TS-2888X
AI-Ready NAS

Engineered for Machine Learning for Model-based Services

The TS-2888X packs everything required for machine learning to help organizations quickly and easily implement AI applications.

Huge Storage Potential
Data is the key to AI, and the TS-2888X provides the potential for huge capacities to remove the storage bottleneck from AI development.


Ready-to-use AI Environment
QNAP’s QuAI computing platform reduces the time and complexity of implementation and management.

Powerful Computing
The TS-2888X supports multiple graphics cards to empower data-intensive machine learning.

Secure & Private
Public cloud training is not suitable for confidential or sensitive data. Keep full control over all your data with the TS-2888X!

Comprehensive Backup Solution
Flexible backup tools and snapshots protect your data while also allowing efficient disaster recovery


The Core. The Power. The Lead.

The Intel® Xeon® W processors in the TS-2888X provide up to 18 cores/36 threads and up to 4.5 GHz with Intel® Turbo Boost Technology 2.0, combined with up to 512 GB DDR4 ECC RDIMM 2666 MHz memory to deliver rapid workload loading and processing.

Quickly build, train and optimize AI models with QuAI
QNAP’s exclusive QuAI enables data scientists and developers to quickly build, train and optimize machine-learning models running on QNAP NAS. Backed by high-performance machine-learning algorithms, the ready-to-use AI environment enables a range of open-source tools and frameworks (including Caffe, TensorFlow, and more) for the AI workflows readily available to you. The combination of system and software helps reduce time and complexity for setup and maintenance of AI environments.


Comprehensive support for AI frameworks

U.2 SSDs offload data preprocessing workloads
Data must be preprocessed before it can be provided to machine-learning algorithms. Data preprocessing can consume a lot of memory, which has the potential to impact system performance. To maintain high performance while tackling data preprocessing you can install up to four U.2 SSDs in the TS-2888X. By configuring a RAID 0 for all four U.2 SSDs, you can obtain maximized bandwidth and higher capacity for preprocessing workloads, leading to improved quality, performance, and higher cost- efficiency.

Verify deep learning inference using OpenVINO™
The Intel® OpenVINO™ (Open Visual Inference & Neural Network Optimization) toolkit helps to streamline and accelerate deep learning inference, bringing computer vision and deep learning capabilities to edge devices smoothly. With the heterogeneous support for inferencing pre-trained models on Intel® architecture hardware targets (CPU/GPU/FPGA/VPU), the OpenVINO™ toolkit allows you to optimize and verify the trained models developed by the TS-2888X before inferencing, and economically distribute vision solutions leveraging AI.

Effortlessly collect massive training data with Download Station
Machine learning requires a large set of data feeds, and Download Station can help to organize downloading tasks for massive training data. For example, more than 9 million images are in the Open Images Dataset v4 (20 TB complete sets, and 600 GB for competition purpose in 2018), and you can create download tasks using Download Station to collect those images automatically and on a scheduled basis.

IOPS-optimized storage infrastructure to boost AI workloads
Even with the best graphics cards at hand, the bottleneck for AI model training tends to be determined by IOPS as data sets grow and the training data comprises of many small data sets (usually more than 10 TB). The TS-2888X supports SSD caching to accelerate IOPS performance and reduce latency for storage volumes, providing many benefits for the metadata-intensive, small file I/O requirements of AI workloads. Qtier™ technology can also empower the TS-2888X with auto tiering to help strike a balance between performance and storage.

Assure data safety of every moment, at any time
Record the state of your system at any moment with snapshots. You can easily keep track of the state of your data, trained models and system at any time, and mitigate disastrous losses of trained data/models. In the event of malware attacks or unwanted changes, your stored data can be quickly reverted and restored to a specific point in time with snapshots to ensure continuity.

Easily back up and synchronize your data
Creating AI-scale data sets takes a long time, and losing them can be catastrophic. The TS-2888X supports Hybrid Backup Sync - QNAP's reliable and cost-effective backup & disaster recovery solution - to assist organizations in routine backup jobs. Hybrid Backup Sync consolidates backup, restoration and synchronization functions (using RTRR, rsync, FTP, CIFS/SMB), allowing you to easily synchronize data across various cloud storage, arrange scheduled backup tasks for onsite/off-site backups to ensure double protection, and to quickly restore data if needed.

Making AI accessible to your business blueprint
The TS-2888X integrates AI-focused hardware and software to reduce the time and complexity for implementing and managing AI tasks, making it the ideal AI solution with great cost-efficiency for most organizations and industries.
Autonomous Vehicles Precision Healthcare Security & Intelligence Smart Manufacturing Consumer Retail Life Sciences Research Language Processing Predictive Analytics Financial Services Fraud Detection Supply Chain Logistics

CPU options : Intel® Xeon® W-2195 18-core 2.3 GHz Processor (burst up to 4.3 GHz)
Intel® Xeon® W-2175 14-core 2.5 GHz Processor (burst up to 4.3 GHz)
Intel® Xeon® W-2145  8-core 3.7 GHz Processor (burst up to 4.5 GHz)
Intel® Xeon® W-2133  6-core 3.6 GHz Processor (burst up to 3.9 GHz)
Intel® Xeon® W-2123  4-core 3.6 GHz Processor (burst up to 3.9 GHz) CPU Architecture 64-bit x86 System Memory options : 512 GB LRDIMM DDR4 ECC (8 x 64GB)
256 GB RDIMM DDR4 ECC (8 x 32GB)
128 GB RDIMM DDR4 ECC (4 x 32GB)
64 GB RDIMM DDR4 ECC (4 x 16B)
32 GB RDIMM DDR4 ECC (4 x 8GB) Maximum Memory 512 GB (8 x 64GB) Memory Slot 8 x DDR4 RDIMM/LRDIMM Flash Memory 4GB (Dual boot OS protection) Drive Bay 8 x 3.5-inch + 20 x 2.5-inch Drive Compatibility 3.5-inch bays: 3.5-inch SATA hard disk drives 2.5-inch SATA hard disk drives 2.5-inch SATA solid state drives 2.5-inch bays: U.2 slot: 2.5-inch U.2 NVMe Gen3 x4/x2 solid state drives SATA slot: 2.5-inch SATA solid state drives Gigabit Ethernet Port (RJ45) 4 10 Gigabit Ethernet Port 2 x 10GBASE-T PCIe Slot 8 Slot 1: PCIe Gen3 x8 or x4 (CPU) Slot 2: PCIe Gen3 x4 (CPU) Slot 3: PCIe Gen3 x8 or x4 (CPU) Slot 4: PCIe Gen3 x4 (CPU) Slot 5: PCIe Gen3 x8 or x4 (CPU) Slot 6: PCIe Gen3 x4 (CPU) Slot 7: PCIe Gen3 x4 (PCH) Slot 8: PCIe Gen3 x4 (PCH) USB 2.0 port 4 USB 3.0 port 6 Audio Input 2 x 3.5mm microphone jacks (dynamic microphones only) Audio Output speaker, 3.5mm line out jack (for amplifier or headphone amplifier) Form Factor Tower LED Indicators 3.5-inch HDD, 2.5-inch SSD, 2.5" U.2 SSD, Status, LAN, USB port status Buttons Power, Reset, USB Copy Dimensions (HxWxD) 612, 374, 448 Weight (Net) 26.7 Weight (Gross) 31.46 Operating temperature 0 - 40 °C (32°F - 104°F) Relative Humidity 5-95% RH non-condensing, wet bulb: 27˚C (80.6˚F) Power Suply Unit 2000W PSU, 200-240Vac, Fan 1 x 9 cm fan (12V DC) 9 x 12 cm fan (12V DC) Sound Level 33.6 dB(A)
Note: Use only QNAP memory modules to maintain system performance and stability. For NAS devices with more than one memory slot, use QNAP modules with identical specifications. Warning: Using unsupported modules may degrade performance, cause errors, or prevent the operating system from starting. 

* Sound Level Test Environment: Refer to ISO 7779; Maximum HDD loaded; Bystander Position; Average data from 1 meter in front of operating NAS.

Designs and specifications are subject to change without notice.