Seoul, South Korea – June 30, 2026 – Rebellions announced on June 30 the acquisition of SqueezeBits, a company specializing in AI inference optimization. With this acquisition, Rebellions is accelerating its evolution into an integrated AI infrastructure company that delivers NPU hardware, software optimization, and inference serving on a single platform.

It encompasses the operational pipeline from receiving a request to executing the model and returning the result, and directly determines a service’s speed, cost, and reliability.

As AI services move into commercial deployment across every industry, how efficiently a model runs in real-world environments is becoming as critical to competitiveness as the performance of the model itself. System-level capabilities that integrate everything from server and rack configuration to serving software into a single infrastructure are emerging as a decisive source of advantage, and this acquisition brings that capability fully in-house at Rebellions.

SqueezeBits is a deep-tech startup founded by researchers with deep expertise in AI semiconductors, deep learning, and model compression. The company has built optimization and lightweighting technologies that accelerate AI models and reduce operating costs across a wide range of hardware environments. In a short time since its founding, SqueezeBits has established a strong track record of collaboration with global AI hardware leaders including Intel and NVIDIA, earning broad recognition for its engineering capability.

The two companies have maintained a close partnership in inference optimization. Since 2024, they have jointly developed model compression technologies and dedicated software for Rebellions’ NPUs, and have driven the growth of an NPU-based open-source AI ecosystem by hosting events and workshops centered on vLLM, an open-source inference optimization framework, for the Korean developer community. Building on the technical trust earned through that work, the two companies expect synergies to materialize immediately once the SqueezeBits team joins Rebellions.

Through this acquisition, Rebellions will maximize the software performance of its AI semiconductors while strengthening its ability to deliver AI infrastructure end-to-end, from hardware to serving. The combined platform is designed to ease the burden of complex optimization at the deployment stage, giving customers an environment in which AI services can be built quickly and efficiently.

In March 2026, Rebellions was selected as the first direct investment of the National Growth Fund, positioning the company as a central player in Korea’s effort to cultivate a “K-NVIDIA.” Choosing SqueezeBits as its first acquisition extends that trajectory beyond hardware, into the advancement of the broader AI industry ecosystem and the build-out of sovereign AI infrastructure. The deal continues the path opened by Rebellions’ 2024 merger with Sapeon Korea — the first integration between domestic AI semiconductor companies — as Rebellions builds global scale and competitiveness through proactive strategic combinations.

Sunghyun Park, CEO of Rebellions, said, “We believe Korea’s AI infrastructure ecosystem opens up new possibilities when technical capability and exceptional talent come together across the boundaries of individual companies.” He added, “By joining forces with SqueezeBits, Rebellions will grow into a company that spans hardware, software, and system-level AI infrastructure at scale, and we intend to prove that conviction in the global market.”

Hyungjun Kim, CEO of SqueezeBits, said, “SqueezeBits’ AI inference optimization technology will significantly expand the reach of the Rebellions NPU ecosystem.” He continued, “Building on our synergy with Rebellions, we will deliver a full-stack AI infrastructure in which hardware and software are optimized together, and help customers run AI services on Rebellions NPUs more easily and more economically.”

About SqueezeBits

Founded in March 2022, SqueezeBits is a deep-tech startup developing AI model compression and optimization technologies to reduce the cost of deploying and operating AI services. The company brings together AI researchers from SNU, POSTECH, and KAIST, along with AI industry experts. Recently, SqueezeBits expanded into heterogeneous computing-based API services, demonstrating full-stack capabilities spanning hardware and software. With two consecutive years of profitability, the company has proven both its technical capability and financial viability. Through strategic collaborations with a range of partners, SqueezeBits is driving innovation in the global AI ecosystem. https://www.squeezebits.com/