GPU · 1283

Starcloud closes a new round of funding with $250 million led by Manhattan West Ventures

Comparatively, according to TechCrunch, the space computing power startup Starcloud announced the completion of a new round of financing of 250 million US dollars, led by Manhattan West Ventures, with Nvidia, Cisco, Benchmark, EQT and other institutions participating. Of these, Nvidia invested about 25 million US dollars in this round of financing. The new funding will be used to expand satellite manufacturing facilities and advance research and development of the next-generation orbital data center satellite Starcloud-3. Starcloud revealed that the company is already running the Nvidia H100 data center GPU in orbit and has completed model training based on that GPU. Currently, most space computing projects use edge computing chips, and Starcloud is collaborating with Nvidia to provide test data for future Vera Rubin Space-1 GPUs designed specifically for space environments. Starcloud CEO Philip Johnston also previously indicated plans to mine Bitcoin in space. This article is sponsored by GENG, Build Your Fortune on GENG (https://geng.one)

4h agoburnking#financing

Tencent's chip leader Gao Jianlin left his job and started a business to target the RISC-V high-performance AI CPU circuit

According to MaxForAI, according to MaxForAI, the core head of Tencent's chip research and development, has recently left Tencent and started a business. He plans to develop high-performance CPUs based on the RISC-V architecture around high-performance AI servers and agentic AI (intelligent AI). Gao Jianlin is regarded as one of the early core promoters of Tencent's self-developed chip system. According to data, he formed an FPGA hardware team within Tencent in 2013, began setting up AI chip research and development in 2018, established the Penglai Laboratory in 2020, and promoted Tencent's development of various AI chips and data center deployment. This startup focuses on CPUs rather than the currently competitive AI GPU market. According to the report, Gao Jianlin believes that with the rapid development of Agentic AI, the AI inference process will involve model call, tool execution, search, database interaction, and large-scale task scheduling, and the CPU will assume a more important scheduling and control role in the AI system. According to reports, Gao Jianlin was involved in RISC-V related research and development during his time at Tencent, and participated in various technical directions such as chip architecture, verification, and back-end. Its new company plans to build high-performance server CPUs based on the open instruction set RISC-V to enter the AI infrastructure market. Currently, the name of Gao Jianlin's new company, financing conditions, and specific product launch dates have not been disclosed. The market is concerned about whether it will become another emerging force in the field of AI chips in China targeting server CPUs and smart body infrastructure.

8h ago

Nvidia Rubin begins mass production and delivery: Microsoft's first devices have arrived

Comparing news, AI News, Microsoft CEO Satya Nadella published a data center site map. Microsoft's first mass-produced version of NVIDIA Vera Rubin has arrived. Nvidia later confirmed that Vera Rubin is entering full mass production. Vera Rubin is a next-generation rack-scale AI computing platform after Blackwell. The NVL72 incorporates 72 Rubin GPUs and 36 Vera CPUs. Nvidia says that compared to the GB200 NVL72, it can reduce the inference cost per million tokens to about one-tenth, and the number of GPUs required to train the MoE model to one-quarter.

12h ago
If it's just tokenized assets and doesn't connect to DeFi, what's left of RWA?

If it's just tokenized assets and doesn't connect to DeFi, what's left of RWA?

Author: Jesus Rodriguez, co-founder of Sentora Compiled by: Luffy, Foresight News Original title: Does RWA still make sense without DeFi? Discussions in the RWA industry often begin with a simple vision: take a treasury bill, fund share, stock, invoice, megawatt hour, or GPU for one hour, then mint a token representing it. Is it useful? It's really useful. But can it be called transformative? It's far from there. This is like putting a bar code on a container and claiming that a global trade problem has been solved. Barcodes make containers recognizable and machine-readable, but they don't create ports, cranes, customs, insurance, financing, shipping routes out of thin air, or bring in buyers from afar. A token is simply an addressable token of interest, and DeFi is a marketplace operating system. The question really worth discussing is not how many types of assets can go on the chain, but how many assets can complete valuation, financing, hedging, transaction monetization, and loss disposal in a stressful environment, and there is no need for offline meetings and coordination every time a transaction occurs. Tokenization completes the representation of equity; what DeFi brings is actual utility. Tokenization is just a bar code, and a similar scene has happened in the history of the supply chain finance market. The reason why mortgages can be scaled up is not as simple as turning a paper document into an electronic record. To actually achieve large-scale expansion, a complete set of operating mechanisms was created around this type of asset: credit review, post-loan services, securitization, credit rating, warehousing and financing, repurchases, hedging, clearing and settlement, and loss allocation rules. RWA also needed to go through the exact same evolutionary process. An asset that can be adapted to DeFi requires six levels: legally enforceable rights, reliable data sources, clear transfer and redemption rules, enforceable secondary market liquidity, collateral parameters that match actual behavior, and a credible settlement and loss disposal path. Most tokenization projects, on the other hand, tend to stop at the top five levels. There is a simple test that can be used to test the maturity of an asset. It only requires answering three questions: How much is this asset currently worth? Can the agreement complete withdrawal and monetization at this point? If the first two judgments are all wrong, who bears the loss? When smart contracts can definitively answer the above three questions, RWA can truly become a basic component of finance. Before that, it was mostly just a digital packaging shell. The deepest technical contradiction of RWA's quadruple time clock is that RWA runs under multiple sets of different time clocks at the same time. The blockchain can complete settlement in seconds and operate uninterrupted for 7 x 24 hours; oracles may update prices every hour or every day; underlying traditional exchanges are closed at night and on weekends; custodians follow bank working days; and the asset redemption process may take 1 day, 5 days, or even 30 days. If you use such a slow-paced RWA asset to support fast-maturing DeFi liabilities, such as stablecoin loans. This is the term shift, and it is also the core model that banks have relied on for hundreds of years: using short-term debt to fund long-term slow assets. This model has practical value, but the risk must be reasonably priced. Imagine a scenario: At 2 a.m. on Sunday, assets hit the liquidation threshold. Smart contracts can seize tokens immediately, but the underlying real-world market won't open until Monday, and the issuer's redemption business will not be processed until Tuesday. On-chain liquidation has been completed, and real-world asset disposal has only just begun. This creates a clearing gap. DeFi requires immediate withdrawal for monetization, but the real world does not allow it. The time difference between the two. This gap has counterintuitive consequences. Even treasury bonds with very low volatility are riskier than native crypto assets that are more volatile when used as collateral. The price of ETH fluctuates drastically, but it can be traded around the clock; the price of RWA assets appears to be stable, and it may only be up to a dozen hours without a new price tag. A flat price sometimes represents safety, and sometimes it's just a disguise of stale data. Liquidity is an exit channel, not TVL. The digital public also has common misunderstandings about liquidity. Liquidity is not equal to TVL, does not equal the existence of a trading pair, nor does it mean that the issuer promises to eventually redeem it according to net worth. Liquidity refers to the ability to convert a position into the settlement asset you need at an acceptable discount within the time window allowed by your debt. Take a crowded theater for example: the size of the hall cannot determine whether it is safe in the event of a fire; what really matters is the width of the exit channel. One copy of RWA to...

1d agoForesight News#DeFi #RWA

Goldman Sachs Raises CoreWeave Price Target to $139, Maintains Neutral Rating

Comparative news, according to Goldman Sachs's August 20 research report, CoreWeave's second-quarter revenue was in line with expectations. The EBIT profit margin was 200 basis points higher than the market consensus, and the 2026 revenue guidance exceeded market expectations by 1%. The revenue backlog increased 5% month-on-month to US$104 billion, adding more than US$25 billion in committed orders since the third quarter. Active electricity installed capacity increased from 1 GW in the first quarter to more than 1.5 GW, and the contracted electricity installed capacity reached 4.2 GW. Goldman Sachs raised its 12-month price target from $121 to $139, which has 53% upside from the current share price and maintains a neutral rating. Goldman Sachs believes that CoreWeave's short-term certainty is clear: demand continues to lead supply, intergenerational pricing for old and new GPUs remains high, and production capacity is expanding as scheduled. Next-generation chips (Blackwell, Vera Rubin) continue to hit new highs, and recent A100 contract deliveries have been extended to 2029. The share of enterprise customers has increased (Caterpillar, IBM, Nissan, ZF), and demand for AI computing power is spreading from tech giants to the real economy. Goldman Sachs expects EBITDA to increase from $3.1 billion in 2025 to $31.3 billion in 2028. A neutral rating reflects waiting for software and platform services to become a more definite contributor to profit margins before making more positive judgments.

1d ago

Bitcoin mining companies' AI capital expenditure in the first half of the year exceeded revenue by 15 times

Comparative news, according to Cointelegraph, the latest report by BlocksBridge Consulting shows that nine listed Bitcoin mining companies had AI and high performance computing (HPC) business capital expenditure of 5.11 billion US dollars in the first half of 2026, with related revenue of only US$340 million during the same period, with an input-output ratio of about 15:1. This highlights the large upfront investment required for transformation. In a sample of 15 mining companies and AI data center companies, the total capital expenditure for the 2026 reporting period reached US$30.7 billion, an increase of 42.6% over the full year of 2025. Despite investment far exceeding revenue, AI and HPC revenue continued to grow at an accelerated pace, with nine mining companies rising 52% month-on-month to $206 million in the second quarter, led by Core Scientific, TeraWolf, and Bitdeer. BlocksBridge points out that although mining companies have power and land advantages, conversion to AI-ready production capacity still requires significant investment in substations, cooling systems, network equipment, and GPUs.

1d ago

CoreWeave signs multi-billion dollar AI cloud computing deal with quantification company Hudson River

Comparatively, quantitative trading company Hudson River Trading (HRT) has reached a multi-year cooperation agreement with CoreWeave, an artificial intelligence cloud computing company, and the two sides will use CoreWeave's AI computing services to develop new transaction research tools and models. CoreWeave Chief Revenue Officer Jon Jones said the agreement was a “major expansion” of the existing partnership between the two parties, but did not disclose specific financial terms. According to people familiar with the matter, the scale of the deal reached the level of several billion dollars. The partnership highlights the rapidly growing demand for high-performance AI infrastructure from financial institutions. As the application of artificial intelligence models in quantitative trading, risk analysis, and market forecasting continues to expand, trading companies are increasing their investment in GPU computing power and dedicated cloud computing resources.

2d ago

Global equity and debt double kill: 30-year US bond yields soared to 5.33%, and the AI industry chain suffered a severe setback

Comparing news, the global financial market has experienced sharp fluctuations. The European, American, Japanese, and South Korean markets have experienced a double slump in stocks and bonds, and the sharp rise in long-term US bond yields has become the focus of market attention. In the Asian market, the sharp decline in the Korean stock market triggered a trading mechanism. The KOSPI index closed down 5.8%, SK Hynix fell nearly 10%, Samsung Electronics fell more than 8%; and the Nikkei 225 index closed down 3.16%. More than 5,000 A-shares fell, leading the decline in AI industry chain sectors such as semiconductors, computing power hardware, PCBs, memory, and CPO. In terms of US stocks, the three major indices fell for the third consecutive trading day, and the Philadelphia Semiconductor Index fell nearly 5% in a single day. Shares of AI-related companies such as Micron Technology, Western Digital, SanDisk, Marvell, AMD, Intel, Coherent, and Credo experienced a sharp correction in stock prices. According to market analysts, the core trigger for this round of sell-off comes from the rapid rise in global long-term bond yields. The 30-year US Treasury yield rose to 5.33% intraday, a record high since 2007. Meanwhile, French, German, British, and Japanese long-term treasury yields have all risen to multi-year highs, and global long-term capital costs are being repriced. According to data from the US Treasury Department, the amount of US debt held by overseas investors fell to 9.299 trillion US dollars in June, a decrease of about 72 billion US dollars compared to May. Among them, Japan's holdings fell to 1.116 trillion US dollars, reducing their holdings by 26.4 billion US dollars in a single month; UK holdings fell to 939.9 billion US dollars. As risk-free interest rates rise, the market is beginning to re-evaluate the AI industry's high capital investment model. Investors are concerned that continued expansion of data centers, GPU procurement, and infrastructure construction will require significant financing, and that higher capital costs may reduce future cash flow estimates for technology companies. Currently, the market is concerned about three major variables: 30-year US Treasury yield trends, the Federal Reserve's judgment on long-term interest rate paths, and whether subsequent earnings reports from tech giants such as Nvidia and Broadcom can verify AI return expectations. The US Treasury Department will issue 16 billion US dollars of 20-year US bonds at 1 a.m. Beijing time on Thursday. The market will pay close attention to the results of this issuance.

3d ago

OpenAI and Anthropic's revenue fell short of market expectations, and bulls were congested and bears increased volatility

Comparing news, US stocks plummeted last night, and AI trading became the center of market sell-off. The NASDAQ fell 1.3%, the S&P 500 fell 0.7%, the NASDAQ 100 fell 1.7%, and the Philadelphia Semiconductor Index fell 5.6%. Storage stocks bore the brunt, with SanDisk falling about 9%, Micron falling about 7%, and Western Digital about 7%; AI chip leader Nvidia fell about 2.3%, Broadcom fell about 3.2%, semiconductor chains such as Marvell, Intel, and ultra-microcomputers also generally fell, and the decline was even deeper in the direction of optical communication and AI network equipment. For this round of the AI bull market, there is currently pressure on revenue expectations and position structures at the same time. The first pressure comes from commercialized data from AI labs. OpenAI disclosed to investors that revenue for the second quarter increased from $5.7 billion to $6.7 billion in the first quarter, up 18% month-on-month, but losses widened further and operating profit margins continued to decline. This growth rate fell short of expectations of some shareholders, and also caused the market to re-evaluate OpenAI's profit path before the IPO. Anthropic has also sparked controversy. As of the end of July, its annualized revenue operating rate is said to have reached 65 billion US dollars. It is still a very high growth rate, but it is lower than the previous optimistic expectations of 70 billion to 80 billion US dollars from some third parties and the market. This is hitting where the AI market is most sensitive. Over the past year, US stock bulls have been willing to renew payments for GPUs, data centers, storage, power, and cloud capital, provided that leading application companies such as OpenAI and Anthropic can continue to prove that demand for terminals is strong enough. However, as revenue growth falls short of optimistic expectations and losses continue to expand, investors will begin to re-examine the return cycle of the AI capital expenditure chain. The position factor amplified this round of decline. Goldman Sachs Prime Brokerage data has shown that the share of bears in typical S&P 500 stocks has risen to a high level since 2011; FactSet and Robinhood statistics also show that most short positions in major US stocks have risen in the past three months. At the same time, the AI infrastructure chain is still gathering large amounts of capital, and some overvalued AI companies have also become overcrowded and shorted targets for hedge funds. Bullish crowding and increasing bears exist at the same time, making the market more sensitive to any negative news. Under this structure, callbacks are amplified. The AI chain increased significantly in the early stages, and the bulls are already making considerable profits; the bears are waiting for cracks in revenue, profit margins, or capital expenditure logic. Once the growth story of OpenAI and Anthropic falls short of the market's highest expectations, capital will first cash out profits, and bears will also follow the trend to lower valuations, eventually forming a market where technology stocks collectively weakened last night.

3d ago