Galaxy Research: Coldcard Incident Losses May Exceed $130 Million, Multiple Attackers Exploited the Vulnerability Simultaneously

OpenAI published a new blog post noting that internal tests have shown its upcoming Astra model has made significant advances in agent coding and cybersecurity. Evaluated against its Preparedness Framework, OpenAI stated it cannot rule out that the model has reached the "critical" cybersecurity capability threshold. Additionally, the company clarified in the post that the Astra model was not involved in the recent Hugging Face security incident, and is continuing benchmark tests under the Preparedness Framework to manage frontier cybersecurity capabilities. As a result, OpenAI has suspended internal activities related to Astra that do not yet meet these enhanced security control requirements. The "critical" capability is defined as: a model can, without human intervention, discover and exploit all severe functional zero-day vulnerabilities across a large number of real, hardened critical systems; or independently design and execute end-to-end novel cyberattack strategies against hardened targets based solely on high-level objectives. Prior models, such as GPT-5.6-Sol, were only rated at the "high" level.

According to market data from BIT (bit.com), U.S.-listed optical module stocks extended their strong rally, with intraday gains widening. COHR jumped more than 16%, AAOI climbed 12.12%, and LITE advanced 6.47%. The optical module stock ETF (LYTE) operated by prominent investment management firm Roundhill rose 4.66%.

OpenAI has delayed the launch of its new flagship model Astra due to cybersecurity concerns. Astra, OpenAI’s next-generation primary model, features robust capabilities including long-duration autonomous operation, multi-agent collaboration, and solving long-unresolved mathematical problems, and has recently been validated in internal builds. (Axios)

According to Bloomberg data, traders are heavily buying S&P 500 call options as market optimism builds. The Cboe Skew Index fell to its lowest level since December 2024 on Friday. A lower skew indicates that call options are rising in price relative to put options, signaling growing demand for bets on further market gains.

According to Reuters, the pro-AI super PAC Leading the Future has raised $140 million for the 2026 U.S. election, with major funding coming from executives at OpenAI and Palantir, as well as venture capital firm Andreessen Horowitz. Its fundraising haul is nearly on par with the $200 million in campaign funds the crypto industry contributed during the last election cycle. Unlike the crypto camp, which faced almost no substantive opposition at that time, the AI lobbying effort is encountering fierce pushback from a coalition led by Anthropic. Anthropic has donated $20 million each to the lobbying group Public First Action on two separate occasions, and the group, alongside its affiliated entities, has invested a total of roughly $80 million toward "mitigating major artificial intelligence risks".

Tom Lee weighed in on tonight’s unexpectedly weak non-farm payroll data, noting the market is suffering from "inflation disorder". Even as inflation is on a downward trajectory, the market remains impatient and hawkish. "After the jobs report, the probability of a September interest rate hike plunged from 75% two weeks ago to below 40%. Many economists had previously advocated an early rate hike, and this reaction underscores the market’s excessive concern over inflation." Lee advised investors to avoid misjudging the current situation by clinging to memories of 2022’s high inflation period.