Semiconductors Startups

Groq designs and manufactures the LPU (Language Processing Unit), a chip purpose-built for AI inference, and operates GroqCloud, a platform for running inference workloads with low latency and cost efficiency across global data centers.

66 investors

Cerebras designs and manufactures wafer-scale AI processors and systems for data centers, enabling high-speed AI training and inference workloads. The company serves AI infrastructure providers, research institutions, and cloud platforms globally.

34 investors

PsiQuantum builds fault-tolerant quantum computers using photonic qubits and semiconductor manufacturing processes. The company develops quantum chips, cryogenic systems, and modular quantum computing platforms for commercial deployment across industries.

34 investors

Starcloud is a deep-tech company building large-scale data center infrastructure designed to operate in space. Its team comes from aerospace and cloud-hardware backgrounds including SpaceX, Amazon's Leo satellite constellation, Microsoft Azure, and fusion energy programs.

34 investors

Etched designs custom inference chips, racks, and software co-designed for running frontier AI models, including large sparse mixture-of-experts models, long-context, and agentic workloads. It builds vertically integrated inference systems aimed at higher throughput, lower latency, and better power efficiency for AI companies, cloud providers, and hyperscalers.

30 investors

Ayar Labs develops co-packaged optics (CPO) technology for AI infrastructure, enabling optical connectivity between thousands of GPUs across data center racks. The company manufactures optical engines (TeraPHY) and light sources (SuperNova) designed for hyperscale AI deployments requiring high bandwidth and low latency.

25 investors

Mojo Vision develops micro-LED technology, positioning it as a platform for AI infrastructure and next-generation displays. The company focuses on micro-LED components and applications for computing and display markets.

23 investors

SambaNova develops an integrated AI platform designed to help enterprises deploy and manage AI workloads efficiently. The company serves enterprises seeking to leverage foundation models and AI infrastructure for transforming their operations.

20 investors

d-Matrix develops AI compute hardware and platforms optimized for generative AI inference, designed to be faster, more energy-efficient, and more accessible than existing solutions.

19 investors

Akash Systems develops thermal management technology called Diamond Cooling, which integrates diamond with semiconductors to dissipate heat and improve efficiency in AI servers. Its products target AI compute, data centers, space, automotive/EV, cellular communications, and radar applications.

17 investors

Marvell designs custom silicon and provides interconnect and network switch products for AI data center infrastructure, serving hyperscalers, enterprises, and telecom carriers. Its portfolio spans custom compute, connectivity platforms, networking, and security for AI infrastructure, enterprise networks, and 5G applications.

16 investors

Extropic builds thermodynamic computing hardware, including thermodynamic sampling units (TSUs) designed to be more energy efficient than GPUs for probabilistic AI workloads. It offers the XTR-0 prototype platform and an open-source Python library for developing and simulating thermodynamic algorithms.

16 investors

Syntiant develops a full-stack edge AI platform combining neural decision processors, MEMS microphones and vibration sensors, and optimized deep learning models to run always-on, low-power intelligence directly on devices. Its technology targets speech, audio, sensor and vision applications across mobile, ear, IoT, and other power-constrained systems.

16 investors

Twist Bioscience develops a silicon-based synthetic DNA manufacturing platform that produces high-quality DNA, genes, and NGS reagents for customers in medicine, agriculture, and industrial chemicals. The platform enables faster and more affordable DNA synthesis at scale with customizable high-throughput capabilities.

16 investors

Tensordyne builds AI hardware, chips, and software systems optimized for efficient inference at reduced energy and computational cost. The company serves organizations seeking advanced AI capabilities with simplified deployment and lower operational overhead.

15 investors

Synopsys provides design automation tools and intellectual property for semiconductor and electronics companies to develop smart, connected devices. The company also operates a responsible business program focused on reducing environmental impact and enabling sustainable innovation across its customer ecosystem.

14 investors

Positron builds purpose-built hardware and inference accelerator silicon for running Transformer-based generative AI models, offering inference appliances that map HuggingFace Transformers models onto the hardware and expose an OpenAI API-compliant endpoint. Its products target AI infrastructure for high-performance, low-power model inference at scale.

14 investors

Conductor develops AI algorithms and control software for quantum computing, helping researchers and engineers optimize qubit architectures and scaling in semiconductor-based quantum systems.

13 investors

Decart builds infrastructure and real-time AI models for live, interactive experiences, optimizing the stack from hardware to live models. Its products include DOS and Cogito as a performance layer and Lucy and Oasis as real-time models for live video transformation and AI-generated game worlds serving consumer and enterprise applications.

13 investors

Graphcore builds processor hardware and systems for machine intelligence and AI workloads. Its products target machine learning and AI computing applications.

13 investors

Intel designs and manufactures semiconductor processors for data center, consumer, and embedded applications, including Xeon server processors, Arc graphics processors, and mobile processors. The company serves enterprises, device manufacturers, and consumers globally.

13 investors

Lightmatter builds photonic interconnect and compute technologies for AI infrastructure, enabling faster data transfer between processors in large AI clusters. The company serves hyperscale AI labs and data center operators by replacing electrical interconnects with photonic systems to solve bandwidth bottlenecks.

13 investors

Neurophos develops photonics and metamaterials-based compute hardware, combining optical systems with analog and digital silicon. Its team draws on backgrounds in AI silicon, memory, and optical systems.

13 investors

BO Materials develops novel materials for industrial applications across trillion-dollar industries. The company creates new material solutions to solve technical problems and reduce conventional trade-offs.

12 investors