FedEx Pilots and Union Reach Tentative Agreement on 40% Pay Increase 
Pony.ai, Uber, and Verne Launch Europe's First Commercial Robotaxi Service in Zagreb 
Rubio Directs U.S. Diplomats to Use X and Military Psyops to Counter Foreign Propaganda 
Apple's Foldable iPhone Faces Engineering Setbacks, Mass Production Timeline at Risk 
Abbott Laboratories Ordered to Pay $53 Million in Premature Infant Formula Lawsuit 
China's AI Stocks Surge as Zhipu and MiniMax Hit Record Highs 
U.S. Automakers Push Back Against EU Rules Blocking American Trucks from European Market 
Kia Cuts EV Sales Target for 2030 Amid Slowing Demand and U.S. Policy Shifts 
Chalco Stock Surges as Q1 2025 Profit Forecast Jumps Up to 58% 
Annie Altman Amends Sexual Abuse Lawsuit Against OpenAI CEO Sam Altman 
Rio Tinto's California Boron Assets Attract Over a Dozen Bidders, Valued at Up to $2 Billion 
Pilots Fear Retaliation for Refusing Middle East Flights Amid Ongoing Conflict 
U.S. Disrupts Russian Military Hackers' Global DNS Hijacking Network 
OpenAI Addresses Security Vulnerability in macOS App Certification Process 
Tokyo Electric Power Attracts Major Investors Amid Billion-Dollar Restructuring Push 
Bank of America Identifies Top Asia-Pacific Semiconductor Stocks Poised for AI-Driven Growth 
MATCH Act: How New U.S. Chip Legislation Could Freeze China's Semiconductor Ambitions 
In a strategic move to enhance AI performance, OpenAI is working with TSMC and incorporating AMD chips into its infrastructure to develop custom silicon. Expected to start production by 2026, this effort could reshape the competitive landscape of AI hardware.
Custom Silicon Partnership with Broadcom and TSMC
According to insiders who spoke with Reuters, OpenAI is allegedly collaborating with Broadcom to create new bespoke silicon that can handle its massive AI inference workloads. The company has also reportedly secured production capacity with TSMC.
Reportedly, OpenAI has assembled a chip development team consisting of approximately twenty individuals, among them are lead engineers who had previously worked on Google's Tensor processors for artificial intelligence.
Nevertheless, according to The Verge, according to the current schedule, the production of the custom-designed hardware might not begin until 2026.
AMD Chips Bolster OpenAI’s Interim Setup
Moreover, OpenAI is reportedly using AMD processors in its Microsoft Azure environment, according to the sources. In a move to catch up to Nvidia, the market leader, AMD has seen its data center business quadruple in the past year, with the introduction of its MI300 chips being a major factor in this development.
In July, The Information reported that OpenAI was in talks with Broadcom and other semiconductor designers regarding the development of an AI chip.
Initial Foundry Plans Shelved Amid High Costs
Earlier this year, Bloomberg said that OpenAI was planning to establish its own network of foundries, but Reuters claims that these plans have been shelved owing to time and money constraints.
Based on the information provided, OpenAI is following the same path as other IT businesses that are attempting to control expenses and gain access to AI server hardware through the use of specially designed chips.
Competing with Industry Giants in AI Hardware
However, OpenAI might require a lot more money to catch up to Google, Microsoft, and Amazon, which are all several generations behind in their endeavors.
