Broadcom Inc. is a global technology company that designs, develops, and supplies a broad range of semiconductor and infrastructure software solutions, serving various industries including data centers, networking, and telecommunications.
Broadcom CEO Hock Tan says capital may become a bigger constraint on AI growth than shortages of chips, memory or power, while Intel CEO Lip-Bu Tan points to the cost of capital and the need for returns.
Broadcom agreed to lend Anthropic up to $42 billion to finance computing infrastructure. The facility could cover about one third of Anthropic's $125.2 billion commitment to lease TPU capacity.
OpenAI offered to invest about $100 million in Hugging Face before NVIDIA agreed to acquire the platform for roughly $13 billion.
Tensordyne has detailed Napier, an AI inference system built on a 3 nm processor and logarithmic arithmetic. The chip has taped out, and the company projects one rack could do the work of nine Nvidia and Groq racks.
Broadcom reported $16.7 billion in third quarter AI semiconductor revenue, up 221% from a year earlier. It expects the figure to reach $21.7 billion in the fourth quarter.
DriveNets has introduced two new AI networking platforms, the 2600SL and 2601S, built on Broadcom’s Tomahawk 6 ASIC, offering up to 102.4Tbps capacity and designed for large-scale AI infrastructure deployments.
OpenAI and Broadcom have unveiled Jalapeño, a new inference processor designed specifically for large language models. The chip was developed in nine months and will be deployed at gigawatt scale in data centers starting in 2026.
OpenAI's plan to produce custom AI chips with Broadcom, aimed at reducing reliance on Nvidia, is facing an $18 billion financing hurdle as Broadcom seeks a purchase commitment from Microsoft for 40% of the chips.
Google is negotiating with Marvell Technology to design two new chips for AI inference, including a memory processing unit to complement its tensor processing units and a new TPU built for running AI models more efficiently, according to The Information.
Meta has announced an expanded partnership with Broadcom to co-develop multiple generations of its Meta Training and Inference Accelerator (MTIA) chips, aiming to strengthen its AI infrastructure across apps and services.
Anthropic has announced Project Glasswing, a cybersecurity collaboration with major technology and financial institutions to use its Claude Mythos AI model for detecting and fixing software vulnerabilities across critical systems worldwide.
OpenAI and Broadcom have entered a multi-year collaboration to co-develop and deploy 10 gigawatts of custom AI accelerators and networking systems between 2026 and 2029. The systems will use Broadcom’s Ethernet-based connectivity solutions and be installed across OpenAI and partner data centers.
OpenAI is collaborating with Broadcom to produce its own AI chips, aiming to reduce its reliance on Nvidia. The chips are expected to be used internally starting in 2026.
The data center semiconductor market is projected to exceed $500 billion by 2029, driven by global AI demand and compute buildouts, according to a report by Futurum.
Vectara has been selected by Broadcom to deliver a conversational AI customer service solution, enhancing support for Broadcom's enterprise clients.
Broadcom has launched its Jericho4 ethernet fabric router, designed to enhance distributed AI computing across data centers, according to Reuters.
Synopsys has announced a successful interoperability milestone with Broadcom's PEX90000 series switch, demonstrating PCIe 6.x IP solutions at PCI-SIG DevCon 2025.
Corning has partnered with Broadcom to enhance AI data center processing by supplying optical components for Broadcom's Bailly co-packaged optics technology, as announced in a press release.
Daily report covering major AI developments and industry news, with both top stories and complete market updates
