OpenAI is making significant strides in both hardware and software as it partners with Broadcom and TSMC to develop its first custom chip, marking a major step toward in-house chip manufacturing.
This collaboration, which includes scaling back OpenAI’s initial foundry ambitions, highlights the company’s intent to enhance AI processing power while reducing reliance on external suppliers.
Simultaneously, OpenAI is rolling out a new feature on ChatGPT’s web platform, a search tool for chat history.
This addition enables users to easily retrieve previous conversations, improving accessibility and user experience across its popular AI chatbot platform.
OpenAI Builds First Chip With Broadcom and TSMC, Scales Back Foundry Ambition
OpenAI is working with Broadcom and TSMC to build its first in-house chip design to support its artificial intelligence systems, while adding AMD chips alongside Nvidia chips to meet its surging infrastructure demands, as per Reuters.
The OpenAI, the fast-growing company behind ChatGPT, has examin a range of options to diversify chip supply and reduce costs.
OpenAI consider building everything in-house and raising capital for an expensive plan to build a network of factories known as “foundries” for chip manufacturing.
The OpenAI has drop the ambitious foundry plans for now due to the costs and time needed to build a network, and plans instead to focus on in-house chip design efforts, according to sources, who request anonymity as they were not authorize to discuss private matters.
OpenAI’s strategy, details here for the first time, highlights how the Silicon Valley startup is leveraging industry partnerships and a mix of internal and external approaches to secure chip supply and manage costs like larger rivals Amazon, Meta, Google and Microsoft.
As one of the largest buyers of chips, OpenAI’s decision to source from a diverse array of chipmakers while developing its customize chip could have broader tech sector implications.
OpenAI, which help commercialize generative AI that produces human-like responses to queries, relies on substantial computing power to train and run its systems.
As one of the largest purchasers of Nvidia’s graphics processing units (GPUs), OpenAI uses AI chips both to train models where the AI learns from data and for inference, applying AI to make predictions or decisions based on new information.
OpenAI has been working for months with Broadcom to build its first AI chip focusing on inference, according to sources.
Demand right now is greater for training chips, but analysts have predict the need for inference chips could surpass them as more AI applications are deployed.
Broadcom helps companies including Alphabet unit Google fine-tune chip designs for manufacturing and also supplies parts of the design that help move information on and off the chips quickly.
This is important in AI systems where tens of thousands of chips are strung together to work in tandem.
OpenAI is still determining whether to develop or acquire other elements for its chip design, and may engage additional partners, said two of the sources.
The OpenAI has assemble a chip team of about 20 people, led by top engineers who have previously built Tensor Processing Units (TPUs) at Google, including Thomas Norrie and Richard Ho.
As per sources, by Broadcom, OpenAI has secure manufacturing capacity with Taiwan Semiconductor Manufacturing Company to make its first custom-designed chip in 2026.
They said the timeline could change.
As of now, Nvidia’s GPUs hold over 80% market share.
But shortages and rising costs have led major customers like Microsoft, Meta, and now OpenAI, to explore in-house or external alternatives.
OpenAI’s plan use of AMD chips through Microsoft’s Azure, first reported here, shows how AMD’s new MI300X chips are trying to gain a slice of the market dominated by Nvidia.
AMD has projected $4.5 billion in 2024 AI chip sales, following the chip’s launch in the fourth quarter of 2023.
Training AI models and operating services like ChatGPT are expensive.
OpenAI has projected a $5 billion loss this year on $3.7 billion in revenue, according to sources.
Compute costs, or expenses for hardware, electricity and cloud services needed to process large datasets and develop models, are the company’s largest expense, prompting efforts to optimize utilization and diversify suppliers.
OpenAI has cautious about poaching talent from Nvidia because it wants to maintain a good rapport with the chip maker it remains committed to working with, especially for accessing its new generation of Blackwell chips, sources said.
OpenAI Rolling Out Chat History Search Feature on ChatGPT Web
The OpenAI announce a new feature for its artificial intelligence (AI) chatbot ChatGPT.
The new feature will allow users to search through their chat history to quickly find relevant information from past conversations.
This is a significant improvement from the current workaround which requires users to scroll through the entire history to find the desired conversation.
As of now, the feature is only available on ChatGPT’s web portal and is being roll out to the paid subscribers of the platform.
In a post on X (formerly known as Twitter), the official handle of OpenAI announce the new feature.
We’re starting to roll out the ability to search through your chat history on ChatGPT web.
— OpenAI (@OpenAI) October 29, 2024
Now you can quickly & easily bring up a chat to reference, or pick up a chat where you left off. pic.twitter.com/YVAOUpFvzJ
The ability to search through past chats has been a highly request feature by users, as currently, it gets increasingly time-consuming to find relevant information.
With the chat history search feature, users can now find a new magnifying glass icon on the side panel of ChatGPT web, which is place on the left side of the window.
You can tap on this icon to open a text field where they can type keywords specific to find a past conversation.
The text field also shows the recent chats to let users quickly pick one of them without needing to type anything.
The feature is currently being roll out to ChatGPT Plus and Team users.
OpenAI also highlight that the Enterprise and Edu subscribers will get access to the chat history search feature within a week.
But, those on the free tier of the platform will have to wait till the next month before they see this feature.
Notably, those using the Android and iOS apps of ChatGPT or are on the Windows or Mac desktop apps will not get this feature for the time being.
OpenAI did not reveal if the feature will also be extend to these platforms in the near future.
OpenAI has quietly adding new features to ChatGPT to keep it ahead of rivals.
Other recent feature adds to the web version of the platform is web search capability.
This is different from the native capability of the GPT-4o model, which can scour the Internet.
With the search the web feature, available as a globe icon within the text field, the AI only uses websites to find the relevant information and displays it with sources for each new piece of information.
In functionality, this is somewhat similar to what Perplexity AI offers.
