The Tool Desk
Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
A February 2024 report described an order for an AI chip combining Samsung’s upcoming 2nm process, HBM3 memory and advanced packaging—but it did not name the buyer or disclose the chip, contract value or production schedule. The report was not enough to identify the customer as Preferred Networks: Samsung announced that company’s 2nm project separately, without tying it to the HBM3 order.
What did the reported deal include?
The February 2024 report said an undisclosed company had contracted with Samsung for an AI chip built on Samsung’s then-upcoming 2nm process. It described the product as a combination of 2nm logic, HBM3 memory and advanced packaging, aimed at data-center-class use. The account was a secondary report, not a Samsung announcement that publicly named a customer or set out commercial terms.
The reported combination matters because it points to an integrated accelerator design rather than a stand-alone consumer processor. The logic chip performs computation; HBM3 is stacked high-bandwidth DRAM placed close to the processor to move data quickly; and advanced packaging brings those components together in a compact package. The report did not publish a detailed package design, so it does not establish precisely how the HBM3 was connected or which packaging technology the order specified.
What remains undisclosed
- Customer: not identified in the February 2024 report.
- Chip model and specifications: not stated in that report.
- Contract value, wafer volume and delivery schedule: not stated.
- Qualification or production status: not established by the public details in that report.
Those omissions matter: an order report is evidence of a reported customer commitment, but it is not enough to establish when a finished chip would ship, how many units Samsung would make, or whether production had reached volume.
#1 Best Overall
![Arduino® UNO™ Q 4GB [ABX00173]- Hybrid Board, Qualcomm Dragonwing QRB2210 microprocessor (MPU) & STM32U585 Microcontroller(MCU), AI Vision, Voice, IoT, Robotics, Linux Debian OS, Wi-Fi 5, USB-C](https://m.media-amazon.com/images/I/51cl4xygwGL._SL160_.jpg)
- Dual-Brain Hybrid Power: Combines the Qualcomm Dragonwing QRB2210 MPU (Quad-core Arm Cortex-A53 @ 2.0 GHz CPU, Adreno GPU, AI acceleration) and the real-time, low-power STM32U585 MCU for advanced applications like object recognition, voice commands, and motion detection.
- AI & Linux Capabilities: Unlocks AI-powered vision and sound solutions; runs Linux Debian OS for coding in Python and supports the Arduino ecosystem with libraries and Sketches; quick start with Arduino App Lab.
- Advanced Features: Equipped with 4 GB LPDDR4 RAM, 32 GB eMMC built-in storage, ideal for single-board computer (SBC) mode, running multiple simultaneous high-level processes, more complex AI or ML models, extensive logs. Dual-band Wi-Fi 5 (2.4/5 GHz), Bluetooth 5.1, and high-speed headers for vision, audio, and display peripherals.
- Seamless Expansion & Connectivity: Features the classic UNO form factor for shields compatibility, an 8x13 LED matrix, and a Qwiic connector for easy expansion with Modulino nodes; power and connect via the USB-C connector.
- Intended Use & Development: The perfect platform for prototyping robotics or IoT projects, empowering innovators with a unified development experience to mix Arduino Sketches, Python scripts, and containerized AI models in a single interface.
Was the HBM3 customer Preferred Networks?
There is no evidence in the cited announcements that Preferred Networks was the unnamed buyer in the HBM3 report. Samsung separately confirmed a turnkey 2nm GAA engagement with Japanese AI company Preferred Networks, including I-Cube S 2.5D packaging for AI chiplet solutions. Samsung’s announcement described a different disclosed customer relationship; it did not say that Preferred Networks placed the earlier HBM3 order.
Samsung Foundry executive Taejoong Song called the Preferred Networks order pivotal because it validated Samsung’s 2nm GAA process and advanced-package technology for next-generation AI accelerators. That statement applies to the Preferred Networks announcement, not to the unidentified customer in the February 2024 report.
Rank #2
- Dual-Brain Hybrid Power: Combines the Qualcomm Dragonwing QRB2210 MPU (Quad-core Arm Cortex-A53 @ 2.0 GHz CPU, Adreno GPU, AI acceleration) and the real-time, low-power STM32U585 MCU for advanced applications like object recognition, voice commands, and motion detection.
- AI & Linux Capabilities: Unlocks AI-powered vision and sound solutions; runs Linux Debian OS for coding in Python and supports the Arduino ecosystem with libraries and Sketches; quick start with Arduino App Lab.
- Advanced Features: Equipped with 2 GB LPDDR4 RAM, 16 GB eMMC built-in storage, ideal to develop in PC-connected mode, running the OS, Python scripts, and basic network services (SSH) without a demanding GUI or heavy multitasking; great for lightweight AI and memory-optimized TinyML applications, needing local storage for basic OS and core libraries. Dual-band Wi-Fi 5 (2.4/5 GHz), Bluetooth 5.1, and high-speed headers for vision, audio, and display peripherals.
- Seamless Expansion & Connectivity: Features the classic UNO form factor for shields compatibility, an 8x13 LED matrix, and a Qwiic connector for easy expansion with Modulino nodes; power and connect via the USB-C connector.
- Intended Use & Development: The perfect platform for prototyping robotics or IoT projects, empowering innovators with a unified development experience to mix Arduino Sketches, Python scripts, and containerized AI models in a single interface.
How the reported order differs from later Samsung announcements
Samsung’s later public statements make the technology context clearer, but they do not retroactively reveal the 2024 buyer or turn separate agreements into one contract.
| Announcement | Process and memory | Packaging or integration | Customer and timing disclosure |
|---|---|---|---|
| Reported HBM3 order, February 2024 (secondary report) | Upcoming 2nm AI logic and HBM3; transistor architecture not stated in the report. | Advanced packaging was part of the reported combination; specific package type not stated. | Customer, contract value, volume and production schedule not stated in the report. |
| Preferred Networks, separate Samsung announcement | Samsung 2nm GAA process; the cited announcement does not identify HBM3 as part of this engagement. | I-Cube S 2.5D packaging and a turnkey approach for AI chiplet solutions. | Customer named; no connection to the reported HBM3 order was announced. |
| Samsung HBM4 collaboration with AMD, 2026 | HBM4 using 1c DRAM and a 4nm logic base die; Samsung stated speeds up to 13 Gbps per pin and bandwidth up to 3.3 TB/s. | The cited announcement concerns HBM4 collaboration; it does not establish that this is the package from the 2024 report. | Samsung described HBM4 mass production as beginning in February 2026; this is a separate collaboration. |
| Samsung–Broadcom MOU, July 2026 | HBM supply alongside 2nm-and-below foundry services; no HBM generation is specified in the cited summary. | 2.3D/2.5D packaging was included in the MOU’s scope. | Separate MOU; Samsung and Broadcom estimated the collaboration’s value at more than $200 billion through 2030. |
The table’s later figures and milestones come from Samsung’s respective 2026 statements, not from the 2024 order report. In particular, the Broadcom figure is an estimate for a later collaboration, not a disclosed value for the unnamed HBM3 order.
Rank #3
- Single core ARM Cortex-A7 32-bit core, integrated with NEON and FPU
- Built in Micro's self-developed 4th generation NPU, with high computational accuracy and support for mixed quantization of int4, int8, and int16. Among them, int8 has a computing power of 0.5 TOPS and int4 has a computing power of up to 1.0 TOPS
- Built in self-developed 3rd generation ISP3.2, supports 4 million pixels, and supports various image enhancement and correction algorithms such as HDR, WDR, and multi-level denoisin
- It has powerful encoding performance, supports intelligent encoding, adapts to save bit rates according to the scene, and saves more than 50% of the bit rate compared to conventional CBR mode, making the captured images high-definition, smaller in size, and doubling the storage space
- The design with built-in RISC-V MCU supports low-power fast startup, 250ms fast capture, and simultaneous loading of AI model library, enabling facial recognition to be completed within 1 second
What changed as Samsung’s roadmap moved from HBM3 to HBM4?
HBM3 in the report identifies the memory generation associated with that particular reported design. It should not be read as a statement about all Samsung AI products or as a current specification for later projects. Samsung’s 2026 HBM4 announcement described a different memory generation, with a 1c DRAM stack and a 4nm logic base die, and gave maximum performance figures of up to 13 Gbps per pin and up to 3.3 TB/s of bandwidth for its AMD collaboration.
Samsung’s FY2025 results also pointed to the next stage of its manufacturing plans: the Memory Business was to focus on HBM4 shipments, while the Foundry Business planned to ramp second-generation 2nm products. These roadmap statements show the direction of later work; they do not supply the missing schedule or outcome for the unnamed 2024 order.
Rank #4
- 【POWERFUL ESP32‑S3 CONTROLLER】Built‑in Xtensa 32‑bit LX7 dual‑core processor, 512KB SRAM, 8MB PSRAM, 16MB Flash for stable AI voice computing and multitask processing.
- 【Preloaded Dual AI Platforms】Comespre-installed with complete Deepseek and OpenAI voice dialogue projects.Experience intelligent voice interaction instantly. (Note: OpenAI functionality requires your own API key.)
- 【STABLE WIRELESS & CLEAR AUDIO】Integrated 2.4GHz Wi‑Fi + Bluetooth 5 (LE); dedicated audio decoding module for natural, responsive voice interaction.
- 【USER‑FRIENDLY VISUAL & PLUG‑AND‑PLAY】2” TFT‑SPI color screen shows real‑time chat; modular design, no extra wiring, ready to use after setup.
- 【FULL LEARNING SUPPORT】45 programmable GPIOs, rich interfaces, online web tutorials, free technical support for beginners & developers.
Why the combination was strategically important
For an AI accelerator, the manufacturing challenge is not only producing fast logic. The system also needs to feed that logic data at high rates, and package design has to bring processing and memory together while meeting power, performance and manufacturing constraints. Pairing a leading-edge logic process with stacked memory and advanced packaging is therefore a foundry-and-integration proposition, not simply a claim about a smaller chip.
The later Samsung–Broadcom MOU illustrates the same broad integrated-business approach at a much more fully described scale: memory supply, 2nm-and-below foundry work and 2.3D/2.5D packaging were all included. Samsung vice chairman and CEO Young Hyun Jun said AI was driving demand for integrated technologies spanning memory, logic and advanced packaging. The MOU is a separate 2026 agreement, however, and should not be used to infer the identity or terms of the earlier customer’s order.
What can readers conclude?
The defensible conclusion is narrow but useful: a February 2024 secondary report described an undisclosed customer order that combined Samsung 2nm AI logic, HBM3 and advanced packaging. It did not identify the customer, reveal commercial or production details, or establish that the buyer was Preferred Networks. Samsung’s later named engagements and HBM4 roadmap show how its integrated AI-chip offering developed, but they remain separate announcements.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
