Skip to content
IFAI

AI compute

Industry tag545 stories

  1. ResearchReach 37

    Google open-sources AlphaChip to design processors in hours

    Google just officially named and released pre-trained weights for AlphaChip, its reinforcement learning system that treats chip floorplanning like a board game. The AI generates superhuman circuit layouts in hours instead of months, and has already designed Google's last three TPU generations as well as MediaTek chips. Why it matters: AI is no longer just running on advanced silicon—it is actively redesigning the hardware that powers it.

  2. IndustryReach 37

    NVIDIA pledges $10M in AI training for developing nations

    NVIDIA just teamed up with the U.S. government and companies like OpenAI and Google on the Partnership for Global Inclusivity on AI. The effort aims to expand access to AI tools in developing nations, with NVIDIA pledging roughly $10 million in free training alongside GPU credits. Why it matters: Frontier AI has been heavily concentrated in wealthy regions, and this effort could help developers in emerging markets build local solutions.

  3. IndustryReach 46

    Sam Altman says superintelligence could arrive in a few thousand days

    OpenAI CEO Sam Altman just published a new essay titled The Intelligence Age, predicting that superintelligence might be just a few thousand days away thanks to deep learning scaling predictably. He argues massive compute and energy infrastructure will be essential to ensure AI drives shared prosperity rather than becoming a resource fought over by the wealthy.

  4. IndustryReach 41

    Qualcomm makes a move on struggling Intel

    Qualcomm just approached Intel about a potential takeover, according to the Wall Street Journal. Intel has been reeling from costly setbacks and slow AI progress, having recently announced $10 billion in cost cuts and over 15,000 layoffs. The megadeal would upend the chip market, though severe antitrust scrutiny makes completion far from guaranteed.

  5. IndustryReach 44

    Microsoft is reviving Three Mile Island for AI power

    Microsoft just signed a 20-year deal with Constellation to restart the shuttered Three Mile Island nuclear plant, exclusively buying 100 percent of its output to fuel expanding AI data centers. The site is slated to come back online in 2028 pending regulatory approval, supplying over 800 megawatts of carbon-free energy. Why it matters: Training and running massive AI models requires enormous electricity, and nuclear power is fast becoming the answer for reliable, clean baseload energy.

  6. IndustryReach 34

    Nvidia puts wireless networks and genAI on the same platform

    Nvidia just launched AI Aerial, a computing suite for designing, simulating, and running AI radio access networks (AI-RAN). It allows telecom providers to optimize 5G and 6G networks while hosting edge generative AI workloads on the exact same infrastructure. Nvidia is also teaming up with T-Mobile, Ericsson, and Nokia to fast-track commercial adoption.

  7. IndustryReach 39

    Intel spins off its chip foundry into independent unit

    Intel just announced plans to spin off its AI-focused foundry business into an independent subsidiary with its own board, clearing the way for outside capital. The stock jumped over 6% following the news. Intel hopes this move and its upcoming 18A process will help turn around steep losses and mount a challenge to Nvidia.

  8. IndustryReach 39

    ByteDance is reportedly designing its own AI chips with TSMC

    ByteDance is reportedly teaming up with TSMC to design two custom AI chips for training and inference, targeting mass production by 2026. The effort looks to bypass costly Nvidia GPUs and navigate tightening U.S. export controls. Why it matters: Going in-house could save billions, though the plans remain in early stages and subject to regulatory shifts.

  9. IndustryReach 62

    Groq taps Aramco for massive AI compute hub

    Groq just partnered with Saudi oil giant Aramco to build a massive AI inference data center in Saudi Arabia. Funded by Aramco with a nine-figure budget, the site will launch by the end of this year with 19,000 LPUs and could scale to 200,000 units. Why it matters: Low energy costs and strategic location could make the Gulf a vital AI compute hub for surrounding regions.

  10. Open sourceReach 54

    Meta flexes Llama's explosive 10x growth

    Meta just dropped new adoption stats for Llama, revealing total downloads on Hugging Face are approaching 350 million—a 10x jump compared to last year. Token volume across major cloud partners more than doubled between May and July 2024, with the massive 405B model driving the highest unique users in August. Open-source AI is no longer just for tinkerers; it's rapidly anchoring enterprise tech stacks.

  11. AI modelsReach 49

    Magic reveals 100M-token context model and $320M in fresh funding

    Magic just announced LTM-2-mini, an ultra-long context model capable of processing 100M tokens—equivalent to roughly 10 million lines of code or 750 novels. The startup claims its sequence-dimension algorithm is 1,000x cheaper per decoded token than Llama 3.1 405B at that context scale, and revealed $320M in new funding alongside a Google Cloud partnership to build supercomputers.

  12. IndustryReach 49

    Nvidia's revenue doubles, but shares dip on sky-high expectations

    Nvidia just reported a massive 122% jump in Q2 revenue to $30.04 billion, driven by explosive 154% growth in its data center division. However, its gross margin dipped slightly to 75.1% due to Blackwell chip design tweaks, sending shares down over 7% after hours as sky-high investor expectations proved tough to beat.

  13. AI toolsReach 41

    Cerebras drops the fastest AI inference on Earth

    Cerebras just launched its new AI inference service powered by the CS-3 system, clocking an eye-watering 1,800 tokens per second on Llama 3.1 8B. That is 20 times faster than hyperscale Nvidia GPUs while retaining full 16-bit precision at just 10 cents per million tokens. Why it matters: lightning-fast generation is critical for multi-step AI agents to finally feel real-time rather than sluggish.

  14. IndustryReach 35

    Qualcomm’s new chip brings on-device AI to midrange phones

    Qualcomm just unveiled the Snapdragon 7s Gen 3, a processor designed to bring generative AI features like on-device translation to $500 to $600 smartphones. Xiaomi will debut the chip next month, trickling down features previously reserved for $800+ flagships.

  15. AI modelsReach 51

    Nvidia puts AI game characters on your RTX PC

    Nvidia just introduced its first on-device small language model, Nemotron-4 4B Instruct, to its ACE digital human suite at Gamescom. Distilled from a 15B model, it is optimized to run locally on RTX GPUs with minimal memory and fast response times. Why it matters: Gamers can now talk to dynamic NPCs in games like Mecha BREAK entirely on local hardware, without relying on cloud servers.

  16. ResearchReach 36

    FinalSpark is renting out living human brain computers for $500 a month

    Swiss startup FinalSpark is letting researchers rent living biocomputers made from human-brain organoids over the internet for $500 a month. Powered by clusters of living neurons trained with dopamine and electrical pulses, the Neuroplatform aims to cut AI energy consumption by 100,000 times compared to conventional silicon chips. Why it matters: As AI power grids strain, synthetic biology could offer a radically green alternative—if scientists can figure out how to program living tissue before the cells die after 100 days.

  17. AI toolsReach 37

    Geekbench launches cross-platform AI benchmark

    Primate Labs just launched Geekbench AI 1.0, bringing standardized performance testing to machine learning workloads across Android, Linux, macOS, and Windows. Serving as the successor to Geekbench ML, the tool gives engineers and enthusiasts a unified way to see how their hardware actually handles AI tasks.

  18. IndustryReach 41

    Microsoft and Palantir bring GPT-4 to classified defense clouds

    Microsoft and Palantir just joined forces to bring Azure OpenAI Service and models like GPT-4 into classified government environments for U.S. defense and intelligence agencies. Palantir will deploy its AIP and Foundry suites across Microsoft's Secret and Top Secret clouds to run operational workloads from logistics to mission planning. Why it matters: This officially pushes frontier LLMs deep into secure military workflows.

  19. IndustryReach 42

    Groq scores $640M to take on Nvidia

    AI chip startup Groq just pulled in a massive $640 million funding round led by Blackrock, valuing the inference specialist at $2.8 billion. The company claims its custom LPUs run generative AI models 10x faster and with a fraction of the energy of conventional GPUs. Meta's Yann LeCun is also joining as a technical advisor, giving the upstart serious ammo against Nvidia.

  20. IndustryReach 43

    Nvidia's flagship AI chip hits a snag

    Nvidia has reportedly delayed its next-gen Blackwell B200 chips by at least three months due to a design flaw discovered late in production. Massive shipments to heavyweights like Microsoft, Meta, and Google are now pushed into early next year. Nvidia says production will still ramp in the second half, but declined to comment on rumors.