# Nvidia Related Articles

HTX News Center provides the latest articles and in-depth analysis on "Nvidia", covering market trends, project updates, tech developments, and regulatory policies in the crypto industry.

GPU Rental Prices Drop 30% in Three Weeks: AI Value Chain Migrating from Nvidia to Memory Chips

GPU rental prices for Nvidia's flagship B200 chip have fallen by approximately 30% over three weeks, dropping from a high of $6.11/hour to $4.22/hour. This decline signals a potential easing of the "compute scarcity" narrative that has long supported AI hardware valuations. Concurrently, the semiconductor market is witnessing a significant divergence: while the VanEck Semiconductor ETF (SMH) has risen 15% in the past month, with memory giants Micron and SanDisk each surging nearly 60%, Nvidia's stock has declined about 3% over the same period. Analysts suggest this shift indicates that the AI value chain's bottleneck and profits are migrating from compute (GPUs) to memory. Demand for high-bandwidth memory (HBM) remains intensely strong, with contract prices soaring over 100% in H1 2026, granting memory manufacturers significant pricing power. In contrast, increased B200 supply from improved manufacturing yields and competitive pressure from new cloud providers are softening GPU rental rates. While long-term contracts, like SpaceX's $30 billion deal with Google, show sustained large-scale demand for Nvidia hardware, the softening spot prices pressure the margins of cloud providers and could eventually impact Nvidia's order flow if chip prices don't adjust. The key takeaway for investors is not a weakening AI thesis, but a recalibration within the sector: pricing power appears to be strengthening for memory chipmakers while showing signs of strain for leading GPU suppliers.

marsbit3h ago

GPU Rental Prices Drop 30% in Three Weeks: AI Value Chain Migrating from Nvidia to Memory Chips

marsbit3h ago

A Chip Company Releases AIDC Energy Storage Certification Standards. Why NVIDIA? Computing Power Reshapes Power Supply Logic. Who's in the Lead and Who's Left Out?

NVIDIA has released a "Battery Energy Storage System Self-Certification Guide," setting strict technical standards for energy storage systems specifically for AI data centers (AIDC). The guide focuses solely on certifying the Power Conversion System (PCS), not the batteries, with 10 mandatory performance metrics and 12 validation tests requiring real-world and simulation comparisons. Key requirements include rapid dynamic response to AI workloads, high-frequency system telemetry, and detailed electromagnetic transient models. The move is driven by the extreme and fluctuating power demands of next-generation AI hardware. Modern AIDCs require energy storage systems to act as intelligent, controllable grid assets, not just passive backup, to manage instantaneous, massive power load shifts that traditional UPS systems cannot handle. This redefines the competitive landscape for energy storage providers, shifting focus from capacity and cost to advanced control capabilities and system integration. While the market potential is significant—with forecasts of hundreds of GWh in new demand by 2030—the certification creates a high barrier to entry. It requires proven PCS delivery volumes and credible plans for rapid capacity scaling, favoring established, well-resourced players. Early movers like Fluence (partnering with Siemens) and several Chinese companies have secured projects ahead of the standard, but new entrants must now navigate this rigorous, costly, and time-intensive certification process to compete in the AIDC energy storage market.

marsbit5h ago

A Chip Company Releases AIDC Energy Storage Certification Standards. Why NVIDIA? Computing Power Reshapes Power Supply Logic. Who's in the Lead and Who's Left Out?

marsbit5h ago

After Missing the 20x, I've Found a 'Dumb' Method for AI Investing

**Missing the 20x Opportunity: A Simple 'Dumb' Approach to AI Investing** The AI boom, driving NVIDIA's revenue from $60B to $216B in two years, creates immense investment pressure. However, like the internet bubble of 2000, the largest AI opportunities likely lie ahead, perhaps after a correction. Instead of rushing in now or waiting paralyzed for a crash, the author proposes a third way: building a "knowledge warehouse" by systematically mapping the AI industry to be ready when opportunities arise. The core of the strategy is understanding AI's four-layer value chain: 1. **Compute Infrastructure (The "Engine"):** This foundational layer, where all money eventually flows, includes: a) **Chip Design:** NVIDIA's dominance via its CUDA ecosystem, b) **Chip Manufacturing/Packaging/Memory:** TSMC's near-monopoly in advanced manufacturing and SK Hynix's lead in High Bandwidth Memory (HBM), c) **Optical Interconnects:** Essential for large-scale AI clusters (e.g., Lumentum, Coherent), d) **Cooling & Power:** Critical for high-density AI data centers (e.g., Vertiv), e) **Servers/Data Centers & Cloud Platforms:** The physical and virtual wholesale providers. 2. **Models & Tools (The "OS"):** The competitive layer of foundation models (OpenAI, Anthropic, Google, Meta, xAI), now generating real revenue. A key shift is the center of gravity moving from **Training** models to **Inference** (running models), which demands different chip characteristics and could challenge NVIDIA's monopoly. 3. **Middleware & Platform ("The Glue"):** Connects models and applications (e.g., Scale AI, Hugging Face). This layer could explode if applications take off. 4. **Vertical Applications ("The Cash Register"):** Where AI meets end-users (e.g., enterprise AI, coding tools, medical AI, robotics). A critical cross-cutting constraint is **Energy**, as AI's massive power consumption drives investment in nuclear and other energy infrastructure. The author identifies four key questions for further research: 1) How will the shift from Training to Inference reshape the competitive landscape? 2) With tech giants spending over $600B on capex, where is the ROI from AI applications? 3) What are the under-the-radar opportunities in the "second" and "third" circles of the value chain (e.g., cooling, specialty foundries)? 4) How will geopolitics (e.g., U.S.-China chip restrictions) bifurcate the supply chain? The conclusion is that missed opportunities stem from insufficient research, not slow timing. By methodically studying each layer—its business models, competition, and valuations—investors can build the "killer intuition" needed to act decisively when the market presents its chance.

marsbit5h ago

After Missing the 20x, I've Found a 'Dumb' Method for AI Investing

marsbit5h ago

Bernstein Research Report Analysis: HBM Prices Must Double Next Year, Memory Becomes a Burden for AI

Bernstein's report indicates that High-Bandwidth Memory (HBM) prices must double (increase 2-2.5x) in 2027. This is because, while standard DRAM prices have surged approximately 4.5x, HBM prices remain locked in annual contracts, making standard DRAM significantly more profitable per wafer for memory makers (2x the revenue, nearly 3x the gross profit). To maintain capacity allocation for HBM, its price must rise sharply. A critical consequence is that HBM is sold packaged within Nvidia GPUs. To protect its 75% gross margin, Nvidia is likely to pass on HBM cost increases by raising GPU/system prices by about 4x the HBM cost hike. Combined with standard DRAM and NAND price increases, this could raise total AI capital expenditure for cloud providers by around 30%. The report maintains Outperform ratings on Samsung, SK Hynix, and Micron, significantly raising price targets due to expected upward earnings revisions. KIOXIA is the main loser, having no HBM exposure. MediaTek could benefit if cloud providers seek more ASIC-based alternatives to avoid Nvidia's markup. Valuations have shifted to a P/E basis, reflecting record-high profitability and cash generation expected in this cycle.

marsbit19h ago

Bernstein Research Report Analysis: HBM Prices Must Double Next Year, Memory Becomes a Burden for AI

marsbit19h ago

A Company Once on the Verge of Collapse Now Has a Market Cap Surpassing Bitcoin

"Once nearly bankrupt, SK Hynix now surpasses Bitcoin in market cap." On June 22, SK Hynix's market value reached $1.35 trillion, exceeding Bitcoin's $1.29 trillion, driven by soaring demand for its High-Bandwidth Memory (HBM), a critical component for AI training. As NVIDIA's primary HBM supplier with over 60% market share, the company reported a 72% operating profit margin in Q1. Its remarkable turnaround stems from a 13-year bet on HBM technology, initiated in 2009 and rescued by SK Group's acquisition in 2012. The company plans a Nasdaq listing, potentially attracting more institutional capital. This milestone highlights a market preference for tangible AI infrastructure assets—like HBM with its physical scarcity and verifiable revenue—over narrative-driven crypto-AI projects, which remain in early developmental stages and lack comparable proof of scalable demand or economic models. The capital shift underscores the premium placed on real-world supply constraints and profitability in the current AI boom.

marsbit21h ago

A Company Once on the Verge of Collapse Now Has a Market Cap Surpassing Bitcoin

marsbit21h ago

A Company Once on the Brink of Bankruptcy Just Surpassed Bitcoin in Market Cap

On June 22nd, driven by rising stock prices, SK Hynix’s market capitalization reached $1.35 trillion, surpassing Bitcoin's total market cap of approximately $1.29 trillion. This temporarily made it South Korea's highest-valued company. The core driver of this surge is HBM (High Bandwidth Memory), for which SK Hynix is the primary supplier to NVIDIA, holding over 60% market share. AI's demand for high memory bandwidth has translated into immense profitability, with SK Hynix reporting a 72% operating profit margin in Q1. The company's success follows a 13-year bet on HBM technology, beginning in 2009. It nearly failed after the 2001 dot-com bubble, was acquired by SK Group in 2012, and was subsequently recapitalized to continue its long-term HBM development. The article contrasts this with the Crypto AI narrative. Capital currently favors AI infrastructure players like SK Hynix due to "real orders, physical barriers, and quantifiable profit margins." In comparison, Crypto AI projects, promising decentralized compute and data markets, remain largely conceptual with limited tangible progress. Examples include Bittensor, whose core mechanisms are still under development, and Bitcoin miners transitioning to AI, who face significant funding gaps and execution challenges. The piece cites analysis suggesting the AI sector has absorbed nearly all new market liquidity since 2022, leaving little for crypto. It concludes that the current AI infrastructure红利 is captured by entities with proven technical barriers and supply capabilities, while crypto networks still need to define their concrete role in the value chain.

链捕手21h ago

A Company Once on the Brink of Bankruptcy Just Surpassed Bitcoin in Market Cap

链捕手21h ago

Musk Targets AI Infrastructure Again: Tesla Set to Sell 'Compute Building Blocks'

Tesla has filed a trademark for "Megapod," a modular AI data center hardware system, signaling a potential move into AI infrastructure. The system, as described in the filing, would integrate computer servers, AI processing hardware, networking equipment, power distribution, and cooling into a single, plug-and-play unit. This follows Tesla's existing "mega" product line like the Megapack energy storage system. The move comes less than a year after Tesla disbanded its in-house Dojo supercomputer team, shifting away from custom AI chip development. Analysts suggest Megapod is likely not a direct challenge to NVIDIA's dominant GPU-based systems, but rather targets the supporting infrastructure layer—power, cooling, and rapid deployment—where Tesla's energy business has expertise. The company's Megapack batteries are already being purchased by AI firms like xAI to manage the intense power demands of data centers. While still an early-stage trademark with no product details, Megapod hints at Tesla's broader strategy to commercialize modular AI infrastructure. It aligns with trends seen in Elon Musk's other ventures, such as SpaceX's lucrative contracts to lease GPU compute capacity. The potential market addresses critical bottlenecks in AI expansion: electricity, thermal management, and deployment speed.

marsbitYesterday 07:35

Musk Targets AI Infrastructure Again: Tesla Set to Sell 'Compute Building Blocks'

marsbitYesterday 07:35

Analysis of the Latest Portfolio Adjustment by the "Top Player" in the U.S. Stock Market: $9 Billion Short on NVIDIA, Shifting Focus to Power and Memory Sectors

AI investor Leopold Aschenbrenner has made a significant portfolio shift, taking a $9 billion nominal short position against top AI infrastructure stocks like NVIDIA, ASML, and Oracle. Simultaneously, he is redirecting capital towards what he sees as the next critical bottlenecks in the AI boom: power, memory, and data center networking, alongside private investments in AI model companies like Anthropic. This move is interpreted not as a call that the AI bubble has burst, but as a rotation within the infrastructure stack. The analysis highlights NVIDIA's recent $25 billion bond issuance as a potential signal, questioning why a cash-rich company would seek external debt despite high profits and increased dividends/buybacks. The core investment thesis is that the initial, crowded "picks and shovels" trade in semiconductors is maturing. The next wave of capital is expected to flow into the physical and logistical constraints of AI expansion: electricity supply, memory chip capacity, data center construction, and enabling technologies like optical networking (fiber) for high-bandwidth communication, where copper remains crucial for short distances. Aschenbrenner's substantial (approx. 20% of fund) private stake in Anthropic is noted as a key part of his strategy—investing directly in the "mine" (AI models) rather than just the "shovels." The discussion concludes that while certain segments may be overvalued, the overarching AI infrastructure demand driven by real product usage remains robust. The most promising long-term investments are seen in essential, non-sexy infrastructure—particularly energy and power companies—whose demand is viewed as a global constant irrespective of AI's cyclicality.

marsbit06/20 03:07

Analysis of the Latest Portfolio Adjustment by the "Top Player" in the U.S. Stock Market: $9 Billion Short on NVIDIA, Shifting Focus to Power and Memory Sectors

marsbit06/20 03:07

CPU Makes a Comeback to the Table, A $170 Billion "Power Seizure" Drama Begins

A new era is dawning for the server CPU (Central Processing Unit), driven by the shift from AI model training to large-scale reasoning and the rise of Agentic AI. This article explores how the CPU is reclaiming a central role in the AI data center. For years, the focus has been on the GPU (Graphics Processing Unit) for AI training. However, as AI moves to the inference and Agent phase—where tasks involve complex, multi-step reasoning, tool calls, and data management—the workload balance is flipping. Studies show CPUs now handle over 70% of the workload in Agentic AI, up from 10-30% in training. This is because Agent tasks generate massive intermediate data (KV Cache) that exceeds GPU memory, forcing it to be offloaded to the CPU's larger, more scalable memory pools. This increased importance is translating into market changes. Major players are taking note: NVIDIA launched its first standalone CPU line, Vera, based on ARM architecture and optimized for Agent performance. AMD doubled its server CPU market forecast to over $1200 billion by 2030. Analyst reports project the total server CPU market could reach $1700 billion by 2030, with AI-driven demand being a primary driver. Furthermore, the classic ratio of CPUs to GPUs in AI servers is rapidly changing, converging from 1:8 toward 1:1 for Agent deployments. This surge in demand has led to a rare industry-wide price increase of 10-15% for server CPUs from Intel and AMD, breaking a decade-long trend of "more performance for the same price." Demand is bifurcating into high-core-count CPUs for in-rack GPU support and moderate-core CPUs for standalone Agent task orchestration. In China, this global trend presents an opportunity for domestic CPU manufacturers like Hygon (海光信息) and Huawei Kunpeng, who are bolstered by both growing AI infrastructure needs and national policies promoting technological self-reliance ("xin chuang"). The maturity of their software ecosystems is also accelerating, evidenced by faster adaptation to new AI models. In conclusion, the narrative is shifting from a GPU-centric view to one where CPU-GPU synergy is critical. The CPU is no longer a peripheral component but a performance-defining bottleneck and a key growth driver in the AI hardware stack, opening a massive new market estimated in the hundreds of billions of dollars.

marsbit06/19 13:41

CPU Makes a Comeback to the Table, A $170 Billion "Power Seizure" Drama Begins

marsbit06/19 13:41

Interview with CoreWeave Co-founders: AI Demand Seems to 'Intensify' Every Day

An Interview with CoreWeave Executives: AI Demand Seems to 'Intensify' Every Day In an interview, CoreWeave executives highlight a structural shift in AI infrastructure demand. While GPU availability remains crucial, the primary bottlenecks are evolving to include powered data center shells, skilled labor (like electricians), and complex supply chain execution. They note that AI demand, particularly for agentic AI and reasoning models, continues to intensify daily, accelerating since Q1 2024. This demand is driving a need for more balanced infrastructure. CoreWeave is redesigning data centers to allocate more space for storage and CPUs alongside GPUs, with significant interest in Nvidia's upcoming Vera CPUs. The company, serving top AI labs and hyperscalers, emphasizes its client-driven model, building precisely to customer specifications. CoreWeave attributes its competitive edge to proven execution, performance, and a mature platform for AI deployment. Pricing is structured to pass component cost increases (e.g., for HBM memory) to customers, protecting margins. Looking ahead, they anticipate Vera Rubin platform deployments to begin meaningfully in late 2025, with a major ramp throughout 2027, mirroring the Blackwell (GB) series rollout pattern. The competition is shifting from merely acquiring chips to holistic engineering and delivery capability.

marsbit06/19 10:37

Interview with CoreWeave Co-founders: AI Demand Seems to 'Intensify' Every Day

marsbit06/19 10:37

活动图片

Hot Categoriesright-arrow

Hot Tagsright-arrow

Market Analysisright-arrow

Industry Newsright-arrow