By, Creus Moreira Carlos,Founder & CEO at WISeKey

Creus Moreira Carlos

As we look towards 2024, it is predicted that advancements in Artificial General Intelligence (AGI) could reach significant milestones. AGI refers to a level of artificial intelligence that can understand, learn, and apply knowledge across a range of tasks, equivalent to the cognitive abilities of a human. Should AGI become a reality, its implications for the semiconductor industry could be profound.

The semiconductor industry would likely experience exponential growth, driven by the increased demand for powerful and efficient chips capable of supporting AGI systems. The proliferation of devices connected to the Internet of Things (IoT) will require advanced semiconductors to process vast quantities of data quickly and reliably.

At the same time, the advent of AGI will necessitate a dramatic escalation in cybersecurity measures. As systems become smarter and more autonomous, the potential for sophisticated cyber threats grows. The industry will need to adopt robust post-quantum cryptography standards to protect against threats that can exploit the vulnerabilities of classical encryption methods in a world where quantum computing is becoming more accessible.

In essence, the semiconductor industry could well be at the heart of this technological revolution, providing the essential components that drive AGI systems while also innovating at the forefront of cybersecurity to protect the vast networks these systems will operate within.

Read the rest of EDACafe Industry Predictions for 2024 -Wisekey

By Koen Verhaege, CEO, Sofics

Koen Verhaege

Navigating Challenges and Embracing Opportunities in 2024

The challenging year 2023 is finally wrapping up… Predictions were that it would be a transition year towards a better 2024. So, let’s welcome 2024!

First, let’s look back. 2022 was notable, albeit for the wrong reasons: warehouses were being filled to brace against new shortages. In 2023, these overstocks created a steep drop in production. Consider the foundry capacity utilization numbers: they are near 80% today, compared to being at all-time highs, above 100%, last year.

The key factors shaping 2024 will likely be geopolitics and the subsequent influx of tax dollars, as well as, the evolving role of Artificial Intelligence in our workplaces.

Geopolitics

Geopolitical tensions will be fought and lost through the use of dollar monsoons – not by war. All (wannabe) powers are releasing tsunamis of dollars (euros, won, yen…) to bring semiconductor production “home”. Politicians are sharpening their scissors to cut ribbons at new manufacturing plants. Many if not all of these plans will fail, at least they will fail to create autonomy or independence.
Read the rest of EDACafe Industry Predictions for 2024 -Sofics

By Coby Hanoch, CEO, Weebit Nano

Semiconductors in 2024: Predictions from an Emerging Memory Perspective

Coby Hanoch

In semiconductors, 2023 was a mixed bag, with some applications like automotive and industrial growing, while overall the industry contracted. Looking ahead, analysts agree that we will see significant growth in 2024 as the industry recovers. The SIA forecasts a 13.1% increase in sales in 2024, and IDC is even more bullish, forecasting 20.2% revenue growth for the semiconductor industry in 2024.

At Weebit, a leading provider of ReRAM, the non-volatile memory (NVM) technology that is set to replace flash memory in the coming years, this growth is a positive sign, since nearly every electronic device needs NVM. Here I will outline some of the industry dynamics in 2024 that could driver further growth.

AI will get even edgier.

We can all agree that 2023 was the year of AI, and 2024 will be even more so. AI technologies are proliferating in every industry, with solutions in the cloud and increasingly at the edge where devices require local brains with power-efficient solutions. In 2024, we will see an increasing number of edge AI designs, particularly TinyML, moving to monolithic integration in 28nm and 22nm. Integrating memory on-chip in an advanced process node can save cost and power, and enhance security. This is important as AI/ML models continue to increase in size and sophistication.

Since embedded flash memory can’t scale below 28nm, what’s needed for these integrated devices is a different embedded NVM that can do the same level of inference as SRAM or DRAM but at extremely low power and cost. This is where ReRAM comes in – used not only for code storage, but also to store the synaptic weights needed for artificial neural network (NN) calculations. As much of the power consumption needed for NNs is related to data movement between a system’s computing elements and memory modules, integrating dense, low-power NVM like ReRAM closer to the computing elements can minimize power and latency. In 2024, we will see more designs moving in this direction.

Read the rest of EDACafe Industry Predictions for 2024 – Weebit Nano

By Mark Himelstein , CTO, RISC-V International

Mark Himelstein

What are the top five trends for the RISC-V open standard ISA in 2024?

This year has been a great one for the RISC-V ecosystem. Our global community of more than 4,100 members from 70 countries across the globe helped to support the continued growth and adoption of the open standard RISC-V ISA. There are now more than 13 billion – yes that’s billion with a B – cores on the market. And we’re not stopping! RISC-V International now hosts more than 75 technical working groups driving progress and advancement on RISC-V standards, software development and tools. We also ratified 16 specifications including the ratification of RISC-V Profiles specification version 1.0 with RVA20, RVI20, and RVA22, RISC-V Cryptography Extensions Volume II and the RISC-V IOMMU Architecture Specification.

We also hosted three anchor events in 2023—RISC-V Summit North America, RISC-V Summit Europe, and RISC-V Summit China. At each event, RISC-V enthusiasts, supporters, adopters, designers, and more joined to address the latest in all things RISC-V. We had a record number of attendees at each event, culminating with more than 2,000 registrants at RISC-V Summit China. Our biggest event yet!

But we are not done. The industry can expect so much more from the RISC-V community in 2024, with key technology advancements, new member products and innovations, and more.

Read the rest of EDACafe Industry Predictions for 2024 – RISC V

By Richard Preston, Hardware Lead, CacheQ Systems

Richard Preston

Technical Innovations, AI integration News and Unanticipated Market Trends in 2024

If 2024 is anything like 2023, it will be full of technical innovations, more news on AI advancements and up-trends in the semiconductor market.

Under the technological innovation category, the semiconductor industry will see more specialized hardware chips for specific workloads such as Large Language Models (LLMs) inferencing, speech processing, image and pattern recognition, and the continued pursuit of autonomous driving and smarter robotics. We will see an increased need for custom compilers to enable software development for these specialized chips. CPUs, GPUs, and FPGAs will continue to serve specific markets, but pioneers will create custom ASICs to differentiate or improve product performance.

Conjecture about AI integration and semiconductor design will continue far into 2024. We can expect more software developers will use AI-generated code in their final products. The potential is also there for AI to auto-generate and self-deploy code for the most agile and responsive applications. The day when AI is permitted to go beyond informing a human, and control operations without human intervention is getting closer. Until then, there are many industries where AI can be an invaluable companion. Some industry experts allude to a day when AI can be an effective doctor, having been trained on a vast amount of information on known medical conditions, symptoms, and treatments. Perhaps 2024 will be the year medical professionals adopt greater use of AI as a companion tool.

Read the rest of EDACafe Industry Predictions for 2024 – CacheQ

By Dr. Ashish Darbari, Founder and CEO, Axiomise

Dr. Ashish Darbari

RISC-V, Chiplets, Formal Verification Adoption in 2024

2024 will offer up a cornucopia technological innovation, AI integration and perhaps unexpected market trends.

It’s hard to argue with the success and adoption of the RISC-V ISA. In 2024, open-source RISC-V implementations for high-performance compute will continue to take wings with interconnect-based SoCs for AI accelerations. Commercial verification tools will begin to emerge in 2024 to support the RISC-V hardware development ecosystems. It will be followed by a robust verification design flow that will quickly become an industry standard.

The semiconductor industry will also see increased adoption of chiplet for AI-based hardware. As the industry moves on to 5nm or less nodes, chiplet would be a key. A chiplet market trend could be the emergence of chiplet suppliers whose only product is a chiplet-based design.

Automotive functional safety though mostly dominated by Arm architecture so far would see new RISC-V players building processors and chipsets that would offer new features while ensuring compliance with ISO26262.

Security continues to be a challenge for hardware design, and here AI will play an interesting role in learning from threat patterns, and we believe adversarial AI would stand out.

Read the rest of EDACafe Industry Predictions for 2024 – Axiomise

By Aleksandr Timofeev,   founder and CEO of POLYN Technology

Aleksandr Timofeev

The trend of integration and deployment of edge processors in IoT, wearables, and consumer electronics applications will accelerate this year. Edge processors based on standard digital architecture such as ARM or RISC-V will be replaced by more energy-efficient specialized edge chips. The expansion of edge chips based on in-memory computing will also accelerate as the advantage of analog edge chips over digital edge chips becomes more apparent thanks to greater energy efficiency, higher performance, and lower price.

A companion trend for 2024 will be the creation of new hybrid digital-to-analog systems and the first attempts to integrate memristors as elements of edge chips.

Watch for proofs of concept targeting integration of ultra-low-power analog processors for sensor data extraction with low power digital MCUs to perform classification. The digital twin role in this integration enables refining the extraction details prior to chip production. These proofs of concept will demonstrate maximum efficiency of the analog chip structure for MAC operations both in power and latency.

Read the rest of EDACafe Industry Predictions for 2024 – POLYN Technology

By Michal Siwinski, CMO, Arteris

Emerging Trends and Future Directions in the 2024 Technology Landscape

Michal Siwinski

As 2024 unfolds, a technological revolution is unfolding. Advancements in artificial intelligence (AI), automotive, compute architecture and industrial electronics are reshaping our world. These evolving trends are pivotal in steering us toward a future with unparalleled opportunities.

The Rise of AI and NLP

The acceleration of AI, particularly at the intersection of Deep Learning (subset of Machine Learning) and Natural Language Processing (NLP) where there is a lot of Generative AI and Large Language Models (LLMs) progress, is evident in the rapid progression from ChatGPT 3.5 to GPT-4 with the expected release of GPT-5 by early 2025. This advancement is paralleled by the remarkable engagement with ChatGPT. The current user base has expanded to over 180 million, marking an 80% increase in just eight months. This represents one of the fastest broad adoption rates of a consumer application.  Generative AI and LLMs have finally transcended academic realms, increasingly emerging as practical tools that are reshaping industries and demonstrating significant potential for innovation, whether directly or indirectly experienced by users of end applications.

 Source: Arteris, Inc

AI’s impact extends to training and inference in enterprise computing and data centers. Companies like Tenstorrent are at the forefront, harnessing AI for innovative compute solutions. The deployment of trained models for inference is crucial in applications ranging from automotive advanced driver-assistance systems (ADAS) to industrial robotics, disrupting traditional markets and introducing new paradigms in how we approach technology. This rapid technology development across multiple platforms signifies the transformative era of AI.

Read the rest of EDACafe Industry Predictions for 2024 – Arteris