News Heat Trend
Project Overview
Ultra is a next-generation PC game distribution platform powered by blockchain, designed to give developers and players the power to embrace the new era of gaming. Ultra describes itself as the first entertainment platform that offers all key gaming industry services and numerous centralized and decentralized apps, all accessible through a single login.Ultra's vision is to provide a complete blockchain solution, including a quick, fee-free layer 1 blockchain that is certified carbon neutral, an app ecosystem, a groundbreaking NFT standard, and an NFT marketplace. Moreover, Ultra Games, Ultra's flagship "killer app", is specifically built as a self-service publishing platform to challenge Valve's Steam, the giant PC game distribution platform.
Analyst: Kimi K3 Ultra-large Model to Be Open-Sourced Soon, AI Infrastructure Platforms May Be the Biggest Beneficiaries
analyst KawzInvests stated that Moonshot AI's upcoming Kimi K3 could become a significant event in the open-source AI space, and the infrastructure demand behind it may drive growth for AI cloud service platforms. Kimi K3 has approximately 2.8 trillion parameters, making it one of the ultra-large-scale open-source models. According to Moonshot's official evaluation, the model's performance is only slightly behind frontier models like Claude Fable 5 and GPT 5.6 Sol, and the full model weights are expected to be released on July 27.KawzInvests pointed out that a model of this scale cannot run on an ordinary laptop or even a single server; users need a computing cluster composed of a large number of GPUs to complete model loading and inference. When top-tier open-source models are made available for free, the biggest beneficiaries might not be ordinary users, but rather platforms that offer model hosting and inference services. For example, $DOCN (DigitalOcean) already supports serverless inference services for models like Kimi K2.6. Developers do not need to deploy hardware; they can call the model via API and pay per Token. Additionally, the platform hosts over 70 models and covers GPU leasing, model fine-tuning, and AI Agent development tools.As more large-scale open-source models are released, developers' demand for low-barrier AI infrastructure will continue to increase. Model hosting, inference services, and GPU cloud platforms may become key beneficiaries in the open-source AI wave.
AI Industry Express: NVIDIA NVL576 Optical Engine Configuration May Increase by 78%, Potentially Benefiting Lumentum, Coherent, etc.
According to Citrini analyst Jukan, FundaAI's latest report reveals that NVIDIA's NVL576 passive co-packaged optics technology nearly doubles the density of optical engines and optical components. The passive co-packaged optics technology increases the configuration volume of 3.2T optical engines per GPU from approximately 2.25 to around 4.0, an increase of 78%. It is estimated that the demand for Rubin Ultra optical engines will reach about 12 million units.This news may directly benefit direct suppliers of optical engines, including Lumentum, Coherent, POET, among others.
Analyst: Kimi K3 Ultra-large Model to Be Open-Sourced Soon, AI Infrastructure Platforms May Be the Biggest Beneficiaries
analyst KawzInvests stated that Moonshot AI's upcoming Kimi K3 could become a significant event in the open-source AI space, and the infrastructure demand behind it may drive growth for AI cloud service platforms. Kimi K3 has approximately 2.8 trillion parameters, making it one of the ultra-large-scale open-source models. According to Moonshot's official evaluation, the model's performance is only slightly behind frontier models like Claude Fable 5 and GPT 5.6 Sol, and the full model weights are expected to be released on July 27.KawzInvests pointed out that a model of this scale cannot run on an ordinary laptop or even a single server; users need a computing cluster composed of a large number of GPUs to complete model loading and inference. When top-tier open-source models are made available for free, the biggest beneficiaries might not be ordinary users, but rather platforms that offer model hosting and inference services. For example, $DOCN (DigitalOcean) already supports serverless inference services for models like Kimi K2.6. Developers do not need to deploy hardware; they can call the model via API and pay per Token. Additionally, the platform hosts over 70 models and covers GPU leasing, model fine-tuning, and AI Agent development tools.As more large-scale open-source models are released, developers' demand for low-barrier AI infrastructure will continue to increase. Model hosting, inference services, and GPU cloud platforms may become key beneficiaries in the open-source AI wave.
Nomura Securities: May Package Substrate Shipments Up 36% YoY to Hit Record High, Rubin Mass Production and HBM4 Technology Roadmap Become Focus
According to TechFlow Research, on July 14, 2026, Nomura Securities released a research report citing data from Japan's Ministry of Economy, Trade and Industry. Japan's packaging substrate shipment value in May reached 27.8 billion yen, a year-on-year increase of 36%, hitting a record high; shipment area increased by 10%, and the average price rose 23% to 1.356 million yen per square meter. The report pointed out that NVIDIA Rubin packaging demand will ramp up this summer (at latest by August), but the variable worth paying more attention to is that Rubin Ultra may shift from a four-die structure to a dual-module structure, as well as the large interposer integration challenges brought by HBM4 wiring upgrades. Intel released EMIB-T technology at ECTC, expected to complete mass production readiness in 2026, capable of integrating up to 12 HBM4 units under an interposer-less solution, with power supply voltage drop improved by 68-80% compared to EMIB; TSMC maintains its lead relying on 3DFabric Alliance and advantages in power delivery and heat dissipation. Nomura Securities believes that the core battlefield of the next-generation packaging competition lies in power delivery, heat dissipation, CPO, and 3D hybrid bonding, and excess returns in the packaging substrate industry will come from the ability to bind to customer technology roadmaps.
Apple turns to NVIDIA as M2 Ultra chip fails to meet advanced AI demands
Apple is shifting its reliance to NVIDIA GPUs to support its artificial intelligence operations, as the company's self-developed M2 Ultra chip has failed to meet expectations when running advanced AI workloads.Sources indicate that Apple previously attempted to use its own servers to run Google's Gemini model to power AI capabilities for the new Siri. However, the performance of its chips fell short of requirements, ultimately forcing the company to turn to NVIDIA GPUs deployed on Google Cloud. Additionally, the development of Apple's next-generation AI server chip, "Baltra," has been delayed, while the M7 Ultra chip, which is expected to rival the performance of NVIDIA's Blackwell architecture, is unlikely to be launched before 2029 at the earliest. (Coin Bureau)
Quickswap adopts Orbs as default perpetual contract infrastructure with 81.8% governance vote
Quickswap has adopted Orbs Perpetual Hub Ultra 2.0 following a community governance vote, making it the default infrastructure for perpetual contract trading across all supported chains, with 81.8% of QUICK token holders supporting the migration. This vote replaces Quickswap's previous Falkor deployment on Polygon PoS, which was supported by Orderly, and expands its Orbs perpetual contract platform launched on Base at the end of 2025. Orbs disclosed that its decentralized Layer 3 network currently has over 1.12 billion ORBS staked.
SemiAnalysis: NVIDIA Kyber NVL144 Delayed Over 12 Months, Launch Postponed to 2028
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.
Intel hikes prices on some PC and server CPUs, Xeon 6980P suggested price rises to $13,955
Odaily Planet Daily reported that Intel adjusted the prices of its PC and server processors on the 3rd. The prices of the Core Ultra 200S Plus and Xeon 6 processors, launched at the end of March, have been increased, while the prices of the Core Ultra 200S processors, released in 2024, have been slightly reduced. The suggested price of the Core Ultra 7 270K Plus has been raised by $50 from its launch price of $299 to $349, and the price of the Core Ultra 5 250K Plus has been increased by $30 to $229. Prices for the server-grade Xeon 6 processors have been raised by up to several thousand dollars, with the suggested price of the Xeon 6980P rising to $13,955. Intel stated that this price adjustment reflects the current market environment, rising supply chain costs, and strong demand for specific product lines. (ZDNet Korea Semiconductor).
Related news
Analyst: Kimi K3 Ultra-large Model to Be Open-Sourced Soon, AI Infrastructure Platforms May Be the Biggest Beneficiaries
analyst KawzInvests stated that Moonshot AI's upcoming Kimi K3 could become a significant event in the open-source AI space, and the infrastructure demand behind it may drive growth for AI cloud service platforms. Kimi K3 has approximately 2.8 trillion parameters, making it one of the ultra-large-scale open-source models. According to Moonshot's official evaluation, the model's performance is only slightly behind frontier models like Claude Fable 5 and GPT 5.6 Sol, and the full model weights are expected to be released on July 27.KawzInvests pointed out that a model of this scale cannot run on an ordinary laptop or even a single server; users need a computing cluster composed of a large number of GPUs to complete model loading and inference. When top-tier open-source models are made available for free, the biggest beneficiaries might not be ordinary users, but rather platforms that offer model hosting and inference services. For example, $DOCN (DigitalOcean) already supports serverless inference services for models like Kimi K2.6. Developers do not need to deploy hardware; they can call the model via API and pay per Token. Additionally, the platform hosts over 70 models and covers GPU leasing, model fine-tuning, and AI Agent development tools.As more large-scale open-source models are released, developers' demand for low-barrier AI infrastructure will continue to increase. Model hosting, inference services, and GPU cloud platforms may become key beneficiaries in the open-source AI wave.
Nomura Securities: May Package Substrate Shipments Up 36% YoY to Hit Record High, Rubin Mass Production and HBM4 Technology Roadmap Become Focus
According to TechFlow Research, on July 14, 2026, Nomura Securities released a research report citing data from Japan's Ministry of Economy, Trade and Industry. Japan's packaging substrate shipment value in May reached 27.8 billion yen, a year-on-year increase of 36%, hitting a record high; shipment area increased by 10%, and the average price rose 23% to 1.356 million yen per square meter. The report pointed out that NVIDIA Rubin packaging demand will ramp up this summer (at latest by August), but the variable worth paying more attention to is that Rubin Ultra may shift from a four-die structure to a dual-module structure, as well as the large interposer integration challenges brought by HBM4 wiring upgrades. Intel released EMIB-T technology at ECTC, expected to complete mass production readiness in 2026, capable of integrating up to 12 HBM4 units under an interposer-less solution, with power supply voltage drop improved by 68-80% compared to EMIB; TSMC maintains its lead relying on 3DFabric Alliance and advantages in power delivery and heat dissipation. Nomura Securities believes that the core battlefield of the next-generation packaging competition lies in power delivery, heat dissipation, CPO, and 3D hybrid bonding, and excess returns in the packaging substrate industry will come from the ability to bind to customer technology roadmaps.
Apple turns to NVIDIA as M2 Ultra chip fails to meet advanced AI demands
Apple is shifting its reliance to NVIDIA GPUs to support its artificial intelligence operations, as the company's self-developed M2 Ultra chip has failed to meet expectations when running advanced AI workloads.Sources indicate that Apple previously attempted to use its own servers to run Google's Gemini model to power AI capabilities for the new Siri. However, the performance of its chips fell short of requirements, ultimately forcing the company to turn to NVIDIA GPUs deployed on Google Cloud. Additionally, the development of Apple's next-generation AI server chip, "Baltra," has been delayed, while the M7 Ultra chip, which is expected to rival the performance of NVIDIA's Blackwell architecture, is unlikely to be launched before 2029 at the earliest. (Coin Bureau)
Quickswap adopts Orbs as default perpetual contract infrastructure with 81.8% governance vote
Quickswap has adopted Orbs Perpetual Hub Ultra 2.0 following a community governance vote, making it the default infrastructure for perpetual contract trading across all supported chains, with 81.8% of QUICK token holders supporting the migration. This vote replaces Quickswap's previous Falkor deployment on Polygon PoS, which was supported by Orderly, and expands its Orbs perpetual contract platform launched on Base at the end of 2025. Orbs disclosed that its decentralized Layer 3 network currently has over 1.12 billion ORBS staked.
OpenAI: GPT-5.6 Sol Ultra Solves a 50-Year-Old Graph Theory Conjecture in 1 Hour
OpenAI researcher Ethan Knight stated on social platform X on the 12th that GPT-5.6 Sol Ultra has generated a proof paper and prompt regarding the unsolved problem in graph theory, the "Cycle Double Cover Conjecture." Under the configuration of up to 64 parallel sub-agents, without using internet search, the model completed the result generation in about 1 hour.The paper publicly released by OpenAI is 3 pages long. It first reduces the problem to cubic graphs, then uses the 8-flow theorem and GF(3) labels, and constructs a structure where each edge is contained in exactly two cycles using linear algebraic methods. This result has not yet been published in an academic journal, undergone peer review, nor passed formal verification systems such as Lean or Coq. Thomas Bloom, a mathematician at the University of Manchester, UK, evaluated the proof as "excellent," but noted its lack of sufficient references in the bibliography. (Source: ZDNet Korea Semiconductor).
TeraWulf Secures Ultra-Long-Term Deal with Anthropic: 20-Year Lease Could Generate $19 Billion in Revenue
: Anthropic has announced that it has signed a 20-year data center lease with TeraWulf to build AI computing infrastructure in Kentucky, USA. The data center is located about an hour's drive southwest of Louisville, with a planned capacity of approximately 400 megawatts. The first phase of power supply is expected to commence in the second half of 2027. According to the agreement, the initial lease is projected to generate approximately $19 billion in revenue for TeraWulf.TeraWulf, a Bitcoin mining company, has been steadily transitioning towards AI data center infrastructure in recent years. Its stock price has risen by over 80% year-to-date, reflecting the market's sustained pricing of demand for AI computing infrastructure. (CNBC)