REDWOOD CITY, Calif., Jan. 31, 2026 /PRNewswire/ — Zilliz, the company behind the leading open-source vector database Milvus, today announced the open-source release of its Bilingual Semantic Highlighting Model, an industry-first AI model designed to dramatically reduce token usage and improve answer quality in production RAG-powered AI applications.

This highlighting model introduces sentence-level relevance filtering, enabling AI developers to remove low-signal context before sending prompts to large language models. This approach directly addresses rising inference costs and accuracy issues caused by oversized context windows in enterprise RAG and RAG-powered AI deployments.

“As RAG systems move into production, teams are running into very real cost and quality limits,” said James Luan, VP of Engineering at Zilliz. “This model gives developers a practical way to reduce prompt size and improve answer accuracy without reworking their existing pipelines.”

Key Innovations and Technical Breakthroughs

• Bilingual relevance by design: Optimized for both English and Chinese, the model addresses cross-lingual relevance challenges common in global RAG deployments. It is built on the MiniCPM-2B architecture, enabling low-latency, production-ready performance.
• Sentence-level context filtering: Rather than scoring entire document chunks, the model evaluates relevance at the sentence level and retains only content that directly supports a user query before sending it to the LLM.
• Lower token usage, higher answer quality: Zilliz reports that sentence-level filtering significantly compresses prompt size while improving downstream response quality, helping teams reduce inference costs and improve generation speed in production environments.

Availability

The Bilingual Semantic Highlighting Model is available today as an open-source release. To learn more about the training methodology and performance benchmarks, visit the Zilliz Technical Blog.

Download: : zilliz/semantic-highlight-bilingual-v1

About Zilliz

Zilliz is the company behind Milvus, the world’s most widely adopted open-source vector database. Zilliz Cloud brings that performance to production with a fully managed, cloud-native platform built for scalable, low-latency vector search and hybrid retrieval. It supports billion-scale workloads with sub-10ms latency, auto-scaling, and optimized indexes for GenAI use cases like semantic search and RAG.

Zilliz is built to make AI not just possible—but practical. With a focus on performance and cost-efficiency, it helps engineering teams move from prototype to production without overprovisioning or complex infrastructure. Over 10,000 organizations worldwide rely on Zilliz to build intelligent applications at scale.

Headquartered in Redwood Shores, California, Zilliz is backed by leading investors, including Aramco’s Prosperity 7 Ventures, Temasek’s Pavilion Capital, Hillhouse Capital, 5Y Capital, Yunqi Partners, Trustbridge Partners, and others. Learn more at  Zilliz.com.

REDWOOD CITY, Calif., Jan. 31, 2026 /PRNewswire/ — Zilliz, the company behind the leading open-source vector database Milvus, today announced the open-source release of its Bilingual Semantic Highlighting Model, an industry-first AI model designed to dramatically reduce token usage and improve answer quality in production RAG-powered AI applications.

This highlighting model introduces sentence-level relevance filtering, enabling AI developers to remove low-signal context before sending prompts to large language models. This approach directly addresses rising inference costs and accuracy issues caused by oversized context windows in enterprise RAG and RAG-powered AI deployments.

“As RAG systems move into production, teams are running into very real cost and quality limits,” said James Luan, VP of Engineering at Zilliz. “This model gives developers a practical way to reduce prompt size and improve answer accuracy without reworking their existing pipelines.”

Key Innovations and Technical Breakthroughs

• Bilingual relevance by design: Optimized for both English and Chinese, the model addresses cross-lingual relevance challenges common in global RAG deployments. It is built on the MiniCPM-2B architecture, enabling low-latency, production-ready performance.
• Sentence-level context filtering: Rather than scoring entire document chunks, the model evaluates relevance at the sentence level and retains only content that directly supports a user query before sending it to the LLM.
• Lower token usage, higher answer quality: Zilliz reports that sentence-level filtering significantly compresses prompt size while improving downstream response quality, helping teams reduce inference costs and improve generation speed in production environments.

Availability

The Bilingual Semantic Highlighting Model is available today as an open-source release. To learn more about the training methodology and performance benchmarks, visit the Zilliz Technical Blog.

Download: : zilliz/semantic-highlight-bilingual-v1

About Zilliz

Zilliz is the company behind Milvus, the world’s most widely adopted open-source vector database. Zilliz Cloud brings that performance to production with a fully managed, cloud-native platform built for scalable, low-latency vector search and hybrid retrieval. It supports billion-scale workloads with sub-10ms latency, auto-scaling, and optimized indexes for GenAI use cases like semantic search and RAG.

Zilliz is built to make AI not just possible—but practical. With a focus on performance and cost-efficiency, it helps engineering teams move from prototype to production without overprovisioning or complex infrastructure. Over 10,000 organizations worldwide rely on Zilliz to build intelligent applications at scale.

Headquartered in Redwood Shores, California, Zilliz is backed by leading investors, including Aramco’s Prosperity 7 Ventures, Temasek’s Pavilion Capital, Hillhouse Capital, 5Y Capital, Yunqi Partners, Trustbridge Partners, and others. Learn more at  Zilliz.com.

SINGAPORE, Jan. 31, 2026 /PRNewswire/ — Photo Dance, a globally leading AI photo dance application, today announced a major product upgrade with the introduction of its new feature, Magic AI Video.

Magic AI Video is a groundbreaking motion-transfer feature that seamlessly blends a user-uploaded static image with the movements and facial expressions from a reference video, producing highly fluid and lifelike dynamic videos. Powered by advanced temporal consistency algorithms, the generated content maintains smooth continuity from frame to frame, completely eliminating frame drops and motion artifacts. This technology lowers the barrier to professional-grade motion capture to zero—any user can bring a person in a single photo to life, perfectly on beat and vividly expressive. With an exceptionally low creation threshold and strong shareability, it is inherently designed for viral growth.

Photo Dance is an AI-powered photo dance app developed by SPARKHUB PTE. LTD. Since its launch, Photo Dance has achieved more than 1.2 million downloads worldwide, with users generating over 1.8 million AI dance videos across global markets.

As conversations around “AI photo dance” continue to gain momentum on social media platforms, Photo Dance‘s continuous innovation and product upgrades are helping to redefine the category. Through measurable performance data, real-world use cases, and a long-term commitment to technological advancement, Photo Dance is setting new standards for creativity and accessibility in AI-generated video content.

What is Photo Dance? What Problem Does It Solve?

Photo Dance is an AI-powered photo-to-dance video application whose core function is transforming any static photograph into a dancing video.

Unlike traditional dance content creation, it doesn’t require users to film videos or dance themselves. Instead, through AI motion transfer technology, Photo Dance brings people, babies, pets, or virtual characters in photos to life with dance movements. For users who prefer not to appear on camera, lack dancing skills, or simply want to effortlessly participate in short-form video trends, Photo Dance offers a low-barrier entry point into dance culture.

Is Photo Dance Easy to Use? Real User Experience

From a user experience standpoint, Photo Dance‘s workflow is designed to be straightforward:

Select template → Choose photo → Generate video. The entire process requires no professional editing or AI operation experience.

Photo Dance offers 500+ dance templates covering trending TikTok choreography and supports continuous updates. Videos are typically generated within 3 minutes and can be directly shared on TikTok, Instagram, YouTube, and other platforms.

To date, Photo Dance has accumulated:

• 1.2 million downloads globally
• 4.4-star App Store rating
• 1.8+ million videos generated

Users consistently highlight keywords like “fun,” “quality exceeded expectations,” “fast rendering speed,” and “simple to operate” in their reviews.

How is Photo Dance Different from Other Photo Dance Apps?

In common user comparisons, Photo Dance is frequently discussed alongside other tools. Photo Dance distinguishes itself on three key points:

• Template Quantity & Update Frequency: Covers more trending TikTok choreography
• Generation Quality & Stability: Reduces motion distortion and unnatural movements
• Creation Efficiency: No lengthy waiting times or complex setup – simple workflow for quick creation

This makes Photo Dance ideal for ordinary users to rapidly produce and share “viral-worthy content,” rather than positioning it as a specialized or experimental AI tool.

What Scenarios Can Photo Dance Be Used For?

Based on user behavior patterns, Photo Dance‘s applications extend beyond entertainment:

• Baby Dance Videos: One of the most popular content types among family users
• Pet Dancing & Funny Expressions: Perfect for lighthearted social sharing
• Anime & Character Dancing: Used for fan-created content and virtual character videos
• Old Photo Animation: Bringing precious memories “back to life”

The last category, in particular, has sparked discussions about “appropriateness” and “boundaries” among some users.

On Controversy: Beyond Technology, Usage Matters Most

Regarding discussions about “whether old photos or photos of deceased loved ones should dance,” the Photo Dance team hasn’t shied away from the topic. They emphasize that the technology itself carries no preset emotional stance – what truly matters is the user’s intention and method of use.

Many users choose to create memorial videos, family memories, or personal emotional content rather than public content. The Photo Dance team encourages users to employ these features with respect, goodwill, and personal judgment.

Photo Dance‘s recently launched Magic AI Video feature further optimizes motion control and naturalness, producing more restrained and authentic results that reduce exaggerated or inappropriate expressions.

Safety & Trustworthiness: One of Users’ Top Concerns

In AI applications, “is it safe? Is it trustworthy?” remains a focus of user inquiries. Photo Dance is currently available exclusively on iOS, adheres to app store privacy and subscription regulations, and clearly discloses subscription and auto-renewal mechanisms. The team has also stated they will continue optimizing product transparency and user experience.

Photo Dance: An AI Creation Tool Built for Short-Form Video

Photo Dance doesn’t position itself as an “all-in-one AI video tool,” but rather focuses on a clear objective: enabling ordinary users to create and express themselves on short-form video platforms at minimal cost.

As AI continues to lower barriers to content creation, Photo Dance represents a broader trend – shifting from “Can I make videos?” to “Do I want to express myself?”

About Photo Dance

Photo Dance is an AI photo dance application developed by SPARKHUB PTE. LTD., enabling users to transform any photograph into a dance video suitable for TikTok, Instagram, YouTube, and other video platforms. Photo Dance targets users aged 25–45, delivering a lightweight, entertaining, and shareable creation experience.

Product Overview:

• Slogan: Turn photos into viral dance videos in seconds.
• Official Website: https://photodanceapp.com/ 
• TikTok Official Account: https://www.tiktok.com/@photodance_ai 
• App Store Link: https://apps.apple.com/us/app/photo-dance-ai-baby-dance/id6474650062 

About SPARKHUB PTE. LTD.

Founded on May 16, 2024, and headquartered in Singapore, SPARKHUB PTE. LTD. is a team with deep expertise in the mobile social space. Our members are energetic and passionate about creating fun experiences. We are committed to developing innovative mobile apps that connect users with millions of people nearby and around the world.

Company Overview:

• Company Name: SPARKHUB PTE. LTD.
• Address: 10 ANSON ROAD #12-08 INTERNATIONAL PLAZA SINGAPORE (079903)
• Official Website: https://sparkhubcorp.com/ 

Contact Information:

• Company: SPARKHUB PTE. LTD.
• Contact Person: Karen
• Email: feedback@sparkhubcorp.com 

SINGAPORE, Jan. 31, 2026 /PRNewswire/ — Photo Dance, a globally leading AI photo dance application, today announced a major product upgrade with the introduction of its new feature, Magic AI Video.

Magic AI Video is a groundbreaking motion-transfer feature that seamlessly blends a user-uploaded static image with the movements and facial expressions from a reference video, producing highly fluid and lifelike dynamic videos. Powered by advanced temporal consistency algorithms, the generated content maintains smooth continuity from frame to frame, completely eliminating frame drops and motion artifacts. This technology lowers the barrier to professional-grade motion capture to zero—any user can bring a person in a single photo to life, perfectly on beat and vividly expressive. With an exceptionally low creation threshold and strong shareability, it is inherently designed for viral growth.

Photo Dance is an AI-powered photo dance app developed by SPARKHUB PTE. LTD. Since its launch, Photo Dance has achieved more than 1.2 million downloads worldwide, with users generating over 1.8 million AI dance videos across global markets.

As conversations around “AI photo dance” continue to gain momentum on social media platforms, Photo Dance‘s continuous innovation and product upgrades are helping to redefine the category. Through measurable performance data, real-world use cases, and a long-term commitment to technological advancement, Photo Dance is setting new standards for creativity and accessibility in AI-generated video content.

What is Photo Dance? What Problem Does It Solve?

Photo Dance is an AI-powered photo-to-dance video application whose core function is transforming any static photograph into a dancing video.

Unlike traditional dance content creation, it doesn’t require users to film videos or dance themselves. Instead, through AI motion transfer technology, Photo Dance brings people, babies, pets, or virtual characters in photos to life with dance movements. For users who prefer not to appear on camera, lack dancing skills, or simply want to effortlessly participate in short-form video trends, Photo Dance offers a low-barrier entry point into dance culture.

Is Photo Dance Easy to Use? Real User Experience

From a user experience standpoint, Photo Dance‘s workflow is designed to be straightforward:

Select template → Choose photo → Generate video. The entire process requires no professional editing or AI operation experience.

Photo Dance offers 500+ dance templates covering trending TikTok choreography and supports continuous updates. Videos are typically generated within 3 minutes and can be directly shared on TikTok, Instagram, YouTube, and other platforms.

To date, Photo Dance has accumulated:

• 1.2 million downloads globally
• 4.4-star App Store rating
• 1.8+ million videos generated

Users consistently highlight keywords like “fun,” “quality exceeded expectations,” “fast rendering speed,” and “simple to operate” in their reviews.

How is Photo Dance Different from Other Photo Dance Apps?

In common user comparisons, Photo Dance is frequently discussed alongside other tools. Photo Dance distinguishes itself on three key points:

• Template Quantity & Update Frequency: Covers more trending TikTok choreography
• Generation Quality & Stability: Reduces motion distortion and unnatural movements
• Creation Efficiency: No lengthy waiting times or complex setup – simple workflow for quick creation

This makes Photo Dance ideal for ordinary users to rapidly produce and share “viral-worthy content,” rather than positioning it as a specialized or experimental AI tool.

What Scenarios Can Photo Dance Be Used For?

Based on user behavior patterns, Photo Dance‘s applications extend beyond entertainment:

• Baby Dance Videos: One of the most popular content types among family users
• Pet Dancing & Funny Expressions: Perfect for lighthearted social sharing
• Anime & Character Dancing: Used for fan-created content and virtual character videos
• Old Photo Animation: Bringing precious memories “back to life”

The last category, in particular, has sparked discussions about “appropriateness” and “boundaries” among some users.

On Controversy: Beyond Technology, Usage Matters Most

Regarding discussions about “whether old photos or photos of deceased loved ones should dance,” the Photo Dance team hasn’t shied away from the topic. They emphasize that the technology itself carries no preset emotional stance – what truly matters is the user’s intention and method of use.

Many users choose to create memorial videos, family memories, or personal emotional content rather than public content. The Photo Dance team encourages users to employ these features with respect, goodwill, and personal judgment.

Photo Dance‘s recently launched Magic AI Video feature further optimizes motion control and naturalness, producing more restrained and authentic results that reduce exaggerated or inappropriate expressions.

Safety & Trustworthiness: One of Users’ Top Concerns

In AI applications, “is it safe? Is it trustworthy?” remains a focus of user inquiries. Photo Dance is currently available exclusively on iOS, adheres to app store privacy and subscription regulations, and clearly discloses subscription and auto-renewal mechanisms. The team has also stated they will continue optimizing product transparency and user experience.

Photo Dance: An AI Creation Tool Built for Short-Form Video

Photo Dance doesn’t position itself as an “all-in-one AI video tool,” but rather focuses on a clear objective: enabling ordinary users to create and express themselves on short-form video platforms at minimal cost.

As AI continues to lower barriers to content creation, Photo Dance represents a broader trend – shifting from “Can I make videos?” to “Do I want to express myself?”

About Photo Dance

Photo Dance is an AI photo dance application developed by SPARKHUB PTE. LTD., enabling users to transform any photograph into a dance video suitable for TikTok, Instagram, YouTube, and other video platforms. Photo Dance targets users aged 25–45, delivering a lightweight, entertaining, and shareable creation experience.

Product Overview:

• Slogan: Turn photos into viral dance videos in seconds.
• Official Website: https://photodanceapp.com/ 
• TikTok Official Account: https://www.tiktok.com/@photodance_ai 
• App Store Link: https://apps.apple.com/us/app/photo-dance-ai-baby-dance/id6474650062 

About SPARKHUB PTE. LTD.

Founded on May 16, 2024, and headquartered in Singapore, SPARKHUB PTE. LTD. is a team with deep expertise in the mobile social space. Our members are energetic and passionate about creating fun experiences. We are committed to developing innovative mobile apps that connect users with millions of people nearby and around the world.

Company Overview:

• Company Name: SPARKHUB PTE. LTD.
• Address: 10 ANSON ROAD #12-08 INTERNATIONAL PLAZA SINGAPORE (079903)
• Official Website: https://sparkhubcorp.com/ 

Contact Information:

• Company: SPARKHUB PTE. LTD.
• Contact Person: Karen
• Email: feedback@sparkhubcorp.com 

DUBAI, UAE, Jan. 30, 2026 /PRNewswire/ — Bybit, the world’s second-largest cryptocurrency exchange by trading volume, has released the latest Bybit x Block Scholes Crypto Derivatives Analytics report, analyzing recent market stress across digital assets as investors reassess expectations for United States monetary policy.

Key findings include:

• Broad crypto market selloff alongside global risk assets
• Sharp rise in short-term implied volatility for Bitcoin and Ether
• Limited response in longer-dated volatility despite steep price declines
• Muted derivatives participation and subdued trading volumes

The report shows that cryptocurrencies have been caught in a wider selloff across global markets, driven by a rapid repricing of the Federal Reserve’s policy outlook and renewed strength in the US dollar. The move erased roughly 4.7 percent of total crypto market capitalization in 24 hours, with losses concentrated among major tokens.

Bitcoin fell to around $81,000, placing it more than 30 percent below its Oct. 6, 2025 high of $126,100, while Ether even dipped below $2,700, dropping well below the $3,000 psychological level. On a year-to-date basis, Bitcoin and Ether are down more than 5 percent and 8 percent, respectively.

That selloff translated into a sharp increase in short-term options implied volatility. One-week at-the-money implied volatility climbed to about 46 percent for Bitcoin and roughly 58 percent for Ether. However, longer-dated volatility did not rise as aggressively, suggesting that traders are primarily pricing in elevated near-term uncertainty rather than a prolonged period of market stress.

Derivatives market participation remains subdued. Open interest in perpetual futures contracts has not fallen sharply alongside spot prices and remains well below levels recorded ahead of the October 2025 liquidation event. Daily trading volumes in perpetual contracts are also significantly lower than those seen during the first three quarters of 2025, indicating continued caution among market participants.

“Cryptos have been caught up in the selloff across global assets, as markets aggressively reprice the Fed policy outlook under a presumably less-dovish Fed Chair,” said Han Tan, Chief market analyst at Bybit Learn. “With US-listed Bitcoin ETFs posting three straight months of net outflows, traders will be keeping a wary eye on the $80,000 level. A sustained break below that psychologically important line may extend Bitcoin’s slump into the mid-$70,000 region, revisiting levels not seen since the aftermath of Liberation Day.”

The report also notes that recent Federal Open Market Committee communications had little impact on longer-term volatility expectations. Despite a slightly hawkish tone and an emphasis on a wait-and-see approach, implied volatility across longer tenors for both Bitcoin and Ether has continued to trend lower since peaking in late 2025.

Overall, the analysis suggests that while short-term volatility has intensified, the absence of a sustained rise in longer-dated volatility and trading activity reflects a market environment marked by restrained participation and cautious positioning. The full Bybit x Block Scholes report is available for download.

#Bybit / #CryptoArk / #BybitLearn

About Bybit

Bybit is the world’s second-largest cryptocurrency exchange by trading volume, serving a global community of over 80 million users. Founded in 2018, Bybit is redefining openness in the decentralized world by creating a simpler, open and equal ecosystem for everyone. With a strong focus on Web3, Bybit partners strategically with leading blockchain protocols to provide robust infrastructure and drive on-chain innovation. Renowned for its secure custody, diverse marketplaces, intuitive user experience, and advanced blockchain tools, Bybit bridges the gap between TradFi and DeFi, empowering builders, creators, and enthusiasts to unlock the full potential of Web3. Discover the future of decentralized finance at Bybit.com.

For more details about Bybit, please visit Bybit Press

For media inquiries, please contact: media@bybit.com

For updates, please follow: Bybit’s Communities and Social Media

Discord | Facebook | Instagram | LinkedIn | Reddit | Telegram | TikTok | X | Youtube

DUBAI, UAE, Jan. 30, 2026 /PRNewswire/ — Bybit, the world’s second-largest cryptocurrency exchange by trading volume, has released the latest Bybit x Block Scholes Crypto Derivatives Analytics report, analyzing recent market stress across digital assets as investors reassess expectations for United States monetary policy.

Key findings include:

• Broad crypto market selloff alongside global risk assets
• Sharp rise in short-term implied volatility for Bitcoin and Ether
• Limited response in longer-dated volatility despite steep price declines
• Muted derivatives participation and subdued trading volumes

The report shows that cryptocurrencies have been caught in a wider selloff across global markets, driven by a rapid repricing of the Federal Reserve’s policy outlook and renewed strength in the US dollar. The move erased roughly 4.7 percent of total crypto market capitalization in 24 hours, with losses concentrated among major tokens.

Bitcoin fell to around $81,000, placing it more than 30 percent below its Oct. 6, 2025 high of $126,100, while Ether even dipped below $2,700, dropping well below the $3,000 psychological level. On a year-to-date basis, Bitcoin and Ether are down more than 5 percent and 8 percent, respectively.

That selloff translated into a sharp increase in short-term options implied volatility. One-week at-the-money implied volatility climbed to about 46 percent for Bitcoin and roughly 58 percent for Ether. However, longer-dated volatility did not rise as aggressively, suggesting that traders are primarily pricing in elevated near-term uncertainty rather than a prolonged period of market stress.

Derivatives market participation remains subdued. Open interest in perpetual futures contracts has not fallen sharply alongside spot prices and remains well below levels recorded ahead of the October 2025 liquidation event. Daily trading volumes in perpetual contracts are also significantly lower than those seen during the first three quarters of 2025, indicating continued caution among market participants.

“Cryptos have been caught up in the selloff across global assets, as markets aggressively reprice the Fed policy outlook under a presumably less-dovish Fed Chair,” said Han Tan, Chief market analyst at Bybit Learn. “With US-listed Bitcoin ETFs posting three straight months of net outflows, traders will be keeping a wary eye on the $80,000 level. A sustained break below that psychologically important line may extend Bitcoin’s slump into the mid-$70,000 region, revisiting levels not seen since the aftermath of Liberation Day.”

The report also notes that recent Federal Open Market Committee communications had little impact on longer-term volatility expectations. Despite a slightly hawkish tone and an emphasis on a wait-and-see approach, implied volatility across longer tenors for both Bitcoin and Ether has continued to trend lower since peaking in late 2025.

Overall, the analysis suggests that while short-term volatility has intensified, the absence of a sustained rise in longer-dated volatility and trading activity reflects a market environment marked by restrained participation and cautious positioning. The full Bybit x Block Scholes report is available for download.

#Bybit / #CryptoArk / #BybitLearn

About Bybit

Bybit is the world’s second-largest cryptocurrency exchange by trading volume, serving a global community of over 80 million users. Founded in 2018, Bybit is redefining openness in the decentralized world by creating a simpler, open and equal ecosystem for everyone. With a strong focus on Web3, Bybit partners strategically with leading blockchain protocols to provide robust infrastructure and drive on-chain innovation. Renowned for its secure custody, diverse marketplaces, intuitive user experience, and advanced blockchain tools, Bybit bridges the gap between TradFi and DeFi, empowering builders, creators, and enthusiasts to unlock the full potential of Web3. Discover the future of decentralized finance at Bybit.com.

For more details about Bybit, please visit Bybit Press

For media inquiries, please contact: media@bybit.com

For updates, please follow: Bybit’s Communities and Social Media

Discord | Facebook | Instagram | LinkedIn | Reddit | Telegram | TikTok | X | Youtube

HONG KONG, Jan. 30, 2026 /PRNewswire/ — SUEZ has been awarded the contract for supplying sludge dewatering solution for the Relocation of Sha Tin Sewage Treatment Works to Caverns by the Drainage Services Department (DSD) of Hong Kong SAR Government. The project will be delivered by SUEZ, leveraging a sludge dewatering technology developed and continuously improved over more than 15 years through SUEZ’s collaboration with Bucher Unipektin AG[1]. This sludge dewatering contract covers design, equipment supply and delivery for the new Sha Tin Cavern Sewage Treatment Works (STCSTW).

Scheduled for completion in 2030, the new facility is set to become one of Asia’s largest sewage treatment plants built inside an artificial cave. With its compact design and advanced process standards, the STCSTW will optimise land use and set a new benchmark for sustainable, green infrastructure in Hong Kong.

New cavern-based wastewater facility to support sanitation needs and relieve land pressure

The existing Sha Tin Sewage Treatment Works has served approximately 700,000 residents for over 40 years. As the facility needed modernisation works, the DSD has decided not only to address the challenge of aging facilities, but also to tackle the acute land scarcity issue and to improve surrounding community environment.

The DSD initiated the relocation of the Sha Tin Sewage Treatment Works into a purpose-built cavern, comprising a 14-hectare site and an upstream sewage collection system. This strategic move allows major benefits for the Hong Kong population: It will limit olfactory nuisance as caverns act as natural barriers that contain and attenuate odours. Furthermore, inside the cavern, odour sources will be enclosed and dedicated de–odourisation facilities will be installed. It will free up 28 hectares of surface land for other beneficial uses.

Long-term technology collaboration between SUEZ and Bucher Unipektin AG supporting cavern operations in Hong Kong

Recognising the constraints of cavern-based facilities, including limited space, controlled access and strict safety requirements, SUEZ will equip the project with 16 units of its proven Dehydris™ Twist sludge dewatering system. Fully enclosed and highly automated, this system supports safer operations in confined areas and helps reduce operator exposure. Its compact design reduces footprint by 25% and offers greater flexibility to allocate other essential equipment within the caverns.

Dehydris™ Twist sludge dewatering solution has proven to be stable and reliable across SUEZ projects in France, Italy, Australia, and the Chinese Mainland. In Hong Kong, seawater flushing for about 85% of the population increases sewage salinity, typically reported around 5,000 – 6,000 mg/L, which can influence dewatering performance.

To confirm results under local conditions, Dehydris™ Twist was piloted in Hong Kong for several years. The trials showed consistent performance, raising sludge solid from around 3% to no less than 30%, reducing sludge volume and associated transport and disposal costs. The system is expected to produce about 120,000 tonnes of dewatered sludge annually, supporting energy recovery and Hong Kong’s circularity and carbon reduction goals.

Commenting on the long-term technology collaboration supporting this project, Daniel Schneider, CEO Bucher Unipektin AG, said: “The implementation of Bucher HPS Technology at the new STCSTW as part of the SUEZ Dehydris™ Twist workshop marks a significant milestone in acknowledging the growing need for advanced biosolids dewatering solutions. Our technology sets new standards in efficiency and sustainability, demonstrating what is now achievable in modern wastewater treatment. There is no doubt that this innovation will play a pivotal role in supporting sustainable and economically efficient development, both in the rapidly expanding markets of Southeast Asia and across the rest of the world.”

Pierre Pauliac, Executive Vice-President of SUEZ, in charge of international activities, said: “This project is an important opportunity for SUEZ to deliver its core sludge dewatering solution for the new cavern-based treatment plant, integrating our global expertise to create local value for Hong Kong. In the Greater Bay Area, we have been committed for decades to providing sustainable environmental solutions: operating two wastewater treatment plants in Hong Kong to improve the water environment; providing quality drinking water services to the entire city of Macao; and supporting industries in Guangdong with advanced wastewater treatment solutions for its green transition. Going forward, we will continue to leverage our expertise in water and waste, working closely with all stakeholders to drive the Greater Bay Area towards a greener, more resilient, and sustainable future.“

About SUEZ:

Faced with growing environmental challenges, SUEZ has been delivering essential services that protect and improve our quality of life for more than 160 years. SUEZ provides its customers with innovative and resilient solutions for water and waste services. With 40 000 employees across 40 countries, the Group works with customers to create value over the full lifecycle of their assets and services, and to drive their low carbon transition. In 2024, SUEZ provided drinking water for 68 million people worldwide and sanitation services for 44 million people. The Group generated 8 TWh of energy from waste and wastewater. In 2024, SUEZ has generated revenues of 9.2 billion euros. For more information:  www.SUEZ.com.

In Asia, SUEZ’s journey began some 70 years ago, first in Southeast Asia before expanding to China 50 years ago. With over 6,500 employees across Asia, SUEZ has built more than 600 water and wastewater treatment plants. Today, SUEZ works with municipal and industrial customers in more than 30 major cities and regions, providing water and waste recycling and recovery services to 25+ million people and 23 industrial parks.