Track Hyper | Can Intel's Dragonfly turn the tide?

Author: Zhou Yuan / Wall Street News

Between general-purpose GPUs and dedicated ASICs, FPGAs, which specialize in this field, have finally "gained freedom" as Altera, originally part of Intel's Programmable Solutions Group (PSG), becomes an independent company.

On January 10th, Altera announced on platform X, "Today, we proudly welcome a milestone moment as we officially raise the Altera flag as an independent FPGA company. With passion, we will drive future innovations with agility and focus, shaping the next era of FPGA technology at altera.com."

This "new company" will maintain a strategic partnership with Intel but will have greater freedom in decision-making in the FPGA field compared to when it was part of Intel PSG.

The relationship between general-purpose GPUs, FPGAs, and dedicated ASIC chips can be simply described: FPGAs sit between the two, being semi-custom compared to standard GPUs, making them particularly suitable for adapting to rapidly changing technological demands without bearing the high costs of ASIC development.

Standardized GPUs have relatively fixed architectures, which may limit efficiency for certain specific non-graphical and non-highly parallel computing tasks.

FPGAs are programmable hardware devices whose internal logic blocks and wiring resources can be reconfigured through programming. This flexibility allows FPGAs to customize hardware circuits according to specific application needs, achieving high performance in certain specific computing tasks.

The advantages of ASICs are evident: they are highly optimized for specific algorithms or tasks, achieving extremely high performance, efficiency, and power consumption ratios; however, their disadvantages are equally significant: once designed and manufactured, they are difficult to modify, thus lacking flexibility, leading to higher design and manufacturing costs, and are usually suitable for specific applications in large-scale production.

FPGAs are mainly used in applications characterized by high customization and flexibility, such as high-speed signal processing (communications), real-time control and signal processing of industrial automation equipment like motors and robots, and data centers that need to accelerate certain specific workloads; GPUs are widely used in graphics processing as well as deep learning and inference; fields that require extremely high performance and power consumption with relatively fixed functions belong to ASICs.

Altera and Xilinx were once the two largest companies in the FPGA field. In 2015, Intel acquired Altera for $16.7 billion. This was Intel's largest acquisition in history.

In the GenAI technology field, FPGAs offer great flexibility, and the cost of adding new instructions and data formats is relatively low. In addition to customizing AI applications, FPGAs can also be used for simulation testing on chips before the launch of AI processors.

This simulation is built on the foundation of Altera's FPGA AI Suite, integrated frameworks like TensorFlow and PyTorch, allowing developers to build optimized solutions compatible with industry standards.

Since Altera has been spun off from Intel's original system, the company will inevitably bundle technical supply with Intel. For example, Altera will use Intel's foundry services (IF) However, this technological collaboration is not exclusive. Altera can also use other foundries, such as TSMC or Samsung Electronics, for chip manufacturing.

The CEO of Altera is Sandra Rivera, and the COO is Shannon Poulin.

Since the standard GPU manufacturer NVIDIA is so successful, and the ASIC-focused Broadcom is also shining brightly, will Altera, which is deeply engaged in FPGA, have a brilliant performance after becoming independent from the Intel system and having the decision-making power that an independent company should have?

Intel should have expectations for this, so at the right time, Altera is likely to push for an IPO.

Currently, Intel, which is deeply mired in financial difficulties, has swiftly replaced its former CEO (Pat Gelsinger) and simultaneously opened the door for the independence of various potential businesses under its umbrella.

In addition to Altera, this includes Intel's perceptual computing business, RealSense.

Recently, there have been reports that Intel plans to operate RealSense as an independent company under Intel Capital, expected to be completed in the first half of this year. In response, Intel stated, "The decision to spin off RealSense is unrelated to its recent financial difficulties."

RealSense is both a business of Intel and a technological framework used for robotics (as shown in the image).

Intel's RealSense technology aims to change the way robots perceive and interact with their environment.

RealSense technology includes hardware and software integration, with hardware mainly consisting of depth cameras equipped with visual processors; the software includes the open-source RealSense SDK 2.0, which supports all Intel RealSense camera products, is compatible with various operating systems, ROS (Robot Operating System), and programming languages, and is easy to integrate with solutions from third-party software providers.

Intel is attempting to set new industry standards for robotic intelligence and vision using RealSense technology, which features high-resolution imaging and depth perception.

Looking at Intel's attitude towards RealSense technology, it is difficult to clearly say what Intel's plans are for it.

In 2012, Intel began investing in perceptual computing research and established a $100 million perceptual computing fund; in early 2014, Intel officially renamed the perceptual computing technology to RealSense and launched it in the market.

In 2016, Lenovo decided to integrate Intel's RealSense technology into multiple system plug-in options at CES and launched a monitor supported by Windows Hello, equipped with a RealSense camera.

In 2020, despite facing numerous challenges, Intel's RealSense applications in robotics, digital signage, 3D scanning, and other fields still saw some development. The bionic robotic dog released by Xiaomi uses RealSense depth sensing technology In August 2021, Intel announced the closure of its RealSense computer vision department; by February 2024, Intel "restarted" the application of RealSense technology in the field of humanoid robots through investments in companies like Figure AI.

Currently, the latest development regarding RealSense technology is that on September 24, 2024, Intel launched the "RealSense" depth camera module D421 that applies this technology.

Now, does Intel want to spin off the RealSense business to become an independent company? Is this again focusing on the future potential of this technology in the field of intelligent robotics?

Intel denied rumors that the spin-off of the RealSense business into an independent company was due to financial difficulties, stating a very eloquent phrase — this aligns with Intel's commitment to "incubating cutting-edge, disruptive technologies and businesses to validate customer demand and market acceptance."