ByteDance’s world model has reportedly debuted, generating navigable 3D worlds from a single sentence.

Polygon Foundation CEO Sandeep Nailwal announced in a post that the Polygon protocol burned 100 million POL tokens two weeks ago, accounting for roughly 1% of the total token supply, with funds for the burn sourced from network-generated transaction fees. Currently, an additional 22 million POL tokens (about 0.2% of total supply) are set to enter the next phase of token burns. Furthermore, Polygon has not yet fully disclosed the impact of OMS on POL, with related details to be released in the near future. Last week, the annualized trading volume processed by OMS exceeded $1 billion for the second consecutive week.

According to data from Token Terminal, the total market capitalization of tokenized stocks has hit a record high of $3.9 billion. Among this, the three major blockchains — BNB Chain, Ethereum Mainnet, and Solana — collectively hold around 70% of the market share.

Beating AI News Flash: The Wall Street Journal, citing people familiar with the matter, reports that Amazon is in talks to acquire AI startup Decart. The potential deal is valued at around $7 billion, though it has not been finalized. Decart specializes in AI chip optimization software, which enables models to run faster and more cost-effectively across different chips, with support for NVIDIA GPUs, Google TPUs, and Amazon Trainium. Founded in 2023, the startup has seen multiple twists in its acquisition rumors. In August, Israeli media outlet Calcalist reported that NVIDIA was close to acquiring Decart, but negotiations were halted after other buyers intervened. SpaceX, owned by Elon Musk, was also speculated as a potential buyer, though no credible reports have confirmed a formal bid. Later, Anthropic considered acquiring Decart for around $6 billion and conducted due diligence, but by September, Anthropic decided to exit the talks, with the reason remaining undisclosed. The potential Amazon deal’s valuation is higher than Decart’s $4 billion funding valuation in May. Amazon and Decart already have an existing partnership: in May, Decart announced Amazon was its strategic client, and the two had run the real-time video model Lucy 2 on Amazon’s self-developed Trainium3 chip. Decart has also developed Oasis, a tool for generating interactive virtual scenes. If the acquisition is completed, Amazon will gain Decart’s cross-chip optimization technology.

According to GMGN market data, Solana ecosystem meme coin BATON briefly exceeded $20 million in market capitalization, hitting an all-time high. It is currently priced at $18.22 million, with a 24-hour gain of 41.11%. BATON takes its name from Baton Corporation, the predecessor of Pump.fun, and is viewed by the community as a meme coin tied to Pump.fun’s early corporate identity. BlockBeats reminds users that most meme coins lack practical use cases, exhibit extreme price volatility, and investors should exercise caution.

JPMorgan Chase analysts said AI stocks currently make up up to half of the weight of Asian indices, and the sector is projected to drive 70% of the region’s profit growth in 2027. In the report, analysts noted that Asia’s AI exposure is concentrated mainly in the hardware sector, and supply tightness is expected to persist through at least 2028, which will grant related companies significant pricing power. They added that the areas facing the most severe shortages include advanced logic chips, chip packaging, semiconductor equipment, Ajinomoto Build-up Film (ABF) substrates, high-end copper clad laminate materials, high-bandwidth memory (HBM), and high-speed optical components. JPMorgan projects that by 2028, the entire AI ecosystem will generate net profit exceeding $1 trillion and free cash flow of $700 billion. The firm maintains an "overweight" rating on tech stocks, the Taiwan (China) stock market, and South Korea’s stock market.

Beating AI News: Microsoft CEO Satya Nadella published a lengthy article advocating that AI models be treated as part of internal enterprise security risk management. Even non-malicious models can make mistakes or be exploited by attackers. Especially when models can access sensitive data and perform critical operations on behalf of users, enterprises cannot rely solely on security promises from model vendors. He called for separating models from the Harness (Agent execution framework) that schedules tasks. Which data a model can read and which tools it can call must be controlled by an external system. Each critical operation should leave tamper-proof, human-auditable records. He also demanded transparency in model chain-of-thought (the displayed reasoning process), while acknowledging that this content may not be reliable. The most stringent requirement is to assume models may be compromised from deployment and implement isolation measures. Authorized personnel must be able to pause or shut down an Agent at any time during task execution, equivalent to an "emergency brake". Models cannot audit their own behavior; enterprises must also continuously test for faults and attacks, and disclose incidents in a timely manner. Microsoft announced this month that its MXC isolation tool is now officially available on Windows. It restricts the files and networks Agents can access via external policies.