Crypto Morning Brief: Meta's Self-Developed AI Chip "Iris" Scheduled to Start Mass Production in September, CXMT Opens IPO Subscription on July 16
TechFlow Selected TechFlow Selected
Crypto Morning Brief: Meta's Self-Developed AI Chip "Iris" Scheduled to Start Mass Production in September, CXMT Opens IPO Subscription on July 16
Micron Technology plans to increase its total investment in the United States to over $250 billion by 2035.
Navigating Web3 tides with focused insightsAuthor: TechFlow
Yesterday's Market Dynamics
Meta's Self-Developed AI Chip "Iris" Scheduled for Mass Production in September, 2027 Computing Power Target Reaches 14 GW
According to Reuters, Meta plans to mass-produce its self-developed data center AI chip "Iris" starting in September as part of its fourth-generation Meta Training and Inference Accelerators project, to enhance AI capabilities on platforms like Facebook and Instagram and reduce reliance on external GPUs from Nvidia, AMD, etc. Internal memos show that Iris completed testing in just 6 weeks with no major defects; Meta plans to deploy 7 GW of computing power this year and increase it to 14 GW by 2027, with 2024 AI infrastructure spending potentially reaching up to $145 billion. To secure expansion, the company has signed long-term supply agreements with Samsung Electronics, SanDisk, and Sumitomo Electric to cope with "price hikes" and shortages of memory and AI chips.
US Military Completes New Round of Strikes on Iran, Hitting Approximately 170 Military Targets Over Two Consecutive Days
The US Central Command stated that the US military completed a new round of strikes on Iran on July 8 local time, to further weaken Iran's ability to attack merchant ships and innocent civilian sailors in the Strait of Hormuz. The US military struck approximately 90 Iranian military targets, including air defense systems, coastal reconnaissance assets, missile and drone storage facilities, naval capabilities, and military logistics infrastructure along the Iranian coast.
The day before, the US military successfully carried out offensive strikes on Iran. On July 7, US Central Command forces struck approximately 80 Iranian military targets, including more than 60 Islamic Revolutionary Guard Corps small speedboats, aiming to impose a heavy cost on Iran for violating the ceasefire by attacking three merchant ships passing through the Strait of Hormuz. The US Central Command stated that the US military remains vigilant and powerful, and is ready to execute actions directed by the Commander-in-Chief at any time.
Zhipu: Plans to Raise Over 31.4 Billion HKD via Share Placement at Approximately 13% Discount
Zhipu (02513.HK) announced on the HKEX that it plans to place 19.78 million shares at 1588 HKD/share (approximately 13% discount) to raise about 31.41 billion HKD.
CXMT Technology Starts New Share Subscription on July 16
July 09, Micron Technology (MU.O): Plans to increase total investment in the US to over $250 billion by 2035.
Micron Technology: Plans to Increase Total Investment in the US to Over $250 Billion by 2035
Apple (AAPL.O): The company newly signed a multi-year agreement with Broadcom (AVGO.O), which is expected to exceed $30 billion, to produce over 15 billion US-made chips.
Decentralized AI Protocol Prime Intellect Completes $130 Million Series A Funding, Led by Radical Ventures
According to official news, Prime Intellect announced the completion of a $130 million Series A funding round, led by Radical Ventures, with participation from Nvidia, Intel Capital, Dell Capital, and existing investors. The company stated it will use the funds to continue building its "Open Superintelligence Technology Stack" to support users in training, deploying, and continuously optimizing their own models.
JPMorgan: Biggest Risk for Bitcoin is Not Strategy Selling Pressure, But Blockchain Adoption Bypassing Public Chains
According to The Block, JPMorgan analysts pointed out in the latest report that although Strategy's Bitcoin selling plan has attracted market attention, it is not the core risk facing Bitcoin. The real structural threat lies in the fact that blockchain applications such as tokenization, payments, and settlements are increasingly occurring on permissioned chains (Permissioned Blockchain), rather than public chains like Ethereum. If this trend continues, the public chain ecosystem will face issues such as declining liquidity and weakened capital inflows, ultimately dragging down Bitcoin valuation.
Analysts also warned that banks building their own blockchain infrastructure and the popularization of tokenized deposits may weaken the status of stablecoins in institutional payments; regulated alternative solutions such as the SWIFT blockchain plan, digital euro, and digital yuan also constitute competitive pressure. However, analysts also pointed out that if hybrid public-private chain models emerge, stablecoin regulation becomes clearer, or Bitcoin continues to be held as "digital gold", the above risks may be mitigated.
Meta Launches Paid AI Developer Model for the First Time
Meta announced the launch of its first paid AI developer model Muse Spark 1.1, which is an industry-competitive agent and programming model. In multiple agent evaluations, its performance is comparable to GPT-5.5 and Opus-4.8. It is available for use via the new Meta Model API and Meta AI platform starting today. Meta CEO Zuckerberg stated that pricing is planned to be lower than competitors.
Bank of Korea Governor: Should Advance Won Stablecoin Legislation ASAP, Advocates Bank-Led Issuance
According to Edaily, Bank of Korea Governor Shin Hyun-song stated at a National Assembly meeting that relevant systems for Won stablecoins and the "Digital Asset Basic Act" legislation should be advanced as soon as possible, emphasizing that a competitive and complementary relationship will form between stablecoins, deposit tokens, and central bank digital currencies.
The Bank of Korea also suggested that Won stablecoins should prioritize issuance by consortia where banks hold more than half the shares, and establish a statutory policy coordination mechanism involving the Bank of Korea, Financial Services Commission, and finance department to reduce potential risks to monetary policy, foreign exchange policy, and financial stability. Currently, there is still divergence in South Korean politics and the market regarding whether banks should lead stablecoin issuance.
Today's Market Performance
Recommended Reading
An Investment Podcast Fed 200 Past Episodes to AI for Review: Bet on Micron for 180% Gain, But Missed Cursor's $6 Billion Acquisition
https://www.techflowpost.com/article/32444
Limitless Podcast has reached its 200th episode. The two hosts did something a bit dangerous: they fed the transcripts of all 200 episodes to Claude and ChatGPT, letting AI help them find hidden investment themes they hadn't discovered, review judgments made over the past year, and predict the next stop for AI. When they pressed the record button for the first episode 14 months ago, SpaceX was still a private company valued at $350 billion, Anthropic valued at $61 billion, OpenAI valued at $300 billion, and GPT-4o was still the flagship model. Now SpaceX's market cap exceeds $2 trillion after listing, Anthropic is close to $1 trillion, and ChatGPT users have more than doubled.
The episode unfolds along two lines. One is a review: When they first shouted "compute power is national security," SanDisk rose 35 times; when betting on memory chips, Micron rose 180%, SK Hynix became the company with the largest market cap in South Korea, and Samsung's profits exceeded Nvidia's. They even said on the program three months ago that they wanted to buy Cursor stock, but Cursor was acquired by SpaceX for $60 billion. If the companies of the founders they interviewed were made into a portfolio, the one-year return was about 4 times, with the best, Valor Atomics, at 13 times. The other line is forward-looking: The number of times big terms like superintelligence and AGI were mentioned in the program plummeted by 90%, while Anthropic's mentions quadrupled and surpassed OpenAI. The three directions they are most bullish on next are space training models, AI implementation in vertical fields, and small models on the edge.
Crypto Old Money Turns: Paradigm Raises $1.2 Billion, Half Bet on AI and Robotics
https://www.techflowpost.com/article/32437
Paradigm's new $1.2 billion fund marks crypto VC shifting from "pure crypto bets" to broader frontier tech investments like AI, robotics, and deep tech. Top capital has not abandoned crypto, but is repositioning its value—future most attractive projects may no longer be single Web3 applications, but super opportunities at the intersection of AI, blockchain, and new infrastructure.
From Frenzy to Flight in Two Months, Energy Funds Lose $3.2 Billion in a Single Week
https://www.techflowpost.com/article/32436
The energy surge brought by the Middle East conflict is rapidly receding, with global funds withdrawing from energy funds at record levels. Falling oil prices and supply recovery have caused energy giants like Exxon and Chevron to pull back significantly. The market is switching from "war premium trading" to "fundamental pricing". The future direction of the energy sector will depend on whether the ceasefire can continue, and whether the new round of power demand from AI data centers can take over the traditional oil narrative.
Gross Margin Exceeds Samsung, Not a Word on HBM: Interpretation of CXMT IPO Details
https://www.techflowpost.com/article/32431
CXMT Technology launched its STAR Market IPO, welcoming a performance explosion thanks to the DRAM cycle reversal, with net profit of 33 billion yuan in Q1 2026 and gross margin exceeding Samsung's, but the prospectus also revealed that its AI HBM layout has not yet been realized. Currently, CXMT core benefits from DDR5 demand growth and supply gaps brought by overseas manufacturers switching to HBM production, but high depreciation, cycle volatility, and capacity digestion pressure determine whether this storage feast can continue.
Viewing the Stablecoin War from Open USD: The Dilemma of Fragmentation and the Next Generation "Dollar Account"
https://www.techflowpost.com/article/32435
The emergence of Open USD marks that stablecoin competition is shifting from "who issues US dollars" to "who controls the US dollar flow network". In the past, stablecoin issuers earned seigniorage from reserve asset interest, but with the addition of distribution channels, payment giants, and financial institutions, real profits may gradually shift to the party controlling user entry points and clearing standards. The core of the future stablecoin war is not to recreate another USDT or USDC, but who can become the default entry point for the next generation of on-chain dollar accounts and financial infrastructure.
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
深潮TechFlow
