LALagrange Harga

According to Beating AI news, OpenAI has announced the latest test results of its self-developed AI inference chip, Jalapeño. The chip outperformed Nvidia's GB200 and GB300 superchips in the InferenceX benchmark tests, achieving 1.5 to 1.9 times the AI workload per watt on models such as GPT-OSS 120B, DeepSeek R1, and Kimi K2.5 1T, while reducing end-to-end latency by 1.7 to 3.6 times. OpenAI stated that Jalapeño utilizes an ASIC architecture developed in collaboration with Broadcom, specifically designed for AI inference, aiming to enhance throughput while reducing latency. The company plans to initiate small-scale deployment by the end of this year, expand production by 2027, and continue developing second and third-generation chips. However, the company emphasized that it will not completely replace chips from partners like Nvidia.

According to Beating AI News, OpenAI has announced the latest test results for its self-developed AI inference chip, Jalapeño. The chip outperformed Nvidia's GB200 and GB300 superchips in the InferenceX benchmark tests, achieving 1.5 to 1.9 times the AI workload per watt on models such as GPT-OSS 120B, DeepSeek R1, and Kimi K2.5 1T, with end-to-end latency reduced by 1.7 to 3.6 times. OpenAI stated that Jalapeño utilizes an ASIC architecture developed in collaboration with Broadcom, specifically designed for AI inference, aiming to enhance throughput while reducing latency. The company plans a small-scale deployment by the end of this year, with plans to scale up production by 2027 and continue developing second and third-generation chips. However, the company emphasized that it will not completely replace chips from partners like Nvidia.

Beating AI News reported that AI security startup Alice announced on Tuesday the completion of a $140 million funding round led by Apax Digital Funds, with participation from Samsung, SentinelOne, Maj Invest, MoreTech, and Phoenix Insurance. Following this round, Alice's total funding has reached $280 million, and Apax Digital will join the company's board of directors. Headquartered in New York and Tel Aviv, Alice, formerly known as ActiveFence, primarily provides security testing and protection for AI models and agents. The company stated that its annual recurring revenue (ARR) is nearing $100 million, with AI business growth exceeding 500% over the past two years. As AI agents increasingly gain access to databases, internal software, APIs, and the ability to execute tasks autonomously, security risks such as prompt injection, jailbreaks, data leaks, and model anomalies continue to rise. Alice currently serves AI companies including Anthropic, Google, and Cohere, claiming to cover 8 out of the top 10 AI model manufacturers globally. Alice's subsidiary, Rabbit Hole, has accumulated years of data on adversarial and harmful content, and the company states that its technology currently provides security protection for platforms covering over 3 billion users. This funding will primarily be used to expand the AI security platform, grow the research team, and scale business operations, with the company currently employing over 150 AI security researchers.