Tenstorrent Valuation, Funding & Investors

PsiQuantum is a quantum computing company that builds fault-tolerant, utility-scale quantum computers on photonic qubits. It was founded by Jeremy O'Brien, Terry Rudolph, Peter Shadbolt and Mark Thompson, who developed the approach as academics in the United Kingdom. The company manufactures its silicon photonic chips in a commercial semiconductor foundry, GlobalFoundries, rather than in a research fab, and designs the cryogenic cooling, networking and control systems needed to link very large numbers of qubits. It is headquartered in Palo Alto, California, and is building large quantum computing sites in Brisbane, Australia and Chicago, Illinois. Customers and partners are governments, national laboratories and large enterprises that want access to quantum machines for chemistry, materials science, drug discovery and other simulation work. The company earns money from government agreements and development contracts and expects to sell access to its systems once they are commercially useful.

Fractile is a London-based startup that designs specialized AI inference chips using an in-memory compute architecture built to run large AI models faster and more cheaply than conventional GPUs. Founded in 2022, the company targets up to 100 times faster model execution and 10 times lower costs, or 20 times better performance per watt, and has hired senior engineers from Nvidia, ARM, and Imagination. Fractile positions its hardware to compete with the GPU-based AI infrastructure of incumbents such as Nvidia.

Positron is a provider of specialized hardware accelerators for transformer model inference in generative AI applications. Its chips deliver high-throughput processing with reduced power usage compared to GPU alternatives. The platform supports every model in the Hugging Face Transformers library and exposes an OpenAI-compatible API endpoint. Developers deploy custom-trained models directly onto Positron hardware for edge and cloud deployments in natural language processing and computer vision tasks.

Lightmatter is a Boston-headquartered designer of photonic computing chips that accelerate matrix multiplications for AI inference and training using silicon photonics. The Passage interconnect links GPU clusters with optical switches, while Envise processors target data center workloads in cloud and telecom sectors.

Neuralink is a neurotechnology company founded in 2016 by Elon Musk and a group of scientists and engineers, focused on developing implantable brain–computer interface (BCI) devices. Its primary product, the N1 chip, is surgically implanted into the skull by a precision robotic system called R1, enabling direct communication between the human brain and external computers or devices. The company's stated mission is to address neurological conditions such as paralysis, blindness, and severe motor impairments, allowing patients to control digital devices using thought alone. In January 2024, Neuralink implanted its first BCI device in a human patient, with the individual subsequently demonstrating the ability to control a computer cursor through neural signals. Headquartered in Fremont, California, the company operates at the intersection of neuroscience, robotics, and semiconductor engineering, and has attracted significant attention for its long-term ambitions around human-AI symbiosis.

Core Scientific Inc is engaged in designing, building, and operating digital infrastructure for high-density colocation services and digital asset mining. The business operates in three operating segments: Digital Asset Self-Mining, consisting of performing digital asset mining for its own account, Digital Asset Hosted Mining, consisting of providing hosting services to third parties for digital asset mining, and Colocation, consisting of providing high-density colocation services to third parties for AI / HPC operations. The majority of revenue is derived from the Digital Asset Self-Mining Segment.

GTA Semiconductor is a Shanghai specialty-process semiconductor manufacturer founded in 2017 around Huada Semiconductor. It develops and runs wafer fabrication lines for automotive electronics, industrial control and high-end consumer electronics, producing microcontrollers, analog circuits, power devices and sensors for chip designers and integrated device makers. Its processes include BCD, IGBT, MOSFET and silicon carbide devices on 6-inch, 8-inch and 12-inch wafers, and it also offers mask, test and multi-project wafer services together with design kits. The company operates fabs in the Lingang area and in Xuhui, Shanghai. It took the Hong Kong-listed Shanghai Advanced Semiconductor Manufacturing Corporation private through a merger by absorption, which gave it a long-established analog foundry. It has financed expansion through state-backed and industry investor rounds and is privately held.

D-Wave Quantum Inc is in the development and delivery of quantum computing systems, software, and services, and it is the commercial supplier of quantum computers and the only company building both annealing quantum computers and gate-model quantum computers. It delivers customer value with practical quantum applications for problems as diverse as logistics, artificial intelligence, materials sciences, drug discovery, scheduling, cybersecurity, fault detection, and financial modeling. Its annealing quantum computers are accessible through the company's LeapTM cloud service.

Ricursive Intelligence is a frontier AI lab founded in December 2025 by former Google DeepMind researchers Dr. Anna Goldie and Dr. Azalia Mirhoseini that automates and accelerates semiconductor design through a self-improving recursive loop. By building platforms that use artificial intelligence to design next-generation silicon chips—which in turn provide the advanced compute needed to train more powerful AI models—the company aims to shift the semiconductor industry toward a "designless" paradigm that bypasses traditional hardware bottlenecks on the path to artificial superintelligence.

Etched is a privately held artificial-intelligence hardware company building chips and rack-scale systems for model inference. It co-designs custom accelerators, interconnects, memory systems, cooling, software and manufacturing processes for large transformer and mixture-of-experts workloads. The company's Sohu architecture is designed specifically for transformer inference rather than general-purpose graphics processing. Etched targets cloud providers, model developers and other enterprise customers that require high throughput, low latency and lower power consumption when serving artificial-intelligence models. Its operations include engineering and prototyping work in San Jose as well as manufacturing activity in Taiwan. The business has assembled a team with experience from semiconductor, cloud-computing and artificial-intelligence companies. Etched expects to earn revenue from sales of chips, accelerator cards, racks and associated computing systems rather than from consumer subscriptions or advertising.

SiFive designs customizable RISC-V processor cores and IP for embedded systems, IoT devices, automotive, and datacenter applications. Its Freedom platform includes microcontrollers, high-performance CPUs, and vector extensions with associated EDA toolchains. Santa Clara-headquartered SiFive licenses cores to Samsung, Western Digital, and Andes Technology. Founded in 2015, the company shipped over a billion RISC-V cores by 2023, powering chips from Qualcomm and NVIDIA partners.

SambaNova is a Palo Alto-headquartered enterprise AI company providing full-stack solutions from Reconfigurable Dataflow Unit silicon like the 2023 SN40L chip to software platforms including Samba Cloud for instant inference on models such as Llama and Qwen, SambaStack for on-premise deployments, and SambaManaged for efficient data center AI services deployable in 90 days.

Kunlun Chip operates as Baidu's dedicated AI semiconductor division. Beijing-based Kunlunxin designs processors such as Kunlun 1 for inference and Kunlun 2 for training, integrated into Baidu Cloud and PaddlePaddle deep learning platform. The chips deliver performance for computer vision, natural language processing, and recommendation systems in data centers.

Sunrise is a Shenzhen-based developer of high-performance GPUs for AI training, edge computing, and graphics processing. The company designs chips like the Sunhope series with up to 512 tensor cores, targeting data centers and autonomous vehicles. Manufacturing through TSMC processes, Sunrise supplies modules to Chinese hyperscalers and exports to Southeast Asia. Founded in 2018, it competes in the domestic semiconductor market amid US export restrictions.

FuriosaAI is a Seoul-based semiconductor firm designing AI accelerators for hyperscale data centers. The company launched the Warboy RTPU in 2023, delivering 40 tokens/second for Llama 70B at 75W TDP with 1,000 TOPS INT8 performance. FuriosaAI supplies chips to Naver Cloud and SK Telecom, scaling to clusters handling 10 PFLOPS for inference workloads.

Xsight Labs is a Tel Aviv-headquartered developer of chiplet-based processors for data center workloads including AI training and inference. The company designs scalable compute architectures using advanced packaging to boost performance per watt in cloud environments. Its Orion platform targets hyperscale data centers with disaggregated cores for machine learning acceleration. Founded in 2021, Xsight Labs partners with TSMC for fabrication and focuses on rack-scale systems deployable by 2025.

Shanghai Enflame Technology Co Ltd is engaged in the research, design and sales of cloud-based AI chips and related products.

IQM Quantum Computers Oyj is engaged in superconducting quantum computing and provides quantum computing systems and cloud platform access to enterprises, research institutions, high-performance computing centers, and national laboratories. Its offerings include quantum systems based on an open and modular architecture. The company has operations in Munich, Germany, and deployments across Europe, Asia, and North America.