Latest news with #GraphicsProcessingUnit
Yahoo
25-07-2025
- Business
- Yahoo
NVIDIA's Q2 2026 Earnings: What to Expect
Valued at a market cap of $4.2 trillion, NVIDIA Corporation (NVDA) is a leading computing infrastructure company headquartered in Santa Clara, California. Initially known for inventing the Graphics Processing Unit (GPU), NVIDIA has evolved into a leader in accelerated computing, driving innovation across various industries, including gaming, data centers, AI, autonomous vehicles, robotics, and healthcare. It is expected to announce its fiscal Q2 earnings for 2026 after the market closes on Wednesday, Aug. 27. Ahead of this event, analysts expect this semiconductor giant to report a profit of $0.94 per share, up 44.6% from $0.65 per share in the year-ago quarter. The company has topped Wall Street's earnings estimates in three of the last four quarters, while missing on another occasion. In Q1, NVDA's EPS of $0.81 fell short of the forecasted figure by 4.7%. More News from Barchart UnitedHealth Stock Spirals Lower Again. Don't Buy the Dip. This Self-Driving Car Stock Is Surging on a Major Nvidia Boost Auto Revenue Keeps Plunging at Tesla. Should You Buy the TSLA Stock Dip or Run Far Away? Markets move fast. Keep up by reading our FREE midday Barchart Brief newsletter for exclusive charts, analysis, and headlines. For fiscal 2026, analysts expect NVDA to report a profit of $4.02 per share, up 37.2% from $2.93 per share in fiscal 2025. Furthermore, its EPS is expected to grow 32.8% year-over-year to $5.34 in fiscal 2027. NVIDIA has rallied 52.1% over the past 52 weeks, considerably outperforming both the S&P 500 Index's ($SPX) 17.3% return and the Technology Select Sector SPDR Fund's (XLK) 21.1% uptick over the same time frame. On May 28, NVIDIA released mixed Q1 results, and its shares surged 3.3% in the following trading session. The company's revenue improved by an impressive 69.2% year-over-year to $44.1 billion, surpassing consensus estimates by 2.7%. However, while its adjusted EPS of $0.81 advanced 32.8% from the same period last year, it missed the analyst expectations by 4.7%. On Apr. 9, NVIDIA was informed by the U.S. government that a license is required for the export of its H20 products to the Chinese market. As a result of these new requirements, NVIDIA incurred a $4.5 billion charge in Q1, which impacted its profitability, leading to the bottom-line miss. Wall Street analysts are highly optimistic about NVDA's stock, with an overall "Strong Buy" rating. Among 45 analysts covering the stock, 39 recommend "Strong Buy," two indicate "Moderate Buy," three advise 'Hold,' and one suggests a "Strong Sell' rating. The mean price target for NVDA is $182.60, implying a 5.1% premium from the current levels. On the date of publication, Neharika Jain did not have (either directly or indirectly) positions in any of the securities mentioned in this article. All information and data in this article is solely for informational purposes. This article was originally published on Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data


Business Recorder
26-06-2025
- Business
- Business Recorder
Pakistan's first AI data centre launched in Karachi
Data Vault Pakistan has officially launched the country's first AI-focused data centre in Karachi, an initiative set to transform how the nation builds, deploys, and scales artificial intelligence. The launch comes at a time when Pakistan is trying to accelerate its national artificial intelligence (AI) strategy. The next-generation facility is designed to fuel the ambitions of Pakistan's AI ecosystem, offering infrastructure for startups, researchers, enterprises, and government agencies, according to a press release from Data Vault. From high-performance Graphics Processing Unit (GPU) computing to secure cloud environments, Data Vault says it is making it possible for AI innovation to happen at home on Pakistani soil, 'powered by clean energy, and protected by sovereign infrastructure'. 'We didn't just build a data centre—we built a launchpad for Pakistan's AI future. Our goal is to give local innovators the tools and confidence they need to create world-class solutions, without depending on foreign platforms. This is about self-reliance, empowerment, and unleashing Pakistan's true digital potential,' Mehwish Salman Ali, founder and CEO of Data Vault Pakistan, said. With Data Vault, AI startups in Pakistan can now access enterprise-grade GPU-as-a-service to train and deploy models 'faster and more affordably'. This means faster go-to-market times, fewer limitations on compute power, and greater independence for local founders. Universities and research institutes can also benefit from partnerships with the facility, tapping into scalable infrastructure that supports deep learning, academic research, and applied AI, according to the press release. At the enterprise and government level, organisations can host sensitive workloads in a fully compliant, high-security environment ensuring data stays within the country and meets global standards for privacy and protection, it added. Data Vault claims to run its system entirely on solar power, making it Pakistan's first green data centre purpose-built for AI. According to the press release, it integrates advanced encryption and zero-trust architecture to protect sensitive workloads and critical infrastructure, and all data is stored locally. 'By providing affordable, secure, and scalable access to computers, Data Vault is helping unlock opportunities for every entrepreneur, engineer, student, and policymaker committed to building Pakistan's AI future,' the company said.


CNBC
20-05-2025
- Business
- CNBC
Kneron's smart robotics, powered by Chat-GPT
Kneron CEO, Albert Liu, unveils a Neutral Processing Unit (NPU) chip designed to outperform traditional Graphics Processing Unit (GPU) chip in AI applications. This enables smart edge devices creating intelligent robots like "Haley".


India Gazette
08-05-2025
- Business
- India Gazette
AI surge requires 50 million Sq Ft additional real estate, 45 TWH incremental power for data centres by 2030: Deloitte
New Delhi [India], May 8 (ANI): India could require an additional 45-50 million square feet of real estate space and 40-45 Terawatt Hours (TWH) of incremental power by 2030 to meet the growing demand for AI, according to a report by multinational professional services firm Deloitte. The report further suggests that targeted policy interventions will be critical in positioning India as a global AI hub. The report identifies six key pillars for India to reach its full potential in building a world-leading, AI-ready ecosystem. These pillars are real estate, power and utilities infrastructure, connectivity and network infrastructure, compute infrastructure, talent and policy framework. 'For India to accelerate its AI capabilities and realise its potential, it is necessary to introduce enabling policies to support the sector. India must develop its AI-ready infrastructure to meet this rising demand for data analytics and processing,' said S Anjani Kumar, Partner, Deloitte India. While India offers an advantage in data centre real estate with lower land and labour costs, significant new constructions are needed to meet the additional AI data centre capacity, highlights the report. From a policy perspective, the report suggests that introducing a separate category for data centres in the National Building Code and recognising them under the Essential Services Maintenance Act could further incentivise specialised infrastructure development. Additionally, fast-tracking and simplifying the approvals by setting up data centre facilitation units can help boost the deployment of data centres, the report adds. The rapid expansion of data centres in India intensifies pressure on the country's power infrastructure, creating an urgent need for enhanced transmission capabilities and grid stability. The report further add that substantial investments are needed to expand generation capacity, upgrade infrastructure and better integrate renewable energy sources, which are supported by advancements in energy-efficient cooling and offshore and onshore power sourcing. Compute infrastructure is a critical component of AI data centre operations. 'A greater focus on building high-performance compute infrastructure, scalable power and cooling systems and efficient networking infrastructure, with a reconsideration of policy framework, can make India a hotspot for AI-powered data centre development in the coming years,' said Neha Aggarwal, Partner, Deloitte India. To stay competitive, the report suggests that India must urgently boost Graphics Processing Unit (GPU) supply, promote GPU-as-a-Service (GPUaaS) and attract foreign investment to build Exaflop-scale capacity with high-performance GPUs. The report emphasises that India's workforce must upskill to fit into AI and data centre ecosystem roles, accelerating sector growth. (ANI)


Zawya
27-03-2025
- Business
- Zawya
Furukawa Electric Launched an epoch compact multi-fiber optical connector suitable for CPO
TOKYO, JAPAN - Media OutReach Newswire - 27 March 2025 - Furukawa Electric Co., Ltd. newly developed the compact 12-fiber optical connector suitable for CPO (Co-Packaged Optics). The rapid spread of generative AI has accelerated the development of CPO for datacenter applications. In CPO, an optical transceiver is placed close to a large-scale LSI (Large Scale Integration) such as a switch ASIC (Application Specific Integrated Circuit) or a GPU (Graphics Processing Unit), and the length of the electrical transmission line can be shortened. Therefore, high-speed communication and low power consumption are expected by CPO. For the deployment of CPO, it is necessary to realize a compact, easy-to-use, detachable optical fiber connector at the edge of the optical transceiver device (Fig. 1). The newly developed 12-fiber optical connector is significantly smaller than conventional multi-fiber optical connectors, with a connection area that is 1/6 MPO connectors (Fig. 2). The semi-automatic connection method using magnets provides excellent connection workability. The connector can withstand 260°C reflow (the process of soldering components to the substrate) that is necessary to be assembled to a CPO substrate. It also employs optical coupling with lenses, there is very little degradation in characteristics when dust adheres to the optical end face, it is also suitable for dusty environments such as data centers. The connection loss is less than 0.5 dB with good repeatability within ± 0.05 dB. This developed connector can also be used as a compact multi-fiber optical connector for connecting optical fibers. In addition to high-density attaching to a glass waveguide substrate that is expected to be applied to the next-generation photonics-electronics convergence form (Fig. 3), it is also applicable to more-fiber structures such as 16-fiber. We are currently developing 40-fiber optical connector. At OFC 2025, which will be held in San Francisco, USA, from March 30 to April 3 this year, we will give an oral presentation (Lecture No. M4J. 2) and exhibit samples (in the booth of Lightera booth No. 2843). This development is partly based on results obtained from the project, "Research and Development Project of the Enhanced Infrastructures for Post-5G Information and Communication Systems" (JPNP20017), commissioned by the New Energy and Industrial Technology Development Organization (NEDO).Hashtag: #FurukawaElectric The issuer is solely responsible for the content of this announcement. Furukawa Electric Co., Ltd.