Ten1010
Founded 2020 · 22 employees on LinkedIn · 2 known investors
TEN builds AIPub, an enterprise AI infrastructure platform that uses spatial GPU partitioning and dynamic resource allocation to raise GPU utilization, plus RA:X workload benchmarking consulting, managed AI cloud, and modular AI data centers. It serves organizations running AI/ML and LLM workloads that want to maximize existing GPU resources and lower infrastructure TCO.
Also known as TEN · TEN Inc.
Investors · 2
Also in the syndicate · 1
Company profile
researched Aug 2026TEN (TEN Inc., operating at ten1010.io) is a Seoul-based software company that builds GPU resource optimization products for enterprise AI infrastructure. Its core platform, AIPub Enterprise, is an orchestration and integrated operations layer for AI infrastructure that applies spatial partitioning of GPU memory and cores rather than time-slicing, so multiple workloads can share a single GPU with independent resource allocation and minimal mutual interference. On top of the partitioning layer, the company offers dynamic resource allocation that redistributes GPUs in real time to match workload demand and automatically reclaims idle capacity, an in-house scheduler for workload prioritization and allocation control, integrated monitoring with per-GPU power control, and metering/FinOps functions that attribute GPU consumption by user, team and project.
Alongside the platform, TEN sells RA:X, a consulting and benchmarking service that sizes AI infrastructure from measured workload behavior in order to determine required GPU performance and capacity before purchase. Two further products are listed as forthcoming: AIPub Cloud, a managed AI cloud bundling GPU infrastructure and operations, and a modular data center offering for rapid AI-dedicated buildouts. Earlier descriptions of the company characterize AIPub as an MLOps platform supporting collaboration among AI modelers, IT engineers and administrators, with container and service management, infrastructure provisioning and service monitoring. The company also maintains a public GitHub organization (ten1010-io) with Kubernetes installers, a project controller, and an SR-IOV/InfiniBand virtualization management tool.
Business model
TEN sells enterprise software and services: a licensed on-premises AI infrastructure orchestration platform (AIPub Enterprise), workload benchmarking and infrastructure sizing consulting (RA:X), and planned managed cloud and modular data center offerings.
Revenue derives from enterprise platform adoption and consulting engagements, with planned managed AI cloud services and modular data center deployments.
Traction
Referenced deployments include a large manufacturer that raised GPU utilization from 10% to over 80% after consolidating distributed GPUs, a public research institution that consolidated more than 60 nodes, a major domestic university implementing multi-tenant GPU sharing with usage-based management, a second large manufacturer running dozens of ML and LLM inference services in one environment, a third manufacturer unifying separate development and production GPU clusters, and a public corporation extending existing VM-based GPU infrastructure to containerized AI environments.
▸Full profile — market position, technology, go-to-market, geography, history
Market position
Positions itself as an AI infrastructure optimization vendor for organizations that already own GPUs, emphasizing avoidance of additional GPU purchases amid rising server prices and long lead times; the site cites validation by global partners including NVIDIA and Cisco.
The company positions spatial GPU partitioning as its differentiator against time-division GPU sharing, claiming independent allocation of GPU memory and cores per workload and materially better workload isolation, and states it was first to develop spatial-partitioning-based dynamic GPU allocation. Additional differentiators cited are a proprietary scheduler for priority-based allocation, per-GPU power control, and pre-purchase workload benchmarking through RA:X.
Technology
AIPub applies spatial partitioning of GPU memory and cores to isolate concurrent workloads, combined with dynamic real-time reallocation and automatic reclamation of idle resources, a proprietary priority-based scheduler, monitoring covering more than 100 indicators at GPU and cluster level, granular per-GPU power control, and metering for cost attribution. RA:X performs workload benchmarking to size GPU performance and capacity. Public repositories include Kubernetes installers, a Java project controller and a Go-based tool for managing SR-IOV virtual functions and InfiniBand P-Key partitions with Kubernetes pod binding monitoring.
Go-to-market
Direct enterprise sales through inbound inquiry forms on the company site, led by a free RA:X GPU infrastructure diagnostic offer used as an entry point, supported by referenced deployments and global technology partners.
Enterprises and institutions running AI/ML and LLM workloads on owned GPU infrastructure, including large manufacturers, public research institutions, universities and public-sector corporations.
Geography
Headquartered in Seoul, South Korea, with customer references drawn from Korean enterprises, public research institutions, universities and public corporations.
History
TEN Inc. was established in 2020 and is based in Seoul, South Korea. In December 2022 it raised pre-Series A funding from Ascendo Ventures, Korea Credit Guarantee Fund, Quantum Ventures Korea and Shinhan Capital, with proceeds intended to expand operations and business reach. As of 2022 the company was led by CEO Sejin Oh.
Compiled by commissioned research from 6 cited public sources — announcements, filings, and press listed under research sources below.
Key figures
latest reportedCompany-reported or press-reported figures, each dated to when it was claimed — not independently audited.
Timeline · 2
launches, deals, and filingsThe company promotes a one-time free RA:X infrastructure diagnostic that benchmarks a prospect's existing GPU environment.
TEN Inc. raised KRW 3 billion in pre-Series A funding from new investors Ascendo Ventures, Korea Credit Guarantee Fund, Quantum Ventures Korea and Shinhan Capital.
Dated company events from announcements, filings, and press; legal rows summarize public dockets and regulator releases.
▸Research sources · 6
primary sources listed
- Ten1010ten1010.io · web
6 public sources were cited for this profile; the first-party ones are listed here.
Frequently asked questions
- What does Ten1010 do?
- Korean enterprise AI infrastructure company whose AIPub platform partitions and dynamically allocates GPUs to raise utilization.
- Who are Ten1010's investors?
- Ten1010's investors include Ascendo Ventures.