Morgan Stanley launches its $1.5 trillion U.S. innovative infrastructure initiative, focusing on cutting-edge technologies including AI.

Strategy CEO Phong Le stated that the company currently holds $4.75 billion in cash reserves, sufficient to cover roughly 2.7 years of dividend payments. He added that while the firm had previously expected investors to highly value Bitcoin’s liquidity and long-term growth potential, after rolling out preferred stock products, it found institutional and short-term capital investors still prioritize cash liquidity. Strategy is transitioning from solely buying and holding Bitcoin to a broader digital credit business. The company has launched preferred stock products including STRC, providing an option for investors seeking Bitcoin-related returns while wanting to mitigate volatility. Le noted that though he would personally prefer holding Bitcoin, the successful operation of these preferred stock products will ultimately benefit MSTR and the company’s Bitcoin strategy. Le also said Strategy aims to become “the JPMorgan of digital finance,” envisioning that other firms could eventually develop new investment tools based on Strategy’s financial products, decentralized finance (DeFi) could further expand into various risk and return profiles, and channel more capital into Bitcoin. Currently, Strategy holds around 840,000 BTC, equivalent to roughly 4% of Bitcoin’s total 21 million token supply. Le remarked: “We have now become a bellwether and also the central bank of Bitcoin.” Furthermore, the company’s traditional software business remains on an upward trajectory, with software revenue rising 7% year-over-year and cloud subscription revenue growing 54%.

According to monitoring by Onchain Lens, a crypto whale has transferred 500 BTC (valued at roughly $32 million) to FalconX, and 274 BTC (worth about $17.5 million) to over-the-counter (OTC) trading platform Cumberland. Another 500 BTC (approximately $32 million) was moved to a new address, with indications it may next be transferred to Galaxy Digital. The total value of the BTC involved stands at around $81.5 million.

According to Bloomberg, sources familiar with the matter revealed that AI startup Anthropic has entered into a $9.1 billion AI data center computing power agreement with Bitcoin mining firm Riot Platforms to meet the growing computing needs of its Claude users. Riot disclosed Monday that it will supply 191MW of computing power at its Rockdale, Texas campus to an AI developer, with a 20-year contract term. The sources confirmed the client is Anthropic. Riot stated the contract will run through June 2048, generating an estimated $9.1 billion in revenue, with options to extend the term by five years twice. If fully exercised, the total contract value could reach up to $16.1 billion. Following the news, Riot’s shares rose 25% in after-hours trading to $24.40, per market data from BIT (bit.com). Anthropic has recently signed a series of large-scale computing power agreements. The firm previously inked a $10 billion deal with infrastructure startup Volta Infra Holdings, and in May agreed to purchase nearly $45 billion in computing power from Elon Musk’s xAI to ease pressure from surging client demand. Riot, which was previously focused on Bitcoin mining, is shifting to AI and cloud computing infrastructure like many other crypto mining firms. It also announced another computing power construction deal with AMD earlier, and its data center business contributed to revenue growth in the second quarter.

CITIC Construction Investment’s research report points out that Unitree’s IPO pricing exceeded expectations, which is expected to drive a valuation reshaping of core robot body manufacturers. Unitree’s IPO is priced at RMB 150.80 per share, corresponding to an issued market value of approximately RMB 61 billion, surpassing prior expectations and poised to reshape valuations for this segment. Domestic industry chain players are actively advancing multi-dimensional capability building in areas such as AI "brain", motion control "cerebellum", and robot "body", while exploring applications across industrial, commercial, and other scenarios. Their shipment volumes continue to expand, and as robots improve their generalization capabilities, their application scenarios are expected to further broaden. Physical AI is the next wave of artificial intelligence, with robots serving as one of AI’s most optimal physical carriers, reflecting clear industry development trends. Upcoming catalysts for the sector include the release and mass production progress of Optimus V3, new product launches by domestic robotics firms, IPO advances by robot companies, and application deployments. It is recommended to focus on high-quality segments within the space.

BlackRock’s Head of Digital Assets, Robert Mitchnick, said Bitcoin market sentiment has seen a “clear yet subtle” shift over the past month or so. Earlier this year, Bitcoin gradually decoupled from U.S. equities; prior to that, BTC had underperformed amid a rally in AI stocks, a decoupling that was once unfavorable for Bitcoin. Mitchnick noted that when AI stocks saw a sharp pullback in July, Bitcoin outperformed U.S. equities significantly. He views this decoupling as healthy, as many investors see Bitcoin as a diversification tool in portfolios and a potential hedge against tail risks facing other assets. He added that Bitcoin ETF investors as a whole remain dominated by fundamental-driven, long-term capital. In terms of fund flows, U.S. spot Bitcoin ETFs recorded net inflows for five consecutive trading days last week, totaling around $853.5 million, marking their best weekly performance since mid-April. Of this total, BlackRock’s IBIT saw net inflows of $693.7 million, accounting for over 80% of all spot Bitcoin ETF net inflows; Fidelity’s related ETFs recorded net inflows of $116.4 million, making up around 13% of the total. Mitchnick said Bitcoin has historically been highly volatile, having gone through five major boom-and-bust cycles so far, with prices at the end of each cycle significantly higher than the previous one, though the process is marked by extreme volatility.

According to Bloomberg, Lambda, an AI cloud computing provider backed by NVIDIA, is seeking a $917 million leveraged loan to finance a NVIDIA-related chip procurement deal. The funds will go toward purchasing and installing GPUs and other infrastructure. Sources familiar with the matter noted that before the formal marketing phase, the loan has already attracted nearly $2 billion in indications of interest. Lambda is part of Neocloud, which offers AI chip and infrastructure leasing services. This financing follows the trend of AI infrastructure firms securing debt market funding for GPU purchases. Bloomberg data shows that global debt raised for AI expansion has approached $600 billion since last year. CoreWeave previously also secured financing via the institutional leveraged loan market, backed by customer contracts and chip-related assets. The Lambda loan is priced at a spread of up to 3.75 percentage points over the benchmark rate, with an issue price of roughly 99% of face value. The loan has a 4.4-year term—significantly shorter than the typical ~7-year term for institutional loans—and features a fully amortizing structure, allowing the debt to be repaid gradually over about 4 years to reduce lenders’ refinancing risk. A prepayment penalty applies if Lambda redeems the loan early. Morgan Stanley is leading the transaction, with the loan issued by Lambda Compute II LLC and Lambda Cloud Canada Inc. Lambda currently operates primarily in North America and uses NVIDIA chips exclusively; besides relying on third-party facilities, the company is also beginning to build its own data centers and has held preliminary discussions with banks about a potential IPO.