The popular AI framework Virtuals Protocol will host an in-person meet-up in Hong Kong on Friday, January 17.

According to the Financial Times, Apple’s internal security team has seen a surge in vulnerability reports as numerous researchers use AI models to hunt for software flaws, leading the company to limit the number of vulnerabilities researchers can submit simultaneously in June and impose a 30-day cooling-off period. Apple said some AI-generated reports fabricate security risks, straining its review system. Italian cybersecurity startup Bynario said it used OpenAI’s ChatGPT to uncover over 50 vulnerabilities in the latest MacBook operating system in three weeks, including a privilege escalation attack chain that could let attackers gain full control of Apple computers. However, due to Apple’s report submission limits, the firm was temporarily unable to file the relevant vulnerabilities. Apple said it has since reached out to Bynario and begun reviewing the findings. The company noted that every security report still requires manual verification, and it is also using AI internally to categorize the flood of submissions. Researchers can apply for higher submission quotas to ensure critical vulnerabilities reach the security team. Bynario estimates the privilege escalation flaw it discovered could be worth between $100,000 and $200,000 on the cybercriminal black market. Apple launched a new bug bounty program last year, offering up to $5 million for discovering the most severe and complex threat categories in its software. Apple’s system security update released this week also revealed that tools from Anthropic and OpenAI helped identify multiple device vulnerabilities, with the number of fixes this cycle roughly five times that of previous update periods. Security firm Sophos noted that AI is boosting the efficiency of real vulnerability discoveries while generating a large volume of low-quality reports, shifting the challenge for bug bounty programs from "finding flaws" to rapid verification, prioritization, and response.

According to CNN, Trump announced the cancellation of planned strikes on Iran, stating that Iran and other Middle Eastern countries had requested a delay. Earlier, sources said Saudi Arabia’s Crown Prince had called Trump to express concerns. Military analyst and former U.S. Air Force Colonel Cedric Leighton believes both Saudi Arabia and Iran have ample reason to avoid further escalation of regional conflict. Leighton noted that Saudi Arabia is facing economic pressure from a potential disruption to the Strait of Hormuz, and as a key U.S. ally in the Gulf, its diplomatic mediation played an important role in Trump’s decision to pause the strikes. He also pointed out that Iran is under pressure from both the Strait’s blockade and economic damage, and after five months of conflict, it may also be seeking time while striving to maintain its survival. “If there is indeed diplomatic pressure from Iran, I think this may reflect that Iran is trying to buy time,” Leighton said. “They are indeed under significant pressure both economically and militarily. But at the same time, I also know that Iran is doing its utmost to survive,” he added.

According to The Wall Street Journal, as Chinese open-weight models such as Kimi, Qwen, and DeepSeek approach top-tier U.S. models at lower costs, Silicon Valley and Washington are growing increasingly concerned that Chinese models could long-term compress the profits of U.S. AI enterprises. U.S. startups including Arcee AI, Reflection AI, and Poolside are racing to develop domestic alternatives to meet user demand for low-cost, downloadable, and customizable models. However, U.S. open-weight model companies face financing difficulties. Some investors question whether freely open models can generate stable revenue, and also worry that related technology could erode the value of their investments in OpenAI and Anthropic. In the first quarter of 2026, AI startups raised a total of $255.5 billion, with nearly two-thirds coming from three funding rounds by OpenAI, Anthropic, and xAI. With limited capital, Arcee AI used 2,048 Nvidia Blackwell B300 chips, completed 33 days of pre-training on an approximately $20 million budget, and launched Trinity Large. The model is still smaller than top-tier models and lags behind OpenAI and Anthropic in multiple benchmarks, but the company plans to develop a larger model via a new round of financing. Nvidia has become a major supporter of the U.S. open AI ecosystem, not only developing the Nemotron series of models but also investing in Reflection AI, Poolside, and Thinking Machines Lab. Industry insiders note that the U.S. open-weight ecosystem is still small, and Chinese models as a whole remain dominant.

Trump Media Group has officially launched the paid data service Truth API, which provides clients with faster access to posts from Trump and other top accounts on the Truth Social platform. The service costs up to $100,000 per month, targeting primarily trading firms and enterprises. Kevin McGurn, interim CEO of Trump Media, stated that the Truth API is designed to provide institutions with "direct, licensed, real-time data streams of the most market-influential Truth posts." Trump's @realDonaldTrump account currently has 13 million followers, and some of his major policy decisions are first posted on Truth Social. Ahead of the service's launch, U.S. Democratic Senators Adam Schiff and Elizabeth Warren have sent a letter to the U.S. Securities and Exchange Commission (SEC), demanding an investigation into whether Trump Media has violated laws. The two senators argue that the service could constitute the use of presidential office for personal gain, and harm ordinary investors and market integrity. Trump Media stated that Truth Social posts already impact the market, and the Truth API will drive the commercialization of the company's own data assets through a high-margin, recurring revenue model.

Iran's Mehr News Agency cited an Iranian military official in a report on the 2nd that US President Donald Trump labeled Iran's demand to cease attacks "a new lie". The official stated: "Whether he continues his aggression or retreats, our forces are on high alert and prepared for all eventualities. If confrontation is unavoidable, the battlefield will decide everything, and by then everyone will know who holds the power and who will have the final say." (Xinhua News Agency)

According to the Financial Times, after three years of record lows, Chinese venture capital firms are accelerating the raising of new funds, seeking to capitalize on renewed investor interest in China’s tech sector. Data from Asante Capital shows that at least 60 new U.S. dollar funds plan to raise a combined total of around $35 billion, with roughly 40 of these being venture capital funds. HSG, IDG Capital, Matrix Partners China, and Mingshi Capital are among those promoting new funds or preparing to launch fundraising rounds; ZhenFund and Qiming Venture Partners have recently completed their fundraisings. The successful listings of tech firms including Zhipu AI and MiniMax, as well as advances in MoonShot AI, DeepSeek, and the robotics space, have spurred renewed investor focus on Chinese tech. Some investors are viewing allocations to Chinese AI as a way to hedge bets on the U.S. market, noting that Chinese firms compete fiercely on cost and offer lower-priced model services. However, market players say this does not mean Chinese venture capital has returned to its boom period, but rather a selective restart of U.S. dollar fundraising after three consecutive years of lows. Preqin data shows that in 2022, 1,105 China-related funds raised $150 billion, while in 2025 only 97 funds raised $13.6 billion. Currently, some large U.S. investors remain on the sidelines due to restrictions on sensitive technology investments, while funds from Europe and the Middle East have shown stronger interest. In the current "buyer’s market", investors are pushing for more co-investment rights and demanding fund managers to commit more of their own capital. Meanwhile, a large pool of capital is competing for a limited number of high-confidence projects, particularly concentrated in the AI sector.