TechFlow’s June 18 Intelligence Bureau edition said its AI Agent patrolled more than 200 global information sources across crypto, AI and technology, filtering out 99% of the noise and retaining the signals it considered most relevant. The roundup covered AI model companies, tokenized equities on-chain, exchange listings, hardware and chip supply chains, technology-company incidents, U.S. equities, macro-finance, geopolitics and new product trends.
AI sector: OpenAI losses, Anthropic comments and security concerns
In the AI and large-model section, Anthropic CEO Dario Amodei spoke openly in a recent interview about why he left OpenAI. He said he was “unable to trust a person whose behavioral pattern was disturbing and dishonest.” TechFlow described the comment as a rare public statement about internal conflict in Silicon Valley’s AI circle. The roundup also noted that community discussion viewed the remark as validating rumors about governance disorder inside OpenAI, with Reddit and video sources cited.
OpenAI’s financial position was another central item. According to the information attributed to BlockNow, OpenAI generated $13 billion in revenue in 2025, but recorded losses of as much as $38.5 billion. TechFlow wrote that revenue had surged, yet uncontrolled costs had led to a massive loss and left the company’s commercialization path under heavy pressure. Its commentary framed the figure in vivid terms, saying OpenAI had lost the equivalent of three and a half Twitter acquisition prices in one year, while Sam Altman’s idea that “AGI needs to burn money” was undergoing a harsh market pressure test.
Security and product issues also featured heavily. Microsoft Copilot was reported to have a “SearchLeak” vulnerability. Security researchers warned that the flaw could be used for data theft and could expose user emails and two-factor authentication codes. Microsoft had not formally responded in the item. DeepSeek launched a vision model, while the United States temporarily refrained from adding it to a blacklist. The roundup said the U.S. listed more than 100 companies as security risks, but DeepSeek was not included for the time being. It also noted that DeepSeek’s web-based “fast mode” showed a clear change in its chain of thought, prompting user discussion on Reuters, DeepSeek Chat and Zhihu-linked threads.
Midjourney was also included for launching a Medical vertical application. The product marks an entry into medical image generation, but TechFlow said compliance and accuracy remained disputed. On Hacker News, 523 comments questioned the ethical boundaries of medical AI generation. Taken together, the AI section placed governance disputes, high operating losses, product expansion and security risks in the same daily frame.
Crypto and Web3: Solana dominates tokenized stock trading
The main crypto and Web3 signal was the rapid expansion of tokenized stock trading on Solana. TechFlow said Solana had become the main venue for on-chain stock trading, with tokenized stocks setting the largest single-day trading-volume record on the previous day at $142 million. Of that volume, 97% took place on Solana. The roundup also said SpaceX’s tokenized stock, identified as $SPCX, accumulated $3.1 billion in trading volume over the nine days surrounding its IPO, and that Hyperliquid was becoming a new venue for pre-IPO price discovery. The sources listed for this section were Twitter/Buffalu and Twitter/Grayscale.
Exchange developments were shorter but specific. OKX announced that the RE token would be listed for spot trading, and that its pre-market futures would be converted into standard perpetual contracts. At the same time, South Korean exchange Upbit placed AERGO under a trading warning. The item cited Twitter/6551News for this exchange-related section and did not provide additional price or volume figures.
Hardware and chips: Ryzen security feature removal, HBM4E samples and AMOLED production
In chips and hardware, AMD was reported to have quietly removed a memory encryption function from consumer-grade Ryzen CPUs. According to Tom’s Hardware, the security feature was silently deleted in a new AGESA firmware release. TechFlow wrote that users, without being aware of the change, could face exposure to vulnerabilities, while AMD engineers remained silent on the matter.
AI memory supply chains were also updated. SK hynix began supplying samples of its next-generation AI memory, HBM4E, using a 12-layer stack. TechFlow characterized the AI chip race as moving into the memory side, adding that tight TSMC capacity had pushed a surge in demand for Samsung foundry services. The roundup also cited industry views that DRAM supply-chain tightness could last until 2027.
China’s display industry appeared in the same hardware block. Cailian Press information said China’s first 8.6-generation AMOLED production line had begun mass production in Chengdu, with a total investment of RMB 63 billion. TechFlow said the line was among the world’s first of that generation to enter production, and described the milestone as China’s panel industry entering the era of high-generation OLED manufacturing.
Technology companies: Apple’s RAM-cost warning, Xiaomi’s AI-rumor case and credential leaks
Among technology-company items, Apple CEO Tim Cook said RAM costs were “unsustainable,” and iPhone prices would rise. The Verge was cited for the information. TechFlow said AI functions were increasing memory requirements, and that Apple had publicly stated it would pass costs on to consumers. The roundup described this as a rare advance warning of price increases in recent years.
In China, Xiaomi’s AI-rumor case led to the arrest of four people. Xi’an police cracked a criminal case involving the use of AI to generate false information. TechFlow described it as the first publicly reported domestic criminal case involving AI-generated rumors, marking the beginning of regulatory action against generative-AI abuse. On Zhihu, the related discussion reached 580,000 views, with users focusing on the legal boundaries of AI-generated rumors.
A separate security item came from Ars Technica. A large-scale data leak exposed thousands of sensitive network credentials. The report said the scope of the leak was broad and involved multiple critical infrastructure networks. TechFlow placed this credential leak alongside the Copilot vulnerability and AMD’s removal of a security feature, presenting them as parts of a broader cost line around fast-growing AI and digital infrastructure.
Macro, geopolitics and new product signals
In the U.S. equities and finance section, NASA selected the rocket company of Eric Schmidt to carry out a Mars mission, creating competition with SpaceX. TechCrunch was cited for the information. TechFlow said the former Google CEO’s aerospace company had received a NASA order, creating the first direct challenge to SpaceX’s monopoly position. At the same time, Sequoia partner Roelof Botha joined SpaceX’s board.
Geopolitics and energy were also prominent. Trump defended the U.S.-Iran agreement by saying he did “not want to see an economic disaster” and did not want to be Hoover, who was associated with the 1929 crash. The United States and Iran reached a memorandum of understanding under which Iran agreed to dilute its enriched uranium stockpile in exchange for sanctions relief. The first batch of Iranian oil had broken through the U.S. blockade and was shipped to market. Republican Senator Bill Cassidy criticized the arrangement as “the worst diplomatic mistake in decades.” TechFlow said market concerns focused on whether changes around the Strait of Hormuz would alter the global energy landscape.
The same section said Ukraine had launched a second airstrike on a Moscow refinery within one week, while Germany sent warships to the Red Sea. CNN and Twitter/FirstSquawk were cited. TechFlow wrote that geopolitical risks were rising and that crude-oil supply chains were again under pressure. Germany’s defense minister said two warships had been dispatched to deal with military action connected to the Strait of Hormuz. TechFlow’s commentary added that Trump did not want to be Hoover, but the market had already started pricing him, moving from “Make America Great Again” to “do not let the economy crash again.”
The new products and new trends section listed three items. Lore, an open-source version-control system designed for scalability, was released and received 1,150 votes and 611 comments on Hacker News; it was described as a challenger to Git. A mother won court support in a lawsuit to inherit 87 game accounts belonging to her deceased son, and the platform’s standard-form terms were ruled invalid. TechFlow described the decision as the first domestic victory in a virtual-property inheritance case, with Zhihu views reaching 8.64 million. The full 1.0 version of The Scroll of Taiwu was also released after eight years of development, with 1.28 million Zhihu views discussing its completion.
TechFlow’s final thread tied the day’s stories together: Apple raising prices because of AI, OpenAI burning cash, Copilot’s vulnerability and AMD removing a security function all pointed to AI’s costs moving from capital markets toward consumers and security baselines. At the same time, record on-chain stock trading and Trump’s compromise with Iran to avoid economic collapse were presented as two signs of the same anxiety: when AI infrastructure costs run out of control and geopolitical risks rise, consumers and taxpayers become the ones paying the bill.