45k user reviews & 1 million votes, over 29k web hosting brands & 85k plans since 2004!
Menu
WHTop → 📋 News → Fractile Secures $220M to Advance AI Inference Chips
May, 2026 : Fractile Secures $220M to Advance AI Inference Chips
📅 - Fractile has raised $220 million to accelerate development of inference hardware designed for large AI models that increasingly spend more time generating outputs than training them. The London-founded startup argues today's chips are already struggling with workloads involving tens of millions of tokens, where responses can take weeks. Investors include Accel, Founders Fund, and Factorial Funds alongside several returning backers globally.
For the past two years, the AI industry has obsessed over training. Bigger clusters, larger parameter counts, endless procurement battles over NVIDIA GPUs. Fractile thinks the more immediate problem sits elsewhere now: inference latency.
The company, [...][... Check source for end of article ...]
📅 - NorthC Restarts Dutch Data Center After Major Fire - NorthC has restored temporary power and cooling at its fire-damaged Almere data center in the Netherlands, enabling customers to begin phased restarts of critical IT systems nearly a week after a major blaze disrupted Dutch digital infrastructure services. Engineers activated the temporary redundant power installation and cooling systems this morning at 09:30 am (CET), while investigators separately launched an inquiry into the cause of the fire and the timeline for reconnecting the facility to the regular power grid.
The immediate crisis at NorthC's Almere facility is shifting into something more operationally awkward. Customers are now being brought back online in phases rather than all [...]
📅 - Saturn Cloud And Mirantis Build AI Stack For GPU Clouds - AI infrastructure providers are discovering that renting GPU capacity is the easy part. Operating a usable AI platform at scale is becoming the harder business problem. Saturn Cloud and Mirantis are now trying to close that gap through a partnership that combines Kubernetes-native infrastructure automation with managed AI development environments for enterprises and neocloud operators worldwide today.
The timing is not accidental. GPU infrastructure demand continues to outpace operational maturity across much of the market. Plenty of providers can rack NVIDIA systems, deploy Kubernetes clusters and expose compute capacity through APIs. Fewer can offer an integrated environment where AI [...]
📅 - HPE Expands Private Cloud Stack for AI and Hybrid Data - HPE is widening its private cloud and data infrastructure portfolio as enterprises struggle with a less glamorous side of AI adoption: operational sprawl. On Tuesday, the company unveiled updates across HPE GreenLake, Alletra Storage, Zerto, and Data Fabric software, targeting customers trying to stitch together virtual machines, containers, AI data pipelines, and ransomware recovery without adding another layer of management overhead or infrastructure debt across sprawling environments today.
HPE's pitch is straightforward enough. AI workloads are creating pressure far beyond GPUs. Enterprises suddenly need faster storage architectures, more predictable recovery systems, Kubernetes [...]
📅 - Casimir Raises $12M to Commercialize Quantum Energy Chips - Quantum energy startup Casimir has raised an oversubscribed $12 million seed round to commercialize semiconductor chips designed to harvest energy from quantum vacuum fields. The company, founded by former NASA propulsion researcher Harold "Sonny" White, says its battery-free MicroSparc chips could power embedded electronics continuously without charging, replacement cycles, cables, or conventional external power infrastructure for devices.
The pitch sounds almost deliberately provocative.
A 5mm by 5mm semiconductor chip producing continuous electrical power without a battery. No charging port. No replacement cycle. Casimir says the first-generation MicroSparc device can generate 1.5 [...]
📅 - Telehouse Canada Retrofits Toronto Campus for AI Density - Telehouse Canada has completed a liquid-cooling overhaul across its downtown Toronto data center campus as AI infrastructure requirements begin reshaping colocation economics in major urban markets. The KDDI-owned operator says the upgrade allows rack densities reaching 120 kilowatts while preserving low-latency interconnection access - a combination increasingly demanded by enterprises deploying inference-heavy workloads near end users globally.
The timing is difficult to ignore.
AI infrastructure is rapidly colliding with older data center assumptions about power density, airflow, and urban real estate constraints. Facilities originally designed around conventional enterprise IT loads [...]
📅 - DataBank Adds Houston Solar Capacity Amid AI Power Surge - DataBank plans to install a 3.15MW rooftop solar array at its Houston HOU3 data center, adding onsite generation capacity as AI-related electricity demand intensifies pressure on US power markets. The company says the installation could become Houston's largest rooftop solar deployment while functioning less as environmental symbolism and more as long-term infrastructure risk management for operators facing volatile electricity economics today.
Data center operators have spent years talking about renewable energy. Usually from a distance. Power purchase agreements somewhere in rural America. Renewable energy certificates detached from the facilities actually consuming electricity. Enough [...]
📅 - 
📅 - 
📅 -