AI & AI Chip Competitors
Articles on emerging AI chip startups and competitors to NVIDIA, including business models, technical architectures, market traction, and contract wins in inference and training markets.
How Etched's Harvard Dropout Founders Turned AI Developer Tools Skepticism Into a $10.3B Chip Empire
Etched raised $300M at a $10.3B valuation led by Sequoia, doubling its value in seven months. Here's what three Harvard dropouts built — and why it doesn't need GPUs.
Qualcomm Bets on 3D-Stacked 'High-Bandwidth Compute' to Break the AI Memory Wall
As AI workloads shift toward memory-intensive inference, Qualcomm is pivoting its datacenter strategy, moving compute closer to memory with a novel 3D-stacked architecture to bypass traditional high-bandwidth memory limitations and improve power efficiency.
AWS's Graviton 5 Is a Chip Triumph Buried Under AI Marketing Noise
Amazon's Graviton 5 delivers 35% faster performance over its predecessor, but AWS and partners keep mislabeling the general-purpose CPU as an "AI chip" — a marketing habit that obscures what Annapurna Labs has actually built and risks the credibility of a $20B+ silicon business.
Beyond the Memory Limit: Transforming LLM Efficiency with Context Compression
Exploring recent technological breakthroughs that enable LLMs to manage long-running agentic tasks by compressing context without accuracy degradation.
Oracle's AI Infrastructure Rush: $70 Billion, $455 Billion Contracts, and a Leap into the Deep End
Oracle is spending $70 billion in reported capex (and an effective $455B in contracted obligations) to build AI datacenters, betting it can outpace rivals in infrastructure speed and scale. Here’s what that means for its balance sheet, margins, and who ends up paying the tab.
FuriosaAI Takes Lisbon — And the Power Math Actually Makes Sense
FuriosaAI's RNGD accelerators are now live at Equinix's Lisbon datacenter. For a South Korean chip startup that's been quietly building power-efficient inference silicon since 2017, the European debut is less about offloading inventory and more about betting on sovereign AI compute — with a chip that sips 180W while delivering serious FP8 throughput.
Dragonfly and the C1000: Qualcomm's Datacenter Gamble
Qualcomm is making a significant push into the datacenter market with its Dragonfly compute platform, aiming to deliver lower total cost of ownership and better performance per watt. The company's C1000 CPU and High-Bandwidth Compute (HBC) technology are set to revolutionize AI datacenters.
Anthropic's Mythos Model: Power Unleashed with New Guardrails
Anthropic has released its Mythos model, previously deemed too dangerous for public release, with enhanced safety measures and a revised data retention policy.
OpenAI's Logging Glitch Causes Unexpected SSD Wear for Codex Users
OpenAI is taking steps to resolve a bug in its Codex platform that causes excessive SQLite logging, leading to unnecessary write cycles on solid-state drives and potentially reducing their total endurance.
Meta's Strategic AI Infrastructure Bet in India: Partnering with Reliance for a 168-Megawatt Data Center
Meta is expanding its AI infrastructure footprint in India through a partnership with Reliance Industries, marking its first major AI data center investment in the country. The collaboration includes a 168-megawatt facility in Jamnagar, Gujarat, and aligns with India's growing prominence as a hub for AI infrastructure.
Rent-a-GPU Outfits Secure Billions in Venture Capital to Meet AI Demand
Rent-a-GPU neoclouds are borrowing billions to keep up with AI demand. Now, Nvidia's new revenue-sharing scheme seeks a direct cut of cloud revenues, shifting risks to emerging providers.
Rebuilding AI from the Ground Up: How Oscillator-Based Computing Could Slash Power Use by 1,000x
Naveen Rao's startup, Unconventional AI, is betting on a radical oscillator-based architecture to cut AI's energy consumption by 1,000x. With a working image-generation model and plans for custom chips, the company aims to redefine AI hardware.