Open side-bar Menu
 EDACafe Editorial
Roberto Frazzoli
Roberto Frazzoli
Roberto Frazzoli is a contributing editor to EDACafe. His interests as a technology journalist focus on the semiconductor ecosystem in all its aspects. Roberto started covering electronics in 1987. His weekly contribution to EDACafe started in early 2019.

2021 Year in Review

 
January 12th, 2022 by Roberto Frazzoli

Let’s start 2022 by quickly recapping some of the major events and trends of 2021, which was another year of the so-called ‘silicon renaissance’ era – and, unfortunately, another year of Covid-19 pandemic.

Record numbers, chip shortage, new fabs, geopolitical tensions

As we all know, the semiconductor industry and its ecosystem thrived in 2021, due to strong demand from multiple end markets. In fact, the demand exceeded the global fab capacity, generating a severe chip shortage that especially hit carmakers. According to market research firm Omdia, in the third quarter of 2021 the semiconductor market surpassed $150 billion. Record numbers were also reported by the Electronic System Design Alliance for the EDA industry, and by Semi for the equipment and material market. As for this latter area, 2021 was an exciting year because of several announcements related to the planned construction of new fabs, such as Samsung’s plant in Texas or Intel’s factories in Arizona. Unfortunately, geopolitical tensions also increased in 2021, with China vowing to pursue reunification with Taiwan.

Intel’s new course

One of last year’s biggest events was the start of a ‘new course’ at Intel with the appointment of Pat Gelsinger as the new CEO in early 2021. Since then, Gelsinger has not been sitting of his hands: the Santa Clara tech giant was in the news multiple times during the year, with a series of announcements. They included two new US fabs and the launch of a foundry business (March); several organizational changes and new appointments (June); the renaming of process nodes and a future transistor architecture called RibbonFET (July); several new products (August); the intention to take Mobileye public (December).

Nvidia-Arm deal still in limbo

The biggest semiconductor deal announced in 2020, the Nvidia acquisition of Arm, did not go through in 2021. Concerns of lessening of competition started to build up in early last year in the form of complains from other company’s CEOs, which were then followed by official investigations undertaken by authorities around the word: announcements came from the UK in August, the European Union in October, and the USA in December. This latter initiative, in particular, is a law enforcement action from the Federal Trade Commission meant to block the acquisition.

AI acceleration attracting investments

Not surprisingly, in 2021 the development of AI acceleration chips/IP/systems continued to attract investments. EDACafe reported about eleven financing rounds in this area, totaling approximately 1.2 billion dollar – and for sure that was not an exhaustive coverage. AI companies getting funding in 2021 included Graphcore, NeuReality, Recogni, Kunlun, SambaNova, Mythic, Sima.ai, Edge Impulse, Deep Vision, Axelera, Houmo.ai.

AI startups, new AI chips, new AI systems

Besides funding, in 2021 the liveliness of the AI space obviously showed up in other announcements, too. Several companies either emerged from stealth or announced upcoming new products: among them AlphaIC, Atlazo, ArchiTek, Expedera, EdgeCortix, Innatera, Roviero. More AI related news came from large and/or already established companies: IBM introduced a 7-nanometer AI chip optimized for low-precision training and inference; Cerebras unveiled its Wafer Scale Engine 2 (WSE-2) chip boasting 2.6 trillion transistors; Google announced the next generation of TPUs, the TPU v4, powered by the v4 chip; Tesla developed its own training chip and system; Intel introduced Loihi 2, its second-generation neuromorphic research chip.

AI in EDA

The use of deep learning technologies in EDA gained more ground in 2021. Among the many relevant events, the publication of the paper from Google researchers about deep learning- based chip floorplanning; Xilinx adding machine learning to Vivado design suite; AI-enabled EDA tools developer Motivo raising $12 million financing; the introduction of Cadence’s Cerebrus Intelligent Chip Explorer, a machine learning-based tool; DARPA awarding funds to Julia Computing to accelerate simulation of analog and mixed-signal circuit models using machine learning and AI techniques.

Risc-V gaining momentum

In 2021, Risc-V kept gaining momentum. Relevant announcements included new chips expanding the presence of this open source ISA to multiple markets, such as the ultra-low power 64-bit Risc-V core from Micro Magic or the “supercomputer-on-a-chip” developed by Esperanto Technologies. At least three significant players, in different ways, chose to adopt Risc-V to replace preexisting processing architectures: MIPS, Imagination, and – in a specific application – Intel. The Santa Clara tech giant adopted Risc-V for the new generation of Nios, the soft processor for its FPGAs. The open-source environment kept growing, too: for example, Alibaba announced that it will open the source code of its Risc-V-based XuanTie IP core series.

Technologies

Technologies obviously kept advancing in 2021, laying the foundations for the next process nodes. In May IBM announced the development of the world’s first chip with 2-nanometer nanosheet technology. Across the semiconductor industry, the concept of moving the power delivery network to the wafer’s backside gained momentum – with equipment vendors such as Applied Materials supporting this innovation, Intel preannouncing the future adoption of this architecture, and research institutes such as Imec developing specific solutions. Packaging technologies also evolved, due to the increasing importance of chiplet-based devices. Samsung Electronics, for example, developed a new 2.5D packaging solution suited to high-performance devices that need to integrate a large number of silicon dies. System-level packaging and assembly also made news in 2021 as critical technologies for power delivery and cooling, especially for high-performance computing equipment such as AI training systems. Tesla’s training system, for example, employs a custom voltage regulator module reflowed directly onto a fan-out wafer.

Acquisitions

The acquisition wave continued throughout 2021 in the entire semiconductor ecosystem, although with a sharp difference from 2020: the lack of mega-deal announcements such as Nvidia-Arm or AMD-Xilinx. The number of transactions, however, was significant: EDACafe reported about more than forty acquisition deals in 2021. As far as the EDA industry is concerned, Cadence acquired Numeca; Silvaco acquired Polyteda Cloud; Siemens Digital Industries Software acquired OneSpin Solutions, Fractal Technologies and the proFPGA family from German company Pro Design; Ansys acquired Phoenix Integration and Zemax. Many more deals, of course, concerned chipmakers. Two of them attracted the analysts’ attention as they involved experienced processor designers: Qualcomm’s acquisition of CPU startup Nuvia, and Intel’s recruitment of the design team of Centaur Technology.

SiC and GaN scenario

Some of last year’s announcements showed the liveliness of the SiC and GaN scenario, with several companies repositioning themselves to compete in this market. Qorvo acquired UnitedSiC; Soitec acquired NovaSiC; Foxconn bought a Macronix SiC fab; STMicroelectronics started manufacturing 200mm SiC wafers. In terms of process technologies, in 2021 researchers demonstrated the feasibility of integrating GaN CMOS logic with GaN power devices.

More events and themes

Obviously, many more industry announcements or research works from 2021 deserve attention. Among them, in no particular order: companies that went public or are planning to go public (Achronix, Arteris IP, GlobalFoundries); Renesas entering the low-cost, low-power FPGA market; more hyperscalers – such as ByteDance – likely starting to design their own chips; new ways to exploit the potential of x86 architectures in AI training tasks; the use of machine learning to prefetch data from memory.

This concludes our quick recap of some of 2021 events and themes – as the world hopes to get back to normal in 2022.

Logged in as . Log out »




© 2024 Internet Business Systems, Inc.
670 Aberdeen Way, Milpitas, CA 95035
+1 (408) 882-6554 — Contact Us, or visit our other sites:
TechJobsCafe - Technical Jobs and Resumes EDACafe - Electronic Design Automation GISCafe - Geographical Information Services  MCADCafe - Mechanical Design and Engineering ShareCG - Share Computer Graphic (CG) Animation, 3D Art and 3D Models
  Privacy PolicyAdvertise