ PODCASTS & VC NEWSLETTERS · MEDIA-REPORTED FIGURES, NOT VERIFIED FILINGS
Market Context
Hardware is experiencing a renaissance driven by two converging forces: AI infrastructure demand is decommoditizing formerly low-margin supply chain players, while a new wave of consumer and developer hardware is emerging around AI-native form factors. The $7.28B deployed across 24 deals in 28 days signals institutional conviction that the silicon, cooling, PCB design, and wearables layers of the AI stack are as investable as software. Investors including Nvidia, Samsung, and Amazon are making direct strategic bets, reflecting a shift from passive chip consumption to vertical integration across the hardware stack.
Investment Activity
- Tenstorrent raised a $1.4B Series C led by Tether, Qualcomm, Amazon, Nvidia, and Bosch, valuing the inference chipmaker at an undisclosed figure as takeover talks with Intel and Qualcomm run in parallel.
- A $520.5M growth round at an $11.6B valuation was secured with backing from Qatar Investment Authority, Nokia, and Solidium.
- TDK agreed to acquire Fabric8Labs for up to $400M, marking a strategic move into advanced hardware manufacturing.
- ZincFive was acquired by Spark I Acquisition Corp. for $600M, bringing energy storage hardware into the public markets.
- Fractile, the custom LLM inference chip company, closed a Series B led by Felicis Ventures.
- A $300M Series C at a $2.4B valuation was raised with backing from Nvidia, Siemens, Applied Materials, General Catalyst, and Atomico.
- A $85M Series A was raised with participation from Samsung, LVMH, CRV, and Inditex.
Key Players
- Tenstorrent: Santa Clara-based inference chipmaker that has now raised over $1.8B in VC and is simultaneously fielding acquisition interest from both Intel and Qualcomm, making it the highest-stakes hardware M&A target in the current cycle.
- Celestica: Identified as the sole supplier of Google TPU servers, holding 50–60% share of cloud Ethernet switching and unique liquid cooling capabilities — a former commodity manufacturer structurally decommoditized by AI workloads.
- Diode Computers: Brooklyn-based AI-native PCB compiler startup whose open-source toolchain, unveiled in June 2026, uses a code-first approach with AI agents to compress circuit board design timelines from months to weeks.
- Fabric8Labs: Advanced hardware manufacturing target acquired by TDK for up to $400M, signaling strategic consolidation in precision manufacturing for AI-era hardware.
- Monako: Launched a 48g waveguide wearable enabling developers to run AI coding agents hands-free, garnering 131 Product Hunt votes and signaling a new form factor for AI-human interaction.
Market Signals
- Nvidia is the most active investor in hardware deals over the past 28 days with 4 deal appearances, underscoring its strategy to control the full AI compute stack from chips to systems.
- Samsung has participated in 3 deals, including the $85M Series A alongside LVMH and CRV, reflecting consumer and enterprise hardware diversification beyond its own product lines.
- General Catalyst appears in 3 deals, signaling top-tier VC conviction in hardware at scale rather than treating it as a niche vertical.
- The PCB design and manufacturing layer is seeing a cluster of activity: Diode Computers, Quilter, and Diode (Brooklyn) each represent AI-native approaches to compressing hardware development timelines, forming a distinct sub-theme.
- Formerly commodity hardware supply chain players — Elite Materials (copper clad laminates, 50–60% CAGR in units), Corning (fiber optic share in AI data centers), and Celestica — are being re-rated as AI infrastructure beneficiaries.
- India-based Tiea Connectors (Dharwad) and Amsterdam-based Schematik (backed by $4.6M from Lightspeed) point to geographic diversification of the hardware startup ecosystem beyond the US and China.
- The wearables and developer hardware layer is accelerating, with Monako and Mira both positioning AI-native glasses as a post-smartphone computing interface.