AI Compute

Twitter

Facebook

Copy Link

Key Facts

Key PlayersNVIDIA, Microsoft, Google, AWS, CoreWeave

Top News

SK hynix unveils iHBM cooling tech that cuts AI memory heat by 30% for next-gen accelerators

SK hynix announced iHBM, a breakthrough thermal architecture that embeds cooling elements directly inside High-Bandwidth Memory packages to tackle AI overheating. The technology reduces thermal resistance by over 30%, enabling stable performance under extreme computational loads and paving the way for next-gen HBM5 accelerators in dense AI data centers.

Technology2 sources

SK hynix unveils iHBM cooling tech that cuts AI memory heat by 30% for next-gen accelerators

Technology2 sources

Chinese chip stocks surge as Huawei unveils plan to match leading chips with 1.4-nanometer technology

Hong Kong semi stocks rallied sharply after Huawei revealed its LogicFolding technology could enable production of 1.4-nanometer chips by 2031. SMIC and Hua Hong Semiconductor jumped over 16% as investors see the major tech leap as China's path to technological self-reliance despite U.S. export restrictions limiting access to advanced processors.

Business2 sources

Chinese chip stocks surge as Huawei unveils plan to match leading chips with 1.4-nanometer technology

Business2 sources

Intel Nova Lake targets edge AI with unusual chip design that ditches performance cores

Intel is developing a specialized Nova Lake processor for edge AI with an unconventional 8E+12Xe configuration—eight efficiency cores paired with a large integrated GPU, completely omitting performance cores. The GPU-focused design prioritizes AI inference and parallel compute over traditional desktop performance.

Technology2 sources

Intel Nova Lake targets edge AI with unusual chip design that ditches performance cores

Technology2 sources

France commits €1.5 billion to quantum computing and advanced microchips in global tech race

French President Emmanuel Macron announced €1.5 billion in funding for quantum computing and advanced microchips, declaring France has the means to win the global technology race. The investment comes one day after the US unveiled $2 billion in quantum computing stakes, intensifying competition in cutting-edge technology that promises breakthroughs in drug discovery, financial modeling, and cryptography.

Policy2 sources

France commits €1.5 billion to quantum computing and advanced microchips in global tech race

Policy2 sources

Lenovo AI revenue surges 84% as company posts record earnings and targets $100 billion sales

Lenovo delivered its strongest quarter ever with $21.6 billion in Q4 revenue, beating estimates by $2.2 billion. AI revenue surged 84% year-over-year and now represents 38% of total sales, while net profit jumped 479% to $521 million. CEO Yuanqing Yang announced plans to reach $100 billion in annual revenue within two years, driven by booming demand for AI servers and Copilot+ devices.

Business4 sources

Lenovo AI revenue surges 84% as company posts record earnings and targets $100 billion sales

Business4 sources

Modal Labs raises $355M at $4.65B valuation as AI coding and compute demand surge

Modal Labs secured $355 million in new funding at a $4.65 billion valuation, up from $1.1 billion in September. The serverless AI infrastructure startup helps companies rent GPUs for AI inference workloads and test AI-generated code, seeing revenue jump from $60 million to $300 million annually as AI coding tools drive unprecedented demand for compute capacity.

Startups2 sources

Modal Labs raises $355M at $4.65B valuation as AI coding and compute demand surge

Startups2 sources

Anthropic explores Microsoft's Maia 200 AI chip to reduce Nvidia dependence amid compute struggles

Anthropic is in talks with Microsoft to use the tech giant's Maia 200 AI chips as the AI startup faces mounting compute capacity challenges. The discussions follow Microsoft's $5 billion investment in Anthropic and represent a strategic push by Microsoft to compete with Amazon and Google in the custom AI chip market.

Technology4 sources

Anthropic explores Microsoft's Maia 200 AI chip to reduce Nvidia dependence amid compute struggles

Technology4 sources

Google commits $15 billion to build massive data center in Missouri next to Amazon's site

Google announced a $15 billion investment to construct a new data center in Montgomery County, Missouri, adjacent to Amazon's $35 billion facility. The move reflects escalating CapEx spending among tech giants racing to secure computing power for AI applications, with Google committing upward of $190 billion this year in infrastructure expansion.

Business2 sources

Google commits $15 billion to build massive data center in Missouri next to Amazon's site

Business2 sources

Bloom Energy and Nebius strike $2.6 billion deal to power AI infrastructure with fuel cells

Bloom Energy has partnered with Nebius, a NVIDIA-backed European AI cloud provider, in a $2.6 billion agreement to deploy fuel-cell technology across data centers in Europe. The deal addresses critical power constraints hampering AI infrastructure buildout, with the first phase expected operational this year featuring 328 MW of installed capacity.

Business2 sources

Bloom Energy and Nebius strike $2.6 billion deal to power AI infrastructure with fuel cells

Business2 sources

AMD EPYC Venice enters production as first 2nm HPC chip with 256 cores and 70% performance leap

AMD has begun production ramp of its next-generation EPYC processor codenamed Venice on TSMC's 2nm process, making it the first high-performance computing product to reach this milestone. The 256-core EPYC Venice promises a 70% compute performance gain over current EPYC Turin chips and will feature Zen 6 architecture, positioning AMD to extend its 46% server CPU revenue share.

Technology3 sources

AMD EPYC Venice enters production as first 2nm HPC chip with 256 cores and 70% performance leap

Technology3 sources

Nvidia Vera Rubin AI systems now cost $7.8 million as memory costs explode 435% to $2 million per rack

Nvidia's next-generation Vera Rubin AI infrastructure faces a dramatic memory price surge, with a single VR200 NVL72 rack now estimated at $7.8 million. Memory costs have jumped 435% compared to previous systems, now accounting for over $2 million—or 25% of the total system cost. The spike reflects surging demand for HBM4 & LPDDR5X memory and 3D NAND storage as cloud service providers race to deploy cutting-edge AI infrastructure.

Business3 sources

Nvidia Vera Rubin AI systems now cost $7.8 million as memory costs explode 435% to $2 million per rack

Business3 sources

Broadcom AI Revenue Projected to Reach $44B by 2026 as TPU Strategy Gains Momentum

Broadcom strengthens its position in the AI semiconductor market through a new packaging partnership with Applied Materials while analysts forecast AI revenue climbing to $44 billion by 2026. The company's TPU strategy and major customer wins from Google, Anthropic, and OpenAI position it as a growing competitor to NVIDIA in AI compute and networking technologies.

Technology2 sources

Broadcom AI Revenue Projected to Reach $44B by 2026 as TPU Strategy Gains Momentum

Technology2 sources

SpaceX files for record IPO, stakes future on $26.5 trillion AI market and orbital data centers

Elon Musk's SpaceX submitted a 400-page S-1 filing revealing ambitious plans to dominate AI through space-based data centers. The company reported $18.67 billion in 2025 revenue but lost $4.94 billion, largely due to AI investments following its xAI acquisition. SpaceX claims a $28.5 trillion total addressable market, with $26.5 trillion coming from AI alone, as it prepares for what could be the largest IPO in history.

Business14 sources

SpaceX files for record IPO, stakes future on $26.5 trillion AI market and orbital data centers

Business14 sources

Anthropic to pay SpaceX $1.25 billion monthly for compute power in $15 billion annual deal

Anthropic will pay SpaceX $1.25 billion per month through May 2029 for access to compute power at Colossus data centers, totaling over $40 billion. The deal emerged from SpaceX's IPO filing and highlights how compute capacity has become a critical bottleneck in AI development. Meanwhile, Anthropic approaches its first quarterly profit with revenue expected to exceed $10 billion.

Business12 sources

Anthropic to pay SpaceX $1.25 billion monthly for compute power in $15 billion annual deal

Business12 sources

Intel CEO bets on agentic AI and foundry revival as chipmaker eyes 10A and 7A process nodes

Intel CEO Lip-Bu Tan is transforming the chipmaker after bankruptcy fears initially scared away top recruits. The company has slashed management layers from 12 to five, secured billions in backing, and is now betting on agentic AI and inference workloads to drive recovery. With 18A yields improving 7% monthly and 14A ahead of schedule, Intel is already mapping out its 10A and 7A roadmap.

Business2 sources

Intel CEO bets on agentic AI and foundry revival as chipmaker eyes 10A and 7A process nodes

Business2 sources

Prev 1…474849…105 Next