Korean startup backed by Samsung and Arm launches rack-sized inference monsters, claims "6x lower power consumption" and up to 75% cheaper acquisition cost compared to Nvidia

Date:
Sat, 04 Apr 2026 16:45:00 +0000

Description:
Rebellions launches rack-scale AI infrastructure systems, claiming lower
power usage and costs, backed by $400 million pre-IPO funding.

FULL STORY ======================================================================Copy link Facebook X Whatsapp Reddit Pinterest Flipboard Threads Email Share this article 0 Join the conversation Follow us Add us as a preferred source on Google Newsletter Tech Radar Pro Are you a pro? Subscribe to our newsletter Sign up to the TechRadar Pro newsletter to get all the top news, opinion, features and guidance your business needs to succeed! Become a Member in Seconds Unlock instant access to exclusive member features. Contact me with news and offers from other Future brands Receive email from us on behalf of our trusted partners or sponsors By submitting your information you agree to the Terms & Conditions and Privacy Policy and are aged 16 or over. You are
now subscribed Your newsletter sign-up was successful Join the club Get full access to premium articles, exclusive features and a growing list of member rewards. Explore An account already exists for this email address, please log in. Subscribe to our newsletter Rebellions launches modular AI systems designed for scalable data center deployment RebelRack operates as a single production-ready rack for AI workloads RebelPOD scales infrastructure into clustered deployments for larger enterprise workloads Rebellions has introduced two rack-scale inference systems, RebelRack and RebelPOD,
extending its platform beyond chip design into fully deployable infrastructure.

These systems are designed to run artificial intelligence workloads directly within data center environments, combining hardware and software into integrated units. RebelRack operates as a single production-ready rack, while RebelPOD scales this model into clustered deployments intended for larger workloads. Article continues below You may like The AI data centers of 2036 wont be filled with GPUs: FuriosaAIs CEO on the future of silicon No Nvidia, No AMD, No Intel, No ARM: Meta plans inference-led RISC-y future without friends as 1700w superchip emerges with 30 PFLOPs performance and half Terabyte (yes 512GB) HBM Nvidia wants to power the next generation of data centers in space Performance and cost claims draw scrutiny The company claims these systems deliver 6x lower power consumption and up to 75% lower acquisition costs compared to Nvidia .

These claims focus on efficiency at the system level, where power usage and total cost of ownership have become central concerns for operators.

While such figures suggest meaningful reductions, they depend on workload conditions and deployment environments, which may vary across use cases.

The infrastructure is built around the Rebel100 neural processing unit and supported by a cloud-native software stack designed for production environments. Are you a pro? Subscribe to our newsletter Sign up to the TechRadar Pro newsletter to get all the top news, opinion, features and guidance your business needs to succeed! Contact me with news and offers from other Future brands Receive email from us on behalf of our trusted partners
or sponsors By submitting your information you agree to the Terms &
Conditions and Privacy Policy and are aged 16 or over.

The platform integrates with widely used frameworks, including PyTorch and Kubernetes-based systems, allowing deployment across different models and infrastructure setups.

The launch reflects a broader shift in the artificial intelligence sector, where the ability to run models efficiently is becoming as important as developing them.

AI is now measured by its ability to operate in the real world at scale, under power constraints, and with clear economic return, says Sunghyun Park, Co-Founder and CEO of Rebellions. What to read next Powering the AI data center boom: the infrastructure upgrades behind innovation 'AI is entering
its next frontier... the foundation of the AI industrial revolution': Nvidia confirms CoreWeave will be among the first to get Vera Rubin chips as it doubles down on financial commitments The battle of the SuperPods: Nvidia challenges Huawei with Vera Rubin powered DGX cluster that can deliver 28.8 Exaflops with only 576 GPUs

Data center operators are increasingly constrained by power availability and infrastructure limits, creating demand for systems that can deliver performance within those boundaries.

Rebellions is accelerating its international presence, with a focus on the United States, where demand for deployable AI infrastructure is rapidly growing.

The company aims to provide ready-to-use systems that integrate seamlessly with existing operations, allowing organizations to deploy AI workloads without long setup periods.

The firm emphasizes end-to-end support, combining hardware, validated software, and ongoing operational assistance to ensure production
reliability.

This approach aims to reduce integration challenges often encountered in data centers running diverse AI workloads.

The recent $400 million pre-IPO funding round, led by Mirae Asset Financial Group and the Korea National Growth Fund, will be used to expand
manufacturing capacity and strengthen supply chains.

This round brings Rebellions total valuation to approximately $2.34 billion, reflecting investor confidence in its strategy and market potential. Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!

And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.



======================================================================
Link to news story: https://www.techradar.com/pro/korean-startup-backed-by-samsung-and-arm-launche s-rack-sized-inference-monsters-claims-6x-lower-power-consumption-and-up-to-75 -percent-cheaper-acquisition-cost-compared-to-nvidia


--- Mystic BBS v1.12 A49 (Linux/64)
* Origin: tqwNet Technology News (1337:1/100)