Ampere Computing

Modal Labs, founded in 2021 in New York, provides cloud-based infrastructure for developers and data teams building generative AI applications.

Series D+ · $1.2B · 2026-10 · Valuation $16B

Starcloud, founded in 2024 by Philip Johnston, Ezra Feilden and Adi Oltean (formerly Lumen Orbit) and based in Redmond, Washington, is a Y Combinator S24 company building AI data centres in low Earth orbit, powered by continuous solar energy and cooled by the vacuum of space. In November 2025 it launched Starcloud-1 carrying an NVIDIA H100 and demonstrated Google Gemini inference and fine-tuning in orbit. Its stated end state is an 88,000-satellite constellation delivering roughly 20 gigawatts of orbital compute.

Series A · $450M · 2026-08 · Valuation $2.3B

D-Matrix, founded in 2019 in Santa Clara, California, USA, develops AI inference chips and systems for data centers.

Series C · $450M · 2025-11 · Valuation $2.0B

Shanghai-based Dongfang Suanxin (founded May 2024, HQ in Zhangjiang, 500+ staff across seven cities) develops large-compute AI chips built on its proprietary software-defined chip and 3D near-memory computing technologies and a fully domestic supply chain, with products spanning chips, system and application software, servers and compute clusters. It launched its flagship DF1000 chip in July 2026.

Series A · 2026-07 · Valuation $1.7B

KONST (KONST Co., LIMITED) is a Taiwan-based AI compute operator running three brands: Konstra AI (GPU data center build-out and hosting), Glows.ai (per-second billed enterprise GPU cloud) and Horizon AI (enterprise AI adoption and model governance). It operates four data centers in Taiwan and is expanding to Indonesia and Malaysia.

Series B · 2026-10

Kunlun Chip is a Baidu subsidiary developing general-purpose cloud AI chips based on XPU/GPU architecture for deep learning and machine learning workloads at cloud and edge scale. Planned products include the M100 for large-scale inference in 2026 and the M300 for hyperscale multimodal training in 2027.

SWI Group (SWI Capital Holding, Euronext Amsterdam: SWICH), formed in March 2025 from the merger of Icona Capital and Stoneweg, converts bitcoin mining sites into AI data centers and HPC facilities and offers GPU compute, with 4GW+ of capacity in its pipeline, alongside real estate and credit businesses.

$1.2B

Graphcore, founded in 2016 in Bristol, UK, designs a novel processor called the Intelligence Processing Unit (IPU), architected from the ground up for AI workloads, positioning it as a more efficient alternative to GPUs for large-scale parallel processing and complex machine learning models. Its customers include JPMorgan, Microsoft, Citadel, and universities such as Imperial College London, UC Berkeley, Oxford, and Stanford.

$682M

Verda, founded in 2020 in Helsinki, Finland, primarily provides AI cloud infrastructure and GPU computing resources for developers and enterprises.

Series B · $450M · 2026-09 · Valuation $1.0B

Founded in 2020 in Seongnam, South Korea, Rebellions designs energy-efficient AI inference chips.

Pre-IPO · $898M · 2026-03 · Valuation $2.3B

Tenstorrent, founded in 2015 in Toronto, Canada, develops advanced processors and software to enhance artificial intelligence and machine learning capabilities.

Undisclosed · $793M · 2024-12

Bitdeer Technologies Group is a Singapore-headquartered Bitcoin mining and high-performance computing infrastructure company, spun off from Bitmain in 2021 by Jihan Wu and listed on Nasdaq (BTDR) in April 2023 via a SPAC merger with Blue Safari Group Acquisition Corp. It designs SEALMINER mining rigs and operates large data centers in the U.S., Norway and Bhutan, and since 2023 has been converting mining sites into NVIDIA GPU data centers and selling AI compute through Bitdeer AI Cloud.