In a market driven by narrative rotation, being first doesn't mean being right—only those who survive matter.

The market is shifting from "storytelling" to "auditing the books," and only those equipped with comprehensive risk management tools can navigate the market's turbulence.

Trading is migrating from legacy systems to new infrastructure, and there is no turning back.

The growth story of AI optical interconnects hits the accounting books of a convertible bond

Key Takeaways: After Meta's acquisition of Manus was blocked by Chinese regulators, the company announced it would resume independent operations and initiate a data cleanup process. On the same day, former Alibaba Qwen lead Lin Junyang founded Pragmatik Labs in Shanghai, focusing on digital and physical world agent research. Both events point to the AI competition shifting focus from model capabilities to agents' actual task execution ability.Key Elements:To meet regulatory requirements, Manus will delete user data generated on and after December 29, 2025. The backup window closes before August 23, with services resuming on August 25. Unaffected users need not take any action.The acquisition deal between Manus and Meta was valued at approximately $2 billion (with a potential total value of $2.5 billion). Announced on December 29, 2025, it was later prohibited under China's foreign investment security review, and the two parties completed operational separation in May.Manus saw significant growth before the acquisition: it raised $75 million from Benchmark in 2025 (at a valuation of approximately $500 million). Eight months after product launch, its annualized recurring revenue exceeded $100 million, with over 147 trillion tokens processed cumulatively and more than 80 million virtual computers created.Manus founding team background: Founder Xiao Hong (graduate in Software Engineering from Huazhong University of Science and Technology, previously founded Nightingale Technology, serving over 2 million business users), co-founder Ji Yichao (previously founded Peak Labs), and co-founder Zhang Tao (previously worked at ByteDance and LightYear.AI).During the unwinding of the deal, the founding team explored raising approximately $1 billion to buy back shares, with early investors ZhenFund and Hony Capital participating in discussions. After resuming independent operations, user data will continue to be stored in the United States and Singapore.Pragmatik Labs, founded by Lin Junyang (former head of Alibaba Qwen technology, stepped down on March 3, 2026), focuses on digital agents (knowledge work, enterprise operations) and physical agents (embodied intelligence), emphasizing reasoning, tool invocation, feedback learning, and coordinated action capabilities.The name Pragmatik Labs derives from the linguistic term "pragmatics," reflecting the philosophy that "AGI must solve real-world problems." Its research direction aligns with Lin Junyang's view of evolving from "training models" to "training agents."

CoreWeave营收翻倍，扩张仍靠融资

The US CFTC requires Kalshi to continue operating its prediction market in New York, disregarding related litigation filed by the New York Attorney General.

Rocket Lab delivered record quarterly revenue and backlog, with the Space Systems business as the primary growth driver, while launch services saw a sequential decline due to revenue recognition timing. The company also raised its next-quarter revenue guidance above market expectations, but continued heavy investment in the Neutron rocket's first flight and production ramp-up has pushed next-quarter operating loss guidance beyond what the market had originally anticipated.

On the Eve of the IPO, How Tech Upstart Wang Xingxing Led Unitree's Market Value to Surge 17,900-Fold

Strategy CEO称公司出售200万美元比特币后BTC当周下跌4%，出售2.16亿美元后BTC上涨6%。

The market begins repricing the AI computing power financing chain

If superintelligence ends up being controlled by a few institutions, all of humanity loses.

Over 40 institutions urge AI labs to open frontier model access to Bitcoin security defenders.

Those who have planted trees know best the direction in which they grow.

Buffett's choice essentially says: in the second half of the AI trade, certainty is worth more than flexibility.

When the cost per inference drops and more businesses start calling on AI, the total volume of calls may actually increase instead.