AI Stocks & Markets

Today's Stories

1. 1

   OpenAI and Broadcom unveil LLM-optimized inference chip

   OpenAI and Broadcom unveil LLM-optimized inference chip

2. 2

   OpenAI and Broadcom jointly designed a custom AI chip to reduce OpenAI's dependence on Nvidia and lower infrastructure costs.

   OpenAI unveiled a custom chip that it designed with Broadcom to power its AI infrastructure. The chip is intended to handle the computational demands of running OpenAI's models and services. Building its own silicon allows OpenAI to reduce reliance on Nvidia's processors and potentially lower the capital expenses tied to AI infrastructure. Hyperscalers (large cloud providers) and AI companies increasingly view custom chips as a way to improve cost efficiency and performance for their specific workloads.

   The timeline and scale of deployment for this custom chip remain to be seen. OpenAI's success with in-house silicon could influence how other AI firms approach their infrastructure strategy.

3. 3

   AMD Named Current Company to Beat in Gartner® AI Vendor Race

   AMD Named Current Company to Beat in Gartner® AI Vendor Race

4. 4

   OpenAI unveils its first custom chip, Jalapeño, made with Broadcom to reduce reliance on Nvidia and handle surging demand for AI inference.

   OpenAI and Broadcom on Wednesday announced Jalapeño, OpenAI's debut custom silicon chip designed to run AI models in production. The companies designed the chip in nine months and plan initial deployment by the end of 2026, with scaling through 2027 and into the first half of 2028. OpenAI says it 'cannot get compute fast enough' and Broadcom's CEO confirmed that demand from his six largest customers is 'simply insatiable.' By designing its own chips alongside the broader computing stack, OpenAI aims to serve more AI capability with greater efficiency, reducing its dependence on Nvidia's expensive GPUs and addressing a critical supply constraint.

   Jalapeño is an ASIC (application-specific integrated circuit), which industry experts say is less flexible than Nvidia's GPU but also less expensive and customizable for specific AI tasks. The companies ultimately aim to build enough capacity to require 10 gigawatts of power, signaling massive infrastructure growth over the coming years.

5. 5

   Micron forecast signals strong AI-driven chip demand will persist through 2027, as memory shortages keep pushing prices higher.

   Micron Technology forecast fourth-quarter revenue of $50 billion（約8兆円）, plus or minus $1 billion（約1600億円）—well above Wall Street's $43.58 billion（約7兆円） estimate—and posted third-quarter results of $41.46 billion（約6.6兆円）, surpassing expectations of $35.85 billion（約5.7兆円）. The company's adjusted earnings per share came in at $25.11, beating the $20.78 estimate. Explosive growth in generative AI has made high-bandwidth memory (HBM) chips critical for large data centers, turning Micron—the only U.S.-based maker of high-end memory chips—into a crucial supplier for AI infrastructure. Demand for HBM chips far outstrips production capacity, and analysts expect supply constraints to persist for the next two to three years, leaving consumer electronics makers scrambling for conventional memory as makers prioritize AI-driven orders.

   Micron expects fourth-quarter capital expenditure of around $10 billion（約1.6兆円） (versus analyst expectations of $8.89 billion（約1.4兆円）) and projects tight supply conditions to persist beyond calendar 2027. CEO Sanjay Mehrotra stated the company expects to increase capital returns while investing heavily in infrastructure expansion to meet soaring demand.

6. 6

   Cerebras, OpenAI's chip designer, saw its stock plummet 20% after reporting lower profit margins ahead, signaling Wall Street's skepticism about the company's path to profitability despite its debut earnings report.

   Cerebras posted its first earnings report on Wednesday, then issued guidance for full-year margins that fell short of expectations, prompting a sharp 20% stock decline. The margin guidance suggests the company is facing pressure to balance growth in data center chip sales against manufacturing and operational costs, a critical test for any AI hardware maker trying to prove it can be profitable at scale.

   Investors will be monitoring whether Cerebras can stabilize or improve those margins in coming quarters, as profitability will likely determine whether it can compete effectively against better-capitalized rivals and retain investor confidence.