# Сопутствующие статьи по теме GPU

Новостной центр HTX предлагает последние статьи и углубленный анализ по "GPU", охватывающие рыночные тренды, новости проектов, развитие технологий и политику регулирования в криптоиндустрии.

Jensen Huang's CES 2026 Keynote: Three Key Themes and a 'Chip Beast'

At CES 2026, NVIDIA CEO Jensen Huang unveiled the Rubin AI computing platform, a major leap in AI infrastructure designed to make AI "think longer" and operate more cost-efficiently. The Rubin architecture includes the Rubin GPU, Vera CPU, NVLink 6, and BlueField-4, working together to significantly reduce inference costs—up to 10x lower than Blackwell—and support longer context and multi-step reasoning. A key innovation is the Inference Context Memory Storage Platform, which uses BlueField-4 and Spectrum-X to manage AI context data at scale, improving token throughput and energy efficiency. NVIDIA also introduced the DGX SuperPOD, scaling the Rubin platform to 576 GPUs for large-scale AI clusters. Huang emphasized the shift toward Reasoning/Agentic AI and announced major updates to NVIDIA’s open-source ecosystem, including models like Nemotron and tools for RAG, safety, and speech. In physical AI, NVIDIA launched Cosmos for physics-aware video generation and prediction, and Alpamayo—an open-source vision-language-action model for autonomous driving, now in production with Mercedes-Benz. The event also highlighted growing adoption of NVIDIA’s robotics platform, Isaac GR00T, by companies like Boston Dynamics and LG, signaling broader commercialization of AI in real-world applications.

marsbit01/06 02:08

Jensen Huang's CES 2026 Keynote: Three Key Themes and a 'Chip Beast'

marsbit01/06 02:08

The Most Centralized Giant in the Crypto World Starts Selling the 'Decentralized AI' Dream

Tether, the highly centralized issuer of the USDT stablecoin, reported $13 billion in profit in 2024—far exceeding the combined revenues and losses of major AI firms like OpenAI and Anthropic. With only 150 employees, Tether earns primarily by investing user funds in U.S. Treasury bonds, profiting from the interest without paying users any yield. Now, Tether is aggressively investing in AI. It loaned over $600 million to Northern Data, Europe’s largest GPU cloud provider with over 10,000 Nvidia H100 GPUs. It also released QVAC Genesis, a massive open-source AI training dataset, and acquired Blackrock Neurotech, a brain-computer interface company, for $200 million. Total AI-related investments approach $1 billion, with potential additional deals in robotics sector. Despite its centralized control over USDT reserves and lack of external audits, Tether promotes a “decentralized AI” vision—advocating for local AI operation and individual data ownership. Critics find this ironic, given Tether’s opaque governance. Tether’s move into AI may stem from concerns over declining Treasury yields and a desire to position itself as a tech innovator. Unlike AI startups burning billions without clear profitability, Tether uses stablecoin profits to fund speculative AI bets—insulating itself from sector risks while gaining influence. The article suggests that in 2026, the best business model in AI might be not doing AI at all, but rather funding it with profits from a separate, lucrative venture.

比推01/05 14:50

The Most Centralized Giant in the Crypto World Starts Selling the 'Decentralized AI' Dream

比推01/05 14:50

AI Industry Welcomes a Cash-Rich Tether

Tether, the company behind the stablecoin USDT, reported a staggering $13 billion profit in 2024, significantly outperforming major AI companies like OpenAI and Anthropic, both of which incurred substantial losses. With only 150 employees, Tether achieved a per capita output 60 times greater than OpenAI’s. Its business model is simple: for every USDT issued, Tether holds one US dollar, primarily investing these reserves in U.S. Treasury bonds. It earns the interest income without paying any to USDT holders. In 2024, interest alone contributed $7 billion to its profits. Now, Tether is aggressively investing in AI. It loaned over $600 million to Northern Data, Europe’s largest GPU cloud provider, acquiring what is effectively an AI training base. It also released QVAC Genesis, a massive open-source AI training dataset. Furthermore, Tether invested $200 million in Blackrock Neurotech, a pioneering brain-computer interface company. Additional investments in robotics could bring its total AI-related spending to nearly $2 billion. Tether’s move into AI may be driven by both anxiety over declining Treasury yields and ambition to establish itself as a tech leader. Ironically, it promotes "decentralized AI" despite being a highly centralized company itself. While OpenAI and Anthropic struggle with profitability and continuous fundraising, Tether leverages its highly profitable stablecoin business to fund its AI ambitions risk-free, making a paradoxical case that the best business model in AI might be not to do AI at all.

marsbit01/05 09:12

marsbit01/05 09:12

Computing Power Subprime Crisis: The AI Infrastructure Debt Wave, Miner Leverage, and the Vanishing 'Liquidation Liquidity'

AI Infrastructure Debt Crisis: A Looming "Compute Subprime" Scenario Beneath the surface of booming AI investment and data center expansion, a severe financial mismatch is brewing. Credit investors are growing alarmed as the industry uses long-term, real-estate-like debt models to finance rapidly depreciating tech assets with an effective shelf life of just 18 months. The core issue is a fundamental asset-liability mismatch. AI compute is inherently deflationary; inference costs are falling 20-40% annually due to technological advances, eroding the future cash flows used to service debt taken out at peak 2024 prices. This risk is amplified by a shift in financing. High-risk, venture-grade tech assets are being packaged into low-risk, utility-grade project finance and asset-backed loans (ABL), transforming potential equity losses into systemic defaults. Crypto miners, often portrayed as successfully "pivoting" to AI, are particularly vulnerable. Many have not deleveraged but have instead taken on double leverage—using volatile crypto holdings as collateral to borrow more dollars to buy GPUs. This creates a dangerous correlation risk where a crypto crash and a drop in AI rental prices could occur simultaneously. The final, critical flaw is the illusion of collateral. Unlike real estate, a defaulting borrower's GPUs are nearly impossible to liquidate. They are physically dependent on specialized infrastructure, face rapid obsolescence, and lack a deep secondary market, meaning the repo market needed for a orderly清算 (liquidation) does not exist. This is not a critique of AI's potential but a warning of a profound credit mispricing, where deflationary tech assets are financed with rigid infrastructure debt, creating a hidden chain of potential defaults.

marsbit12/18 11:04

Computing Power Subprime Crisis: The AI Infrastructure Debt Wave, Miner Leverage, and the Vanishing 'Liquidation Liquidity'

marsbit12/18 11:04

Abandoning Mining for the Cloud: Bitcoin Miners Reshape Business Models in the AI Era

Bitcoin miners are facing a survival crisis due to declining Bitcoin prices and rising operational costs, with the average cost to mine one Bitcoin now exceeding its market value. In response, many are pivoting to leasing data center space to major tech companies for AI computational needs. This shift leverages their existing infrastructure—including power capacity, cooling systems, and high-performance hardware—to meet the explosive demand for AI data processing. Companies like Core Scientific have successfully transitioned, turning losses into profits by repurposing their facilities. This move not only provides stable revenue streams but also enhances the industry’s overall resilience, signaling a broader evolution from pure cryptocurrency mining toward diversified digital asset technology services.

比推12/16 13:19

Abandoning Mining for the Cloud: Bitcoin Miners Reshape Business Models in the AI Era

比推12/16 13:19

Axe Compute (NASDAQ: AGPU) Completes Corporate Restructuring (formerly POAI), Enterprise-Grade Decentralized GPU Computing Power Aethir Officially Enters Mainstream Market

Predictive Oncology has officially rebranded as Axe Compute (NASDAQ: AGPU), marking its transition into commercializing Aethir’s decentralized GPU network to provide enterprise-grade, guaranteed computational power for global AI companies. The core infrastructure is supported by the Aethir Strategic Compute Reserve (SCR), which offers predictable GPU reservations, dedicated computing clusters, and enterprise-level SLAs to address AI training, inference, and data-intensive workload demands. This move represents the first time decentralized GPU infrastructure has entered mainstream capital markets via a U.S. publicly listed company. Axe Compute will serve as the enterprise-facing entity, delivering compliant and scalable computational resources, while Aethir continues to power the underlying decentralized GPU-as-a-Service infrastructure. The structure bridges Web3 decentralized networks with Web2 enterprise needs, allowing businesses to utilize distributed GPU resources within familiar procurement and compliance frameworks. Aethir’s network currently spans 93 countries and over 200 regions, with more than 435,000 GPU containers deployed, supporting high-end hardware like NVIDIA H100, H200, B200, and B300. Axe Compute’s model aims to mitigate industry challenges such as long GPU procurement cycles, centralized cloud queuing, and pricing volatility by offering reserved GPU access, bare-metal performance, multi-region deployment, and enterprise SLAs. This listing is seen as a significant milestone in scaling decentralized AI infrastructure into enterprise markets, providing a publicly evaluable model for the commercial adoption of distributed computational resources.

marsbit12/12 13:35

Axe Compute [NASDAQ: AGPU] Completes Corporate Restructuring (formerly POAI), Enterprise-Grade Decentralized GPU Computing Power Aethir Officially Enters Mainstream Market

Predictive Oncology has officially rebranded as Axe Compute and will trade on NASDAQ under the ticker AGPU. This rebranding signifies the company's shift to operating as an enterprise-level provider, commercializing Aethir's decentralized GPU network to deliver guaranteed computational power for global AI enterprises. Axe Compute's infrastructure is supported by the Aethir Strategic Compute Reserve (SCR), which offers predictable GPU reservations, dedicated computing clusters, and enterprise-grade SLAs to address computational bottlenecks in AI training, inference, and data-intensive workloads. This move marks the first time decentralized GPU infrastructure has entered mainstream capital markets via a U.S. publicly listed company. Axe Compute will serve as the enterprise-facing delivery and contracting entity, while Aethir continues to operate as the underlying decentralized GPU-as-a-Service infrastructure. This structure bridges Web3 decentralized networks with Web2 enterprise demand, allowing businesses to use distributed GPU resources within familiar compliance and procurement frameworks. Aethir's network currently spans 93 countries, over 200 regions, and deploys more than 435,000 GPU containers, supporting high-end hardware like NVIDIA H100, H200, B200, and B300. Axe Compute's model aims to provide guaranteed GPU reservations, dedicated clusters, bare-metal performance, multi-region deployment, and enterprise SLAs—addressing common industry challenges such as long procurement cycles, centralized cloud queues, and price volatility. This represents a significant step in scaling decentralized AI infrastructure for commercial use.

深潮12/12 13:14

Others

1Web3er Travel Safety Manual: a16z Security Lead's 6-Step Checklist