OpenAI Unveils Aurora Analytics Tool and Jalapeño Chip for AI Applications

OpenAI has made significant strides in AI development, unveiling its smart analytics tool, Aurora, and a custom inference chip, Jalapeño. Aurora aims to help businesses understand customer data better, while Jalapeño achieves record speed and efficiency in AI inference. The chip, designed with AI-assisted tools, provides higher throughput and lower latency than existing hardware.

OpenAI plans to deploy Jalapeño within its compute infrastructure by the end of the year, challenging Nvidia's dominance in the AI hardware market. The company shared the first measured performance results of Jalapeño, which delivered more peak throughput per kilowatt and lower token latency than commercial systems.

Meanwhile, Anthropic, the creator of AI chatbot Claude, is in talks to lease a massive data center in Bastrop County, Texas, which could bring hundreds of millions of dollars in investment to the region. Perplexity has partnered with Nvidia to launch Portable Computer, a fully local AI agent with zero token costs, which can run entirely on a local device with no need for cloud connectivity.

Other notable developments include BlackBerry's QNX software being used in 275 million vehicles, with a royalty backlog of $950 million, part of which comes from robotics. Additionally, Algorhythm Holdings is evaluating strategic alternatives for its SemiCab AI logistics business after receiving multiple indications of interest for a potential sale.

Key Takeaways

['• OpenAI unveils Aurora analytics platform and Jalapeño custom inference chip for AI applications.', "• Jalapeño chip achieves record speed and efficiency in AI inference, challenging Nvidia's dominance.", '• Anthropic in talks to lease a massive data center in Bastrop County, Texas, for AI development.', '• Perplexity launches Portable Computer, a fully local AI agent with zero token costs, in partnership with Nvidia.', "• BlackBerry's QNX software used in 275 million vehicles, with a royalty backlog of $950 million.", '• Algorhythm Holdings evaluates strategic alternatives for SemiCab AI logistics business.', "• OpenAI's Jalapeño chip delivers 1.5 to 1.9 times more AI work per watt and 1.7 to 3.6 times lower end-to-end latency than comparison systems.", '• DeepSeek R1 and Kimi K2 show strong performance on Jalapeño chip.', "• The estate of Carl Sagan sues Luma AI for copyright and false endorsement over the use of Sagan's voice in an advertisement.", '• IBM develops Spyre accelerator card for AI inference, which can run entirely on IBM LinuxONE.']

OpenAI unveils smart analytics tool Aurora

OpenAI has launched its new Aurora analytics platform that uses AI to help businesses understand customer data better. The tool is part of OpenAI's compute strategy, which aims to improve the entire system of AI through an integrated approach. This approach includes data centers, chips, frontier models, and developer platforms. OpenAI's goal is to stay on the Pareto frontier, continually seeking the strongest mix of capability, speed, reliability, efficiency, and cost for each workload.

OpenAI's new chip boosts AI inference

OpenAI has developed a new custom inference chip called Jalapeño, which achieves record speed and efficiency in AI inference. The chip was designed with the help of AI-assisted tools and provides higher throughput and lower latency than existing hardware. OpenAI plans to deploy Jalapeño within its compute infrastructure by the end of the year.

OpenAI unveils Jalapeño chip

OpenAI has unveiled its custom AI inference chip, Jalapeño, demonstrating industry-leading performance and signaling a move towards greater infrastructure control. The chip was designed to optimize the entire system for throughput, latency, energy efficiency, and cost. OpenAI's strategy involves maintaining a diverse portfolio of compute partners and technologies to stay on the Pareto frontier.

OpenAI shares Jalapeño chip performance

OpenAI has shared the first measured performance results of its custom inference chip, Jalapeño. The chip delivered more peak throughput per kilowatt and lower token latency than commercial systems in the comparison. Jalapeño also performed strongly on DeepSeek R1 and Kimi K2, showing its gains extend across model families.

Jalapeño chip shows industry-leading speed

OpenAI's Jalapeño chip has demonstrated industry-leading speed and efficiency in AI inference. The chip can serve more AI work per unit of power while also returning responses more quickly. Jalapeño delivered 1.5 to 1.9 times more AI work per watt and 1.7 to 3.6 times lower end-to-end latency than comparison systems.

OpenAI challenges Nvidia with new chip

OpenAI's new custom inference chip, Jalapeño, challenges Nvidia's dominance in the AI hardware market. The chip was designed with a novel architecture that focuses on reducing data movement and optimizing algorithms to achieve low latency and high throughput. OpenAI plans to use Jalapeño to reduce inference costs for customers.

Estate sues Luma AI over use of Carl Sagan's voice

The estate of Carl Sagan is suing Luma AI for copyright and false endorsement over the alleged unlicensed use of Sagan's voice in an advertisement for its AI product Ray 3.14. The estate claims that Luma used an eight-second clip of Sagan's voice without permission.

Estate sues AI startup over ad using Carl Sagan's voice

The estate of Carl Sagan is suing AI startup Luma AI for using his voice without permission in an advertisement. The estate claims that Luma used Sagan's voice to 'dragoon him into acting as an unwitting spokesperson.' Luma AI has not responded to the lawsuit.

Perplexity launches Portable Computer

Perplexity has launched Portable Computer, a local-first build of its agentic Computer platform that runs on NVIDIA DGX Spark. The system includes a local model, inference engine, tool sandbox, and app connectors, and can run entirely offline with zero per-token cost for local steps.

Perplexity partners with Nvidia to launch Portable Computer

Perplexity has partnered with Nvidia to launch Portable Computer, a fully local AI agent with zero token costs. The system can run entirely on a local device, with no need for cloud connectivity. Perplexity aims to bring serious AI agent workloads off the cloud and onto local devices.

Anthropic to lease data center

Anthropic, the creator of the AI chatbot Claude, is in talks to lease a massive data center in Bastrop County, Texas. The project could bring hundreds of millions of dollars in investment to the region and create new opportunities for AI development.

Will AI take your job?

The increasing use of AI tools like Claude, created by Anthropic, is changing the way professionals work. A senior banker used Claude to write a report, but the director had to check it with another AI tool.

BlackBerry sees growth in robotics

BlackBerry CEO John Giamatteo said that robotics is one of the company's fastest-growing businesses. BlackBerry's QNX software is used in 275 million vehicles and has a royalty backlog of $950 million, part of which comes from robotics.

Algorhythm considers strategic alternatives

Algorhythm Holdings is evaluating strategic alternatives for its SemiCab AI logistics business after receiving multiple indications of interest for a potential sale.

DGrid AI token jumps 93%

DGrid AI's DGAI token nearly doubled in its first day of trading as the decentralized AI inference project rolled out its native token.

Evaluating skills in AI systems

Researchers have developed a framework for evaluating skills in AI systems, called Agentic Continuous Evaluation of Skills (ACES).

IBM's Spyre accelerator for AI inference

IBM has developed a new accelerator card, Spyre, for AI inference that can run entirely on IBM LinuxONE, reducing latency, security, and regulatory exposure.

Sources

NOTE:

This news brief was generated using AI technology (including, but not limited to, Google Gemini API, Llama, Grok, and Mistral) from aggregated news articles, with minimal to no human editing/review. It is provided for informational purposes only and may contain inaccuracies or biases. This is not financial, investment, or professional advice. If you have any questions or concerns, please verify all information with the linked original articles in the Sources section below.

OpenAI Aurora AI Analytics Jalapeño Chip Inference Nvidia Dominance AI Hardware Carl Sagan Estate Luma AI Copyright False Endorsement Perplexity Portable Computer Local AI Agent Anthropic Claude AI Chatbot Data Center Robotics BlackBerry QNX Software Royalty Backlog Algorhythm SemiCab AI Logistics DGrid AI DGAI Token Decentralized AI Inference Agentic Continuous Evaluation of Skills ACES IBM Spyre Accelerator AI Inference LinuxONE

Comments

Loading...