Taiwan's aetherAI Injects Smart Medicine into Endoscopy and Hematology with GIGABYTE's Edge AI Mini-PC

Smart medicine is a branch of digital health that utilizes IT advancements such as artificial intelligence (AI) and the internet of things (IoT) to enhance and expedite healthcare delivery. This is especially pertinent as the world's population grows older: the United Nations estimates that by the mid-2030s, the number of individuals aged 80 and above will outnumber infants; by 2080, people aged 65 and above will surpass the number of children under 18. This intractable trend exacerbates the already critical shortage of doctors and nurses worldwide: the World Health Organization (WHO) projects a shortfall of 11 million health workers by 2030. It should come as no surprise, then, that the healthcare profession has emerged as a leader in enterprise AI adoption. A Forbes survey in 2025 found that 22% of U.S. healthcare organizations had already implemented domain-specific AI tools—more than twice the rate of AI deployment in the broader economy (9%).

Although AI can make an impact in various healthcare segments—a topic we covered extensively in our deep dive into the benefits of AI in healthcare and medicine—it is imperative that IT inventions contribute significantly to the speed, convenience, and precision of diagnosis. Otherwise, health workers may be reluctant to adopt AI tools, which in turn will stymie the proliferation of smart medicine and leave other pertinent issues like financing and payment unresolved.

"This is one field of digital health where standalone digitization does not create a lot of value," says Dr. Joe Yeh, Founder and CEO of the Taiwanese medical AI company aetherAI. "If the doctor is just looking at a monitor rather than a microscope, digitizing medical images only adds a layer of work that may bog down the process."

Founded in 2015, aetherAI is a digital pathology leader with over 70% market share in Taiwan and a global clientele spanning Germany, Japan, and the United States. Its portfolio comprises dozens of AI-powered diagnostic software that expedite image analysis and management in pathology to help operators complete diagnoses more efficiently, while also offering consulting services and total solutions combining hardware and software to upgrade the proficiency and efficacy of medical imaging in hospitals and labs. In the fields of endoscopy and hematology, aetherAI has produced a selection of advanced software, and it is now working with GIGABYTE Technology and its subsidiary GIGAIPC, which focuses on industrial computer manufacturing, to deliver edge AI and medical AI solutions for endoscopy ("aetherAI Endo") and hematology ("aetherAl Hema").

On the outside, aetherAI Endo and aetherAl Hema look like the quintessential mini-PCs for a professional setting, packing an impressive array of I/O ports in compact form factors measuring less than a dozen centimeters long and wide. In fact, they are the joining of an edge AI platform—GIGABYTE's GB-BRU7-255H BRIX Ultra Compact PC—with medical AI software developed by aetherAI. The solution for hematology, aetherAI Hema, is the bread-and-butter GB-BRU7-255H operating bone marrow smear analysis software on reliable hardware with powerful processing capabilities, offering automatic classification and counting reports for 15 types of bone marrow cells with automatic counting results consistent with the output of a senior human analyst 94% of the time, greatly optimizing the efficiency of data analysis. For aetherAl Endo, GIGABYTE customized the mini-PC to include M.2 expansion, additional I/O ports, and support for a video capture card, or "frame grabber", to enable real-time image inference and preliminary classification during colonoscopies. The resulting 96% accuracy and ultralow latency (<0.06s) impart a boost to computer-aided detection (CADe) with increased discovery rates for adenomas, polyps, and lesions.

"Our mission has always been to utilize state-of-the-art information technology to improve healthcare standards and quality of life," says Dr. Yeh. "We found that the key to implementing smart health not only lies in hospitals and doctors, but also in the perception of the patients. When people understand that AI can make diagnoses more accurate, they will be more inclined to pay for smart services, which in turn will lead to a broader adoption of AI tools in the healthcare industry."

Dr. Yeh and his team worked with renowned health centers such as NTU Hospital and Cathay Hospital to train AI models using hundreds of thousands of annotated medical images collected from thousands of patients representing diverse clinical scenarios. These models make it possible for the software to pinpoint minuscule markers of dangerous illness: for example, aetherAI Hema can automatically detect an unhealthy ratio of stained marrow cells, assisting operators to quickly arrive at the overall tally and ratio of cells; aetherAI Endo can help find tiny, flattened precancerous neoplastic polyps in the digestive tract that may need to be removed. The team fine-tuned the products to more closely hew to user requirements. For instance, aetherAl Hema streamlines the workflow into three easy steps to avoid human error under stress, removing the burden of counting through man-machine collaboration and allowing humans to focus on important analysis and diagnostic decisions. Similarly, aetherAI Endo is engineered to function flawlessly with mainstream endoscopy video processors and a hospital's existing IT infrastructure.

GIGABYTE BRIX is a family of mini-PCs suitable for multipurpose functionality in a far-reaching spectrum of professional user scenarios, including enterprise, industry, and retail. GB-BRU7-255H represents a new breed of ultra-compact AI mini-PCs that leverage high-performance, cost-efficient AI processors to infuse workspaces with intelligence—even without cloud resources. It became aetherAI's launchpad for providing convenient and ubiquitous access to AI for endoscopy and hematology.

"Both 'carrot' and 'stick' are needed for the faster adoption of medical AI," explains Dr. Yeh. "The 'carrot' may come in the form of better policies or payment schemes. The 'stick' may be staffing shortage; but it may also be a positive, market-driven force for change, such as first movers who introduce AI-enhanced services that make consultation more effective, that revolutionize the way we receive medical care—then you will see a new industry standard centered on smart health tools."

With a view to helping to initiate smart health transformation, GIGABYTE GB-BRU7-255H boasts three key features that are conducive to the rollout of medical AI.

● Built-in AI Inference Engine for High-precision CADe on Network Edge

GIGABYTE GB-BRU7-255H combines Intel® Core™ Ultra 7 processor 255H with Intel® AI Boost Neural Processing Unit (NPU) to achieve the benefits of heterogeneous computing, transforming the min-PC into a dedicated inference engine capable of minimizing CPU load while maximizing system stability. The AI pipeline is further optimized with top-tier memory bandwidth and storage that can support continuous high-resolution video streaming without buffering or frame drops, so operators can engage in real-time medical image analysis on a fast and reliable local device.

Intel's strides in both hardware and software technology convinced aetherAI to adopt mini-PCs built around Intel's seminal microprocessors, even though it took the team some time to adjust the models from floating point operations to integer operations so they can run on the CISC architecture utilized by Intel's x86 chips. Energized by the topline processing power that came at a competitive price, aetherAI was able to accelerate the deployment of its smart health solutions.

● Versatile I/O Setup for Seamless Infrastructure Integration 

On top of connecting to inspection instruments like an endoscope host, GB-BRU7-255H can merge smoothly into existing IT infrastructure through its diverse selection of superbly compatible I/O ports. To illustrate, the USB4/USB Type-C connector can link up with top-grade storage devices to efficiently process high-resolution medical images. The 2.5GbE LAN offers stable interconnectivity with a facility's picture archiving and communication system (PACS) and hospital information system (HIS). Additional expansion slots and I/O ports can be used to turn the mini-PC into the central hub of an AI-enhanced clinical environment.

For aetherAI Endo, GIGABYTE customized GB-BRU7-255H to support M.2 expansion and 3G-SDI broadband transmissions. This enabled the device to offer plug-and-play integration with major endoscopy systems and HDMI/DP video capture cards. As a result, medical professionals can be empowered by AI to treat patients more effectively without having to enact noticeable changes to established practices.

● Low-latency On-prem Computing for Diagnostic Efficiency and Data Privacy

A pivotal advantage in fielding an all-in-one edge AI platform is that latency is virtually eliminated, ensuring speedy diagnoses that help staff efficiently deal with a large inflow of patients. Data privacy and security, another cornerstone of responsible digital health practice, are protected by processing all data on a single local device that does not require cloud support. Many of the hurdles that impede smart medicine rollout are thus resolved with cutting-edge on-prem AI computing.

GIGABYTE GB-BRU7-255H stands out in the market by packing all these perks into a compact 0.4L form factor with a modest power draw that's perfect for mobile deployment—almost like a physical AI assistant that can go into the field to provide care for patients who need it the most. GIGABYTE BRIX's thermal management design guarantees ultralow noise while in operation, making it well-suited for sustained use close to patients in clinics, hospitals, health centers, and beyond.

"An important reason we chose to work with GIGABYTE is because they are serious about nurturing their brand, which they accomplish through commitment to their customers and a hands-on approach to developing the market," states Dr. Yeh. "Whether it's GIGABYTE, GIGAIPC, or Giga Computing (which provides servers and data center solutions), the team is always ready to invest in the market and discover new opportunities with the client, which is precisely the kind of partner we need to promote Bio-ICT in the Taiwanese market and beyond."