News linked to both this project and an event.
According to Reuters, Chinese AI startup DeepSeek is developing its own AI chips, three informed sources revealed. The chip is designed specifically for inference scenarios, rather than for model training. The project was launched approximately one year ago and remains in the early stages. The company has engaged with chip design, wafer foundry, and storage enterprises, and has quietly increased recruitment of chip design engineers without publicly posting job listings. If successfully developed, DeepSeek will reduce its reliance on Nvidia and Huawei Ascend chips, following the trend of global AI giants such as OpenAI and Anthropic developing their own hardware. Affected by U.S. export controls, DeepSeek previously shifted from Nvidia H800 to Huawei chips. This self-developed chip is regarded as a significant strategic transformation. Meanwhile, DeepSeek also plans to complete its first round of external financing, with a fundraising scale of approximately $7 billion, and a valuation reaching $52 billion to $59 billion.
SemiAnalysis posted on X, stating that just three months after Jensen's GTC demonstration of the Kyber NVL144, the product has encountered a major setback and is delayed by more than 12 months, with its launch postponed to 2028. NVIDIA's NVL72x2 back-to-back rack architecture has also been cancelled, thereby limiting the expansion domain of the Rubin Ultra.
Serenity has released an exclusive analysis of the AI ASIC market on the X platform, presenting the core thesis that "NVIDIA is the kingmaker of the ASIC market." It proposes a set of industry reasoning logic, arguing that NVIDIA CEO Jensen Huang is not pleased with Broadcom monopolizing the custom ASIC track. With implicit support from the NVIDIA ecosystem, companies such as Marvell, MediaTek, AlChip, and GUC are steadily capturing market share originally held by Broadcom, taking on more custom chip projects for hyperscale cloud vendors. This landscape is comparable to the rise of emerging cloud service providers last year, serving as an important means for NVIDIA to hedge against the moat created by leading cloud vendors developing their own ASICs.Serenity suggests this could represent a two-year trading opportunity but does not constitute investment advice. It also predicts that after 2030, major companies like Google will internalize a significant amount of chip design work. It added that NVIDIA has the ability to reshape the valuation of the industry chain, and there have already been market expectations that Marvell could potentially reach a trillion-dollar market cap.
according to an official announcement, Binance CrossMargin, PortfolioMargin, and PortfolioMargin Pro have added 15 new bStocks tokens as qualified collateral assets, including Circle (CRCLB), Micron (MUB), NVIDIA (NVDAB), Sandisk (SNDKB), Tesla (TSLAB), SpaceX (SPCXB), Advanced Micro Devices (AMDB), iShares MSCI South Korea ETF (EWYB), Intel (INTCB), Strategy (MSTRB), Lumentum (LITEB), Meta (METAB), Microsoft (MSFTB), Palantir (PLTRB), and Invesco QQQ Trust (QQQB). The corresponding trading pairs are now available for margin trading. This feature is only available to VIP3 and above users in eligible regions, and loans are not currently supported.
According to TechFlow Research, Nomura Securities confirms the AI cycle peak is extended to 2028. Global data center construction is just beginning to scale up, with the number of projects increasing from 240 to 280, and gigawatt-level projects increasing from 40 to 50. The deployment peak will only begin to decline after 2027, pushing back the entire cycle. Server market growth rates were significantly upwardly revised, with global figures rising from 43% to 74%/65% (2026/2027), and AI servers from 58% to 78%/76% (2026/2027), indicating capacity demand far exceeds prior expectations. However, there are real bottlenecks on the capacity side: Wafer-level Substrate (WoS) has become a constraint. TSMC's capacity target is 2,000kpcs, while Nomura estimates actual output is only 1,800kpcs. Additionally, CPU demand is systematically underestimated; AMD, Arm, and NVIDIA are all expanding the CPU market scale, which will further extend the cycle.
Odaily reports: AI chip startup Etched has completed a roughly $800 million funding round, with investors including quantitative trading giant Jane Street and a venture capital firm affiliated with Taiwan Semiconductor Manufacturing Company (TSMC). The company is currently testing its AI inference chip product and plans to begin shipping to select customers this summer. It has also signed sales contracts totaling approximately $1 billion, though specific customers were not disclosed.Founded in 2022, the company positions itself as a potential competitor to NVIDIA in the field of AI computing chips, focusing on designing customized chip architectures for large model inference scenarios. It is collaborating with TSMC to develop "low-voltage inference" technology aimed at reducing energy consumption and heat dissipation pressure.This funding round, previously reported to have a valuation of around $500 million, includes participation from Stripes, funds associated with Peter Thiel, and several quantitative firms. Jane Street is said to have invested over $100 million in total, with subsequent additional contributions. (Bloomberg)
Ondo Finance has announced the integration of its tokenized securities platform, Ondo Global Markets, with Uniswap. Over 430 tokenized stocks and ETFs on the platform are now simultaneously available on Ethereum and BNB Chain. Users can trade via the Uniswap interface or the UniswapX API.Compliant users can trade tokenized shares of companies such as SpaceX, Tesla, NVIDIA, Apple, Microsoft, and Amazon, as well as major ETF assets like SPY and QQQ, directly on-chain. Trades leverage UniswapX to benefit from optimal liquidity routing, MEV protection, and gas-free transactions. All wallets and protocols integrated with the UniswapX API can support Ondo's full suite of tokenized securities without requiring additional development. Data shows that since its launch in September 2025, Ondo Global Markets has covered over 430 tokenized securities, with its TVL surpassing $1 billion and cumulative trading volume exceeding $20 billion.
NVIDIA officially announced the launch of NVIDIA Halos for Robotics, a full-stack safety system designed for robotics and Physical AI. It aims to unify AI computing capabilities with robot safety architecture, accelerating the large-scale deployment of next-generation autonomous robots in factories, warehouses, and logistics scenarios. It is reported that humanoid robot company Agility Robotics has become one of the first adopters of NVIDIA Halos for Robotics, applying it to the safety system construction of its humanoid robot Digit.
cloud computing startup Runpod has announced the completion of a $100 million funding round and stated it has rejected multiple acquisition offers. Specific investor information has not yet been disclosed. The company focuses on providing GPU computing power rental services for developers, supporting open-source models and AI application deployment. Currently, this sector generally relies on NVIDIA GPU servers for computing resources. As AI applications expand, market demand for low-cost, highly flexible computing infrastructure continues to rise, driving up both the valuations and funding activity of cloud service providers like Runpod. (The Information)
Bitget has listed 7 stock tokens including rSCHD (US Dividend Stock ETF), rYUM (Yum! Brands), rASST (Strive), and rDHI (D.R. Horton), covering popular sectors such as finance and consumer goods.It is reported that rTokens, identified by the prefix "r" + stock ticker symbol (e.g., rNVDA for NVIDIA), are issued by Reality, a licensed RWA protocol under Bitget. By partnering with the compliant broker Alpaca, these tokens are directly connected to global liquidity pools such as Nasdaq and NYSE. Key features include: 1:1 reserve backing of underlying assets held by licensed custodians, stock dividends distributed 1:1 in token form, synchronized mapping of corporate actions (such as stock splits and reverse splits), and the ability to use the holdings as cross-collateral for unified accounts and USDT-margined perpetual contracts, allowing users to hold global stock assets while flexibly managing their funds.
“White-Haired Stock God” Serenity stated on X platform that he believes the "photonics theme + continuous-wave (CW) laser bottleneck" is one of the most important investment directions in the current AI infrastructure field.Serenity pointed out that the market seems to have overlooked Lumentum's (LITE) market cap growth from about $30 billion in 2024 to over $65 billion, with the core reason being the EML laser supply bottleneck driven by NVIDIA and changes in the optical network architecture.He believes the market is currently experiencing a similar logic, where continuous-wave lasers, 1.6T optical modules, and CPO (Co-Packaged Optics) architecture have become the focus of a new round of industrial upgrades. As NVIDIA continues to sign long-term supply agreements, AMD and other cloud service providers are competing for supply chain resources for continuous-wave lasers and optical components.Serenity cited Goldman Sachs research, stating that the total addressable market (TAM) for the optical communications market could grow to $154 billion by 2028, while the CPO market size is expected to surge from nearly zero to $91 billion within the next two and a half years.He stated that companies including Applied Optoelectronics (AAOI), SIVE, and other continuous-wave laser-related enterprises all hold strategic value, and believes the market may once again replicate the investment opportunities brought by the EML laser supply crunch. He is currently focusing on related assets in the continuous-wave laser supply chain, such as SOI, Tower Semiconductor (TSEM), and SIVE.
Bitcoin mining company HIVE Digital announced that its wholly owned subsidiary, BUZZ High Performance Computing (BUZZ HPC), has entered into a sovereign AI infrastructure partnership with Bell Canada and Cohere. The collaboration will integrate Bell AI Fabric’s data center and network infrastructure, Cohere’s large language models and enterprise-grade AI capabilities, and BUZZ HPC’s computing platform—built on NVIDIA-accelerated GPU cloud and AI factory technologies. BUZZ HPC disclosed that it has signed a three-year GPU cloud contract valued at up to $220 million and will procure GB200 NVL72 rack-scale systems comprising 2,304 NVIDIA Grace Blackwell GPUs. This AI cluster will provide compute power for Cohere’s foundational models and enterprise AI applications.
Sharon AI has announced the issuance of convertible senior notes to raise $700 million, with the proceeds intended to support its computing agreement with NVIDIA.
Bitget, in collaboration with Block Scholes, has released an RWA perpetual contract liquidity report. Data shows that Bitget’s tokenized stock and commodity markets have continued to mature since 2026. Taking NVIDIA (NVDA-USDT) as an example, as of mid-May, the order book liquidity within a ±2% spread reached approximately $4.1 million—roughly 75% of Bitget’s Bitcoin spot market depth.
according to the official announcement, SpaceX (SPCX) was listed on NASDAQ today, and Bitget has simultaneously opened its US stock spot token rSPCX trading channel. From now until 12:00 PM EST on June 13, trading rSPCX enjoys a 0 fee promotion, supports fractional share trading, with a minimum of 10 USDT required. Additionally, users can participate in leveraged trading through Bitget stock contracts.It is reported that rTokens, identified by the letter r followed by the stock symbol (e.g., rNVDA for NVIDIA), are issued by Reality, a licensed RWA protocol under Bitget. Through a partnership with the compliant broker Alpaca, they directly connect to global liquidity pools such as NASDAQ and the NYSE. Their features include: 1:1 reserve of underlying assets held by a licensed custodian, 1:1 distribution of stock dividends in token form, synchronous mapping of corporate actions like stock splits and reverse splits, and the ability for holdings to serve as joint margin for unified accounts and USDT-margined contracts, allowing users to flexibly manage funds while holding global stock assets.
: According to an official announcement, to meet users' round-the-clock trading needs, Bitget Stock Spot (rToken) has officially launched its weekend trading feature. The first batch supports 20 popular U.S. stock options, including rAAPL (Apple), rNVDA (NVIDIA), and rTSLA (Tesla).During U.S. stock market closures, the platform will generate continuous weekend trading prices based on Friday's closing prices, market maker quotes, and market expectations. This feature supports limit orders and stop-loss/take-profit settings. To prevent abnormal fluctuations, the system implements an approximate ±20% protection limit for weekend limit orders.rToken, identified by the letter "r" followed by the stock ticker (e.g., rNVDA for NVIDIA), is issued by Reality, a licensed RWA protocol under Bitget. Through a partnership with compliant broker Alpaca, it connects directly to global liquidity pools such as Nasdaq and the New York Stock Exchange. Its features include: 1:1 reserves of the underlying assets held by a licensed custodian, stock dividends distributed 1:1 in token form, corporate actions such as stock splits and reverse splits synchronously reflected, and holdings can be used as joint margin for unified accounts and U-margined contracts, allowing users to flexibly manage funds while holding global stock assets.
Sharon AI Holdings, a Nasdaq-listed company, announced a six-year strategic computing cooperation agreement with NVIDIA to jointly expand AI infrastructure capabilities in Australia.Under the agreement, the two parties will collaborate to build approximately 72MW of data center computing capacity and deploy infrastructure based on the NVIDIA DSX AI Factory architecture. The plan is to gradually scale up to 40,000 Grace Blackwell GB300 GPUs to meet the computing needs of AI startups, enterprise clients, and research institutions.The cooperation model adopts a structure combining revenue generation with credit support: Sharon AI will be responsible for selling cloud services based on NVIDIA's computing power, while NVIDIA, upon receiving hardware and basic product revenue, will also participate in a share of cloud service revenue, forming a sustainable "usage-driven revenue model."Sharon AI stated that this collaboration will significantly enhance its capital efficiency, enabling it to expand AI infrastructure capabilities without relying on traditional heavy-asset financing and accelerate the deployment of "sovereign AI computing power" in Australia. With this partnership, Sharon AI's total AI factory capacity will increase to 132MW, of which approximately 102MW is already contracted by clients. The company expects to deploy over 55,000 NVIDIA GPUs by mid-2027. (Businesswire)
sources say NVIDIA has begun pitching its first independent central processing unit (CPU) product, Vera, to Chinese clients. Designed specifically for Agentic AI systems, the chip has entered mass production, marking NVIDIA's attempt to further expand its presence in the Chinese market with a CPU offering.According to sources, some Chinese clients have already shown interest in Vera. One major Chinese cloud computing company plans to procure over 300 servers equipped with dual Vera CPUs for testing, and will decide whether to expand procurement after the tests are completed.Built on the Arm Holdings architecture, Vera is NVIDIA's first independent CPU product. NVIDIA has previously stated that Vera's performance in AI agent-related computing tasks is 1.8 times that of comparable competitor products, and expects the product to contribute approximately $20 billion in revenue by the end of this fiscal year (ending January next year).The report notes that as the AI industry's focus gradually shifts from model training to inference computing, CPUs and custom chips are gaining more attention. Vera also positions NVIDIA to directly compete with Intel and Advanced Micro Devices (AMD), which have long dominated the server CPU market.Sources indicate that due to strict U.S. export restrictions on high-end GPUs, CPUs face relatively smaller regulatory hurdles in the Chinese market compared to GPU products. Currently, some Chinese clients plan to first deploy Vera chips for testing in overseas data centers. Meanwhile, software ecosystem compatibility and existing domestic AI chip deployment frameworks may still impact the subsequent large-scale adoption of Vera. (Reuters)
According to The Information, OpenAI is negotiating a 20-year lease agreement, which would represent its largest infrastructure investment commitment to date. NVIDIA has discussed providing credit support for this project.
Odaily Odaily, according to an official announcement, Bitget has added 15 US stock tokens as margin currencies in its unified account and USDT-margined contract cross-margin mode.The newly supported assets include: core tech assets such as rAAPL, rAMZN, rMETA, rMU, rTSLA, rGOOGL, rNVDA, rINTC, rMSFT, rASML, rAVGO, rTSM, rSNDK, as well as the broad market index tokens rQQQ and rSPY.Following the update, users can hold spot positions in these related US stock tokens and use them directly as margin to participate in contract trading, enabling more flexible capital management and cross-market risk hedging. For more details, please refer to the official Bitget platform announcement.