Next-gen networking tech, sometimes powered by light instead of electricity, is emerging as a critical piece of AI infrastructure. The new era of Silicon Valley runs on networking--and not the kind you find on LinkedIn. As the tech industry funnels billions into AI data centers, chip makers both big and small are ramping up innovation around the technology that connects chips to other chips, and server racks to other server racks. Networking technology has been around since the dawn of the computer, critically connecting mainframes so they can share data. In the world of semiconductors, networking plays a part at almost every level of the stack--from the interconnect between transistors on the chip itself, to the external connections made between boxes or racks of chips.

Welcome back to, TIME's new twice-weekly newsletter about AI. If you're reading this in your browser, why not subscribe to have the next one delivered straight to your inbox? This summer, I found myself in the strange position of watching a humanoid robot try to load laundry. It squatted beside a washer-dryer unit, reached with one hand into a laundry basket that it was holding with the other, and put some clothes into the drum. But twice in a row, it dropped a piece of clothing and couldn't pick it back up. An engineer with a litter-picker grabbed the fallen cloth and sheepishly moved it behind the machine, out of my line of sight.

This article is a continuation of Broadcom's blog series: 2023 Tech Trends That Transform IT. Stay tuned for future blogs that dive into the technology behind these trends from more of Broadcom's industry-leading experts. It is clear that artificial intelligence, machine learning, and automation have been growing exponentially in use--across almost everything from smart consumer devices to robotics to cybersecurity to semiconductors. In 2023, there is no doubt that artificial intelligence and automation will permeate every aspect of our lives. As Michael Dell predicts, "Building systems that are built for AI first is really inevitable."

Broadcom Inc. announced delivery of the Trident 4C Ethernet switch ASIC, a 12.8 terabits/second security switch capable of analyzing all traffic at line rate. Trident 4C detects flow anomalies in real time as part of a rapid cyber-threat defense. Network bandwidth has been increasing rapidly and security analytics until now have been forced to rely on statistical sampling. Typically one random packet out of a thousand is analyzed, while the remaining 99.9 percent pass through unchecked. Trident 4C eliminates the need to compromise between performance and security.

Broadcom Inc. has announced expanding opportunities for organizations to gain greater value from the company's advanced AI, security, and hybrid cloud solutions with "Day One" support for IBM's new z16. Broadcom's suite of software solutions, services, and unique "beyond code" programs provide clients an advantage to succeed in an increasingly challenging business environment. "Our strategic investments position clients to exploit the z16 along with advances in AI, cybersecurity, cloud integration, and agility," said Greg Lotko, senior VP and GM, Mainframe Software Division, Broadcom. What distinguishes Broadcom is our deep investment in technology and how we work side-by-side in partnership with our clients to overcome their unique challenges and create new opportunities." As a member of the z16 Early Ship Program, Broadcom collaborated with IBM to ensure clients can capitalize on the full range of our mainframe software solutions on the new platform to drive progress toward their innovation and business goals. "Nothing can match the transaction performance of a mainframe, and the way that we manage the platform using Broadcom technology is a real differentiator for us," said Johan Bosch, executive director for iOCO Infrastructure Services. "We can deliver our services at 25 percent of the cost when measured against standalone banking environments.

Chatbots have a checkered past of often not delivering the performance their providers have promised. This is especially true in the IT service management (ITSM) and multilingual NLP spaces, where service desks found support teams deluged with complaints -- yes, about the support chatbots. Just getting English language nuance right and how enterprises communicate often require chatbots to be custom programmed with constraint and logic workflows supported with natural language processing (NLP) and machine learning. If that sounds like a science project, it is, and IT users are the test subjects. Because of their complexity, chatbots were contributing to already overflowing trouble-ticket queues.

A new GamesBeat event is around the corner! Learn more about what comes next. EdgeQ revealed today it has begun sampling a 5G base station-on-a-chip that allows AI inference engines to run at the network edge. The goal is to make it less costly to build enterprise-grade 5G access points, as well as radio units and distributed units that make up an open radio access network. The choice EdgeQ made to create a base station that can be deployed as a system-on-chip (SOC) platform also reduces the time and effort providers of wireless networks need to create physical layer software that governs all the essential protocols and features of an integrated 4G/5G network, said EdgeQ CEO Vinay Ravuri. At the core of the product, simply named EdgeQ 5G Base Station-on-a-Chip, is a specialized baseband processor RISC-V ISA that EdgeQ created by extending an open source RISC-V Open Instruction Set with more than 50 custom instructions.

BROOMFIELD, Colo., Sept. 4, 2020 -- Liqid, provider of the composable infrastructure platform, announced a collaboration with global Cloud and Data Center infrastructure leader Broadcom to deliver the first PCI-Express (PCIe) Gen 4.0 Fabric Reference Design Kit (RDK). To further drive data center adoption of the high-speed PCIe connectivity, the two companies have built a managed-fabric solution powered by Liqid Command Center orchestration software and Broadcom PEX88000 PCIe Gen 4.0 switches to enable research and development departments to quickly evaluate and accelerate deployment of Gen 4.0 solutions in the marketplace for next generation data center applications. "Edge and 5G computing, AI ML workloads, high-value HPC applications, and composable infrastructure deployments are all use cases that benefit significantly from high-speed data transfers and low latency PCIe Gen 4.0 solutions," said Jas Tremblay, Vice President and General Manager of the Data Center Storage Group, Broadcom. "Through our joint efforts, Liqid and Broadcom empower companies to deliver PCIe Gen 4.0 solutions that offer scalable performance, composable architecture, ease of management, and superior power efficiency for applications driving innovation across vertical markets." PCIe Gen 4.0 doubles the throughput over Gen 3.0, achieving bandwidth of up to 256 GT/s per port, substantially reducing transfer time for big data applications.

Broadcom released new switches and a retimer for data centers and cloud providers. The PEX88000 Switches and PEX88T32 Retimer, which are part of Broadcom's suite of peripheral component interconnect express (PCIe) switches, are designed to address the rising demand for high performance computing and storage, according to the vendor. The PEX88000 family of switches features a range from 26 to 98 lanes and ports, which Broadcom claims is "ideal for high-throughput and low-latency applications such as deep learning, high-performance computing and NVMe [non-volatile memory access] all-flash arrays." The PEX88T32 retimer is based on Broadcom's PCIe Gen 4.0 SerDes IP and designed to increase the deployment of the PCIe switches in applications that make use of artificial intelligence, machine learning, and high-performance computing. Will Townsend, senior analyst at Moor Insights & Strategy, says he expects the new products to be impactful in the white-box market with companies like Arrcus, Cumulus, and others that focus on merchant silicon solutions.

We know history repeats itself, and remembering that can prove useful in determining how interfaces that are used to control application-specific integration circuits in switches can spur SDN development. The four interfaces -- Switch Abstraction Interface, the Fast Data Plane Project, Broadcom's Software Development Kit Logical Table and the P4 programming language -- work with data plane chipsets and switch ASICs. But they also have the potential to drive SDN and machine learning growth by advancing open interfaces and programmability capabilities. To fully grasp how, however, let's first look at the history of compute systems and hard drive interfaces. Going back in time, one of the seminal moments in compute history was the standardization and adoption of the International Standard Architecture (ISA) in the 1980s.