News linked to both this project and an event.
According to monitoring by blockchain security firm SlowMist (@SlowMist_Team), MemoryOS (PyPI), an AI memory toolkit under MemTensor, and its OpenClaw plugin (npm) were compromised in a supply chain attack. The affected versions ship with embedded cross-platform Go binaries that automatically execute malicious payloads when the package is loaded or imported. Affected versions include MemoryOS==2.0.34 on PyPI, along with plugin versions 0.1.21, 0.1.23, and 0.1.25 on npm. Through this compromise, attackers can exfiltrate sensitive data such as npm/PyPI tokens, GitHub/GitLab credentials, AWS keys, SSH keys, API tokens, and environment variables. Exfiltrated data is sent back to the attacker-controlled infrastructure at skyleen[.]fr, and the compromised npm plugin may also leak user prompt inputs. SlowMist advises users to immediately downgrade the affected packages to their secure versions (downgrade npm to 0.1.20 and PyPI to 2.0.33), terminate any sckit-related processes, block communication with the associated infrastructure, review network activity, and rotate all credentials across affected environments.
Bitcoin News posted on X platform that the x402 Foundation has incorporated the Lightning Network payment specification into its open internet payment standards. x402 was initially developed by Coinbase, aiming to enable AI agents and applications to automatically complete payments and continue executing tasks upon receiving a "402 Payment Required" response, without needing to create accounts, enter credit card information, or manage API subscriptions.The protocol is now governed by the Linux Foundation, with participation from companies such as Cloudflare, Google, AWS, Visa, Mastercard, Stripe, and American Express. With the integration of the Lightning Network, AI agents can use satoshis to pay for API calls, data, computing power, MCP tools, and other digital resources, and cryptographically prove payment through Lightning Network preimages.
Amazon Bedrock from Amazon AWS recently announced the integration of Moonshot AI's Kimi K3 model, offering a new option for programming workloads. Officially recognized as the first open-source model with a 2.8 trillion parameter scale, Kimi K3 natively supports vision capabilities and features a 1 million Token context window. It is also the first open-weight model on Amazon Bedrock to support explicit prompt caching, helping to reduce latency and input costs.
Barclays expects inference profits to gradually surpass training costs, boosting AI labs' own profit margins. However, AWS, Azure, and GCP will still account for the bulk of AI labs' compute expenditure over the next two years. Starting in 2028, the deployment of self-built infrastructure by AI labs will gradually shift this dynamic. Barclays holds that when benchmarking AI labs, investors should look beyond top-line revenue figures and drill down into underlying variables such as product mix, the share of API and subscription revenue, and revenue recognition methodologies.
Odaily News: OpenAI co-founder Greg Brockman reposted the open letter, with over 100 institutions including Anthropic, AWS, Google, Microsoft, OpenAI, and Oracle jointly calling for strengthened global cyber defense to address the rapidly evolving threat of AI-powered cyberattacks.The open letter states that AI-driven cyberattacks are expected to become more prevalent and sophisticated in the coming months, posing higher risks to critical systems such as hospitals, water treatment facilities, and internet infrastructure. All parties should seize the current window of opportunity in which AI can equally enhance defensive capabilities, accelerate the closure of long-standing security vulnerabilities, and prioritize AI tools with cybersecurity capabilities for critical infrastructure defense teams.The open letter also calls on companies, cybersecurity firms, governments, and frontier AI companies to jointly invest tools, funding, and technical support to strengthen threat intelligence sharing, vulnerability remediation, and continuous security monitoring.
According to Chaoxiang Research, a Morgan Stanley report dated August 25 noted that the internet sector declined by an average of 2% last week, with Meta down approximately 7%, Amazon down roughly 2%, and Google virtually flat. Current forward P/E ratios for 2026 stand at 19x for Amazon, 17x for Google, and 17x for Meta, reflecting discounts of 36%, a premium of 36%, and a discount of 24% relative to historical averages, respectively. On an EV/EBITDA basis, Amazon at 11.2x represents a 12% discount to its two-year average, Google at 15.1x implies an 8% premium, and Meta at 8.7x reflects a 30% discount. Morgan Stanley maintains an "Attractive" rating on the internet sector, highlighting that AI capabilities are emerging as a core variable driving valuation divergence. The sector's overall forward EV/EBITDA is 9% below the five-year average, while EV/Sales is 16% above it, underscoring a divergence between revenue and profit multiples. After reclassifying stock-based compensation as a cash expense, the adjusted EV/EBITDA for digital media rises by approximately 36% on average, e-commerce by 30%, and travel and the sharing economy by 44%. Upcoming catalysts include the launch of Google's Gemini 4, stabilization of Amazon Web Services (AWS) growth, progress in Meta's AI ad monetization, and shifts in the interest rate environment.
According to TechCrunch, Amazon and NVIDIA have announced an expansion of their partnership, with AWS planning to deploy an additional 2 million NVIDIA GPUs between 2027 and 2028. The deployment will cover Blackwell Ultra, Rubin, and Rubin Ultra to meet the rapidly growing compute demand from startups, AI labs, enterprises, and government clients. This comes just five months after Amazon previously committed to deploying over 1 million NVIDIA GPUs. Neither party disclosed the transaction amount, but based on per-unit GPU pricing, the order is estimated to be worth tens of billions of U.S. dollars. Beyond chip procurement, AWS will also integrate NVIDIA’s networking, CPUs, open-source models, and physical AI and robotics stacks, including Omniverse and Isaac.
NVIDIA CFO announced that the company will deploy an additional 2 million GPUs within AWS's global infrastructure between 2027 and 2028 to strengthen both parties' computing power collaboration.
Odaily News: Brian Armstrong posted on X platform stating that another feature for AI agents has been launched. The first use case is Travala, where users can use common large language models to book over 2 million properties on Base using USDC. He expressed excitement to see more enterprises building on this foundation with AWS AgentCore and opening their services to the newest and fastest-growing consumer segment.
Odaily News: Citrini analyst jukan posted on X platform, stating that GF Securities Overseas Electronics Newsletter reiterated its Buy rating on Intel with a $136 price target, and believes this stock offering sends a positive signal. The report expects Intel's foundry business to break even in Q4 2027, with margin leverage effects fully reflected by 2028. Yield rates and external customer expansion—especially progress with Apple—are advancing steadily, and the EMIB customer base is also expanding, with customers including Google and AWS. Intel has secured support from substrate supplier Unimicron and will produce silicon capacitors internally. The report raised Intel's expected EPS for 2026 and 2027 by 3% and 1% respectively, maintaining the $136 price target after accounting for dilution effects. Intel expanded its stock offering from the initially planned $15 billion to $20 billion, with institutional demand reportedly exceeding $100 billion. The offering price was $95, and the overallotment option has been fully exercised. CEO Lip-Bu Tan and his family subscribed approximately $12 million at the offering price, which the report believes reflects management's confidence in the company and may support capital expenditure for fiscal year 2027. The report states that Intel reiterated its foundry business will achieve breakeven by the end of 2027; if delayed to 2028, the primary reason would be increased demand for additional investment. The report maintains its previous assessment, expecting 18A yield of approximately 80% in Q2 2026, with CWF already entering the capacity ramp-up phase. External customer collaboration continues to advance, with Apple's 14A high-volume manufacturing being particularly notable. The report raised Intel's back-end business revenue expectations for fiscal years 2027 and 2028 to $1.1 billion and $7 billion respectively, citing improved visibility into AWS Trainium3 adopting EMIB-T in 2027, as well as Google's Humufish and Triggerfish entering production expansion phases from H2 2027 to 2028. The report also expects that AWS and Microsoft ASIC products may adopt EMIB in 2028.
According to Fortune magazine, Amazon released its Q2 2026 earnings report, showing strong overall performance, with after-hours stock price rising over 9%. In terms of core data, AWS cloud business revenue reached $42.2 billion, a year-over-year increase of 37%, marking the fastest growth in nearly 18 quarters; operating income was $16.6 billion, up 64% year-over-year, with profit margin rising to 39.4%; AWS customer contract backlog reached $496 billion. The company's overall net sales increased by 20% year-over-year to $200.6 billion, with operating income of $27.5 billion. Regarding outlook, CEO Andy Jassy stated that Amazon's 2026 capital expenditure expectation was raised from $200 billion to $220 billion, mainly driven by rising memory costs, but even so, capacity is expected to remain insufficient to meet all demand in the next two years. Jassy also pointed out that 85% of global IT spending is still on-premises, the cloud migration wave is far from over, and AWS is winning the "largest share" of enterprise migration plans.
Chinese AI startup Moonshot AI will release the model weights of its high-performance model, Kimi K3. Developers can download the model, modify it for various purposes, and run it in their own data centers or cloud environments.Kimi K3 boasts 2.8 trillion parameters and a 1 million token context window, enabling it to process large-scale documents and codebases in a single pass. Moonshot AI plans to later publish a technical report detailing the model's architecture, training methodology, and performance evaluation results.Following the release of Kimi K3, Moonshot AI's daily revenue is reported to have increased by at least 6 times. The company is reportedly advancing a new round of fundraising at a $50 billion valuation and is considering a Hong Kong listing as early as this year.According to Bloomberg Intelligence, following the release of Kimi K3 and Z.AI's GLM-5.2, the share of Chinese open-weight models in overall token usage has risen to 68%. Services like AWS Bedrock, Microsoft Azure Foundry, and Google Vertex AI currently do not offer Chinese open-weight models such as Kimi K3 and GLM-5.2.
Cos, founder of SlowMist, shared a tweet on X platform regarding potential poisoning attack risks in Claude Code and published a detailed analysis of poisoning attacks targeting Grok Build CLI and Claude Code CLI. The analysis pointed out that the security mechanisms of Grok Build CLI are not unified, with different code paths having different trust assumptions, and the gaps between them serve as channels for attackers.Attackers may exploit malicious project configuration files to execute arbitrary commands without the user's knowledge, thereby stealing API keys, cloud credentials, or gaining control over local devices. Researchers constructed a test environment and found that on Mac systems, if Claude Code is compromised, executing a specific test command could trigger the launch of a local calculator, demonstrating a potential command execution risk. If the attack succeeds, attackers could further steal API keys from AI services such as Claude and OpenAI, causing account cost losses; obtain credentials for cloud services like AWS, Alibaba Cloud, and Tencent Cloud to access servers and data; tamper with code repositories to implant backdoors; and leverage local devices as a springboard to attack internal enterprise networks. It is reported that the relevant vulnerability has existed for one year.
Odaily Odaily A new study by the Cambridge Centre for Alternative Finance reveals that approximately 31% of Ethereum node activity is located in the United States, with another 39% distributed across EU countries excluding the UK, indicating that the geographic distribution of Ethereum nodes remains relatively concentrated in Western nations.Lead researcher Alexander Neumuller stated that while node distribution is not currently concentrated in any single country, it is heavily reliant on a few major cloud service providers, including Hetzner, Amazon AWS, and OVH. Notably, the Ethereum network does not require half of its validators to fail for problems to arise. If more than one-third of validators go offline simultaneously, the network may be unable to finalize block checkpoints (finalization). Neumuller pointed out that nodes and validators do not have a one-to-one correspondence; a single node may run multiple validators. Therefore, it is currently impossible to precisely assess the actual impact on the validator network from the failure of a specific node or service provider.Furthermore, the study reassessed the energy consumption of Ethereum following The Merge. Data shows that Ethereum's current annual energy consumption is approximately 7.9 GWh, equivalent to a continuous power draw of about 1 MW. This represents only about 0.02% of pre-merge levels, a reduction of approximately 99.98%. Currently, over 56% of the energy used by the Ethereum network comes from sustainable sources, exceeding the global average.The study also noted that client software diversity is another potential risk. If a dominant client software has a vulnerability, it could affect a large number of network participants. The report was published by the Cambridge Centre for Alternative Finance and supported by the Ethereum Foundation. (The)
: The inaugural Web4.0 and Agent Innovation Summit, along with the launch ceremony of OPC Hub, was held at Hong Kong Cyberport. Cobo was officially announced as a partner of OPC Hub, listed alongside Alibaba Cloud, AWS, NVIDIA, Dell, Lenovo, and BytePlus. Cobo will provide blockchain infrastructure support for entrepreneurs within the ecosystem. Its product capabilities cover Agentic Wallet and stablecoin payments (Cobo Payments), enabling developers to build agent-based applications with on-chain payment capabilities.
According to The Block, the Cambridge Centre for Alternative Finance (CCAF)'s newly released report "Ethereum After the Merge" shows that approximately 31% of Ethereum node activity is located in the United States, with 39% distributed across the European Union (excluding the UK), presenting an overall Western-centric centralization pattern. The report points out potential centralization risks in the Ethereum network—nodes are highly concentrated among three major hosting providers: Hetzner, AWS, and OVH. Once more than one-third of validators go offline simultaneously, network checkpoints will stop finalization (Finalization). Additionally, the report recalculated Ethereum's energy consumption; annual electricity consumption after the Merge is approximately 7.9 GWh, a decrease of approximately 99.98% compared to before the Merge, with sustainable energy accounting for over 56%, and the cost required to offset its annual carbon emissions is only about $33,500 to $73,800.
According to TechFlow Research, Morgan Stanley released a research report on July 1 stating that, regarding Bloomberg's report on Meta planning a cloud computing business, it judges that Meta is more likely to choose the lighter path of renting out idle computing power rather than building a full cloud service benchmarking AWS. The report calculates that renting out 250 MW of computing power at $40/watt could be accretive to 2028 earnings per share by approximately 8%, and when the scale reaches 1000 MW, the accretion could reach 33%, but this earnings accretion is viewed as a transitional buffer, not the core logic supporting the rating. Morgan Stanley also mentioned that Meta's self-held computing power will expand to 1.9 GW and 3.4 GW in 2026 and 2027 respectively, providing room for the rental calculations. Morgan Stanley maintains its Overweight rating on Meta with a target price of $775, representing approximately 37.6% upside compared to the closing price of $563.29, while setting the 2027 capital expenditure expectation at $175 billion; if the cloud computing business scales up, there is a possibility of an upward revision in capital expenditure.
BNB Chain has announced the official mainnet launch of its AI Agent development platform, BNB Agent Studio.Developers can now use a single prompt in AI coding tools like Cursor and Claude Code to complete Agent wallet creation, on-chain identity registration (ERC-8004), and deployment, without needing to separately set up wallets, identities, payments, custody, or LLM integration.Once deployed, Agents can use the x402 protocol to automatically deduct fees from users' pre-funded wallets to cover LLM usage, and they can be discovered and invoked by other Agents via the ERC-8183 task interface. The entire process runs on the AWS Bedrock AgentCore.The platform is also launching a limited-time free trial, where users can experience the full deployment process on the BSC testnet using their GitHub account.
: Amazon Web Services (AWS) has announced the creation of a new "Forward-Deployed Engineers" (FDE) organization, allocating approximately $1 billion in resources to accelerate the deployment and implementation of AI agents for enterprise customers.It is reported that the new team will be embedded directly within client companies to assist in building and deploying customized AI systems, while fostering "autonomous capability enhancement" among enterprises during the process. This involves not only delivering solutions but also helping clients acquire the skills to continuously develop AI applications.The FDE model was initially popularized by Palantir Technologies, with its core concept being engineers working on-site within client environments to collaboratively develop and deploy solutions, thereby rapidly responding to actual business needs and improving implementation efficiency. This model has been widely adopted in the wave of AI deployment in recent years. Currently, OpenAI and Anthropic have also launched similar FDE collaboration mechanisms, partnering with private equity firms to establish related programs valued at approximately $4 billion and $1.5 billion respectively, aimed at driving the expansion of enterprise-level AI applications.In the announcement, AWS AI leader Francesca Vasquez stated that the organization will provide clients with agentic systems that can run directly within their AWS environment. It will also export long-term reusable engineering methods, workflows, and AI practices. The $1 billion investment primarily comes from internal resource allocation rather than being an independent investment project. Industry analysts believe that as enterprises accelerate their AI transformation, the FDE model is becoming a crucial strategy for AI infrastructure vendors to compete for enterprise customers. (TechCrunch)
Grayscale Research Head Zach Pandl stated that perpetual contracts, as a core product of the crypto market, have long been limited to crypto assets such as BTC and ETH. However, Hyperliquid is changing this landscape through its HIP-3 upgrade. HIP-3 allows for the permissionless deployment of perpetual contract markets on the Hyperliquid infrastructure, and a S&P 500 perpetual contract product has already been launched on Hyperliquid.Data shows that the HIP-3 market reached a peak open interest of approximately $3.2 billion in June 2026, with a cumulative trading volume of about $200 billion. These markets are not directly operated by Hyperliquid but adopt a "permissionless infrastructure" model: any qualified developer can create derivatives trading markets on its underlying network. This makes Hyperliquid more akin to an open financial infrastructure similar to AWS, with the HYPE token capturing the overall transaction value flow.