AI news flash: OpenAI launched the more cost-effective GPT-5.6 model versions Luna and Terra, significantly reducing usage costs compared to previous versions, aiming to let enterprises deploy AI workflows at scale in production environments with a lower barrier. These two models continue OpenAI’s push on the price-performance frontier…
AI news brief: Chinese optical communications supplier Zhongji Innolight listed on the Hong Kong Stock Exchange on July 30, with shares dropping on debut and breaking below the issue price right at open. The IPO raised HK$53.4 billion (about US$6.8 billion), making it the second-largest new listing in Asia this year. Zhongji Innolight is a key AI-related equipment supplier, primarily making optical communications components used for high-speed data transmission in data centers and AI servers…
AI News Brief: This news piece is very short, only an introductory paragraph with limited substantive details. The Trump administration once proposed banning Chinese open-weight AI models from use in the US, but before China’s tech industry and officials could take a stance, Silicon Valley had already stepped forward to express opposition. The report points out that the Trump administration and US tech companies hold different views on how to respond to the rise of Chinese AI models: the government favors restrictive measures like bans to block them, but Silicon Valley players including Nvidia believe that amid growing AI bubble concerns and market caution over growth momentum, the US tech ecosystem itself still depends on open-weight models to sustain innovation and expansion. A hasty ban on Chinese open-weight models could end up hurting the development foundation and competitiveness of America’s own AI industry. The original summary only provides event background and an overview of opposing positions, without further detailing the specific proposal contents, timeline, or the full arguments from Silicon Valley’s opposition. Please refer to the original link for details.
AI News Brief: A Bitcoin (BTC) mining stock hedge fund is seeking capital amid an AI stock selloff, the UK’s Financial Times reports. Situational Awareness, a hedge fund founded in 2024 by former OpenAI researcher Leopold Aschenbrenner, has approached investors and lenders for fresh capital after suffering major losses during the recent AI stock selloff. According to the Wall Street Journal, the fund had roughly $20 billion in AUM as of June 8. Citing insiders and a July 24 investor letter, the FT noted the fund also offered some investors the option to buy back portfolio assets, though the size of losses and capital raised were not disclosed. The letter said the fund was up 439% net of fees through June, but the FT pointed out the fund used leverage to amplify positions, which magnified losses during the July AI crash. Aschenbrenner argued the selloff created attractive investment opportunities. Cointelegraph previously reported that the fund made heavy bets on power and data centers supporting AI, including Bitcoin miners pivoting to AI compute. March SEC filings showed the fund held about $1.11 billion across 7 Bitcoin mining stocks, including IREN, Core Scientific, Riot Platforms, and CleanSpark. Aschenbrenner published a series of AGI essays in mid-2024, predicting AGI machines would surpass college graduates’ capabilities by the end of the decade. Cointelegraph reached out to Situational Awareness for comment but had not received a response at the time of publication.
AI news brief: Zhongji Innolight (中際旭創) is a Chinese optical communications component supplier that mainly produces AI-related key equipment. The company debuted on the Hong Kong Stock Exchange on July 30, with shares falling on the first day. The IPO raised HK$53.4 billion (about US$6.8 billion), making it Asia’s second-largest listing so far this year. The bell ceremony was attended by Chairman and President Liu Sheng along with other executives. The weak first-day stock performance reflects market concerns about…
AI News Brief: Japan has posted a trade deficit for two consecutive months, while the Philippines’ monthly deficit continues to widen—Thursday’s data from both governments lays bare a divergence in Asia’s trade landscape. Rising crude oil prices are squeezing trade balances and currencies across multiple Asian economies, and the numbers show a clear split: some economies are riding tech export growth hard enough to offset the energy cost shock, while others can’t fully absorb it, and their trade deficits are widening as a result. Japan and the Philippines fall squarely in the latter camp—both saw their trade accounts deteriorate in June, reflecting the dual pressure on trade balances and local currencies from rising oil prices layered on top of uneven tech export momentum. The original summary does not provide specific deficit figures, year-over-year changes, or which tech export items are driving other Asian economies’ performance, so details will need to be verified against the original source.
AI News Brief: Teleperformance, a France-based business process outsourcing (BPO) group, revealed to Nikkei Asia that the company plans to open its first site in Vietnam before the end of this year. The move primarily targets demand related to AI data services, as AI technology is reshaping how the call center industry operates, prompting the company to adjust its strategic positioning in response to structural market shifts. The original summary only outlines high-level plans without further details on site scale, investment amount, staffing, or specific AI data service content — see the original link for full details.
AI News Brief: SoftBank Corp. and NTT are partnering with dozens of Japanese companies and research institutions to drive a nationwide AI and data sharing initiative, aiming to build a foundational technology platform for cross-industry data integration. This move is seen as part of Japan’s strategy to catch up with the US and China in AI. Since Japan lags behind the US and China in AI model development, the government and industry hope to leverage the scattered data resources held by various companies and institutions to build AI infrastructure with scale and differentiation…
AI News Brief: SK Hynix’s AI chip boom benefits exporters like South Korea, but the US, despite investing hundreds of billions in AI infrastructure, sees limited economic growth. The report points out that while US data center construction pushes up semiconductor prices, most of the profits flow to chip exporters like SK Hynix rather than showing up in US GDP growth. In short, the AI infrastructure boom’s short-term lift to the US economy is weaker than expected, with upstream chip-exporting countries capturing the biggest dividends.
AI News Brief: OpenAI found that adjusting two API settings can nearly triple GPT-5.6’s scores on the ARC-AGI-3 benchmark while also improving efficiency. The first setting retains reasoning across multi-turn tasks so the model continues its reasoning chain instead of rethinking from scratch; the second enables compaction, which compresses prior reasoning before handing it to subsequent steps to slash context length and compute without losing key info. Combined, they show that managing a model’s memory and reasoning continuity is itself a direct lever for boosting both performance and cost — not just swapping models or scaling compute.