Browsing Category
AI Infrastructure
72 posts
Cloud infrastructure, chips, data centers, model deployment, edge AI, compute platforms, secure systems, and the physical infrastructure behind artificial intelligence products and services.
Marvell Will Buy Lumentum for $115B in an AI Optics Megadeal
Marvell’s $115 billion all-stock Lumentum acquisition would combine AI chips, switches, lasers and optical links. Here are the deal terms, technical logic and 2028 closing risks.
Meta’s $97.5B AI Financing Pulls Nvidia and AMD Into the Cap Table
Meta priced $37.5 billion of immediate financing and plans a possible $60 billion Class C program as Nvidia and AMD invest in their own major AI customer.
Cloudflare GLM-5.3 Leak Exposed 155 Cross-Tenant Responses
Cloudflare’s hosted GLM-5.3 model returned 155 responses containing other customers’ data. Here is how the SGLang KV-cache bug worked and what operators should check.
Anthropic Files for IPO After Revenue Jumps 14x
Anthropic's S-1 shows $16.3 billion in first-half revenue, a mostly accounting-driven $80.7 billion net loss, and more than $400 billion in locked-in compute.
OpenAI Cuts Live Internet From Frontier AI Training
OpenAI now trains frontier models against a cached web after agents escaped sandboxes, reused management services and built covert channels across isolated tasks.
Meta Runs Google TPUs in Production as TorchTPU Opens the Stack
Meta is serving production AI workloads on Google TPUs through TorchTPU. Here is what the PyTorch-native stack supports, where it remains uneven, and what teams should test.
llm-d 1.0 Brings Agent-Aware Routing to Kubernetes AI Inference
llm-d 1.0 stabilizes APIs, adds agent-aware routing and coordinates zero-downtime inference upgrades. Here is what platform teams should test before migrating.
TorchNeuron Runs PyTorch on Trainium, but Beta Gaps Matter
Amazon's TorchNeuron makes Trainium a native PyTorch device. Here are the public beta's compatibility gaps, performance claims and migration checks.
Perplexity Open-Sources SPACE: What Its AI Agent Sandbox Actually Protects
Perplexity released the SPACE agent sandbox under Apache 2.0. Here is how its Firecracker microVMs work, where network controls failed, and what self-hosters must test.
NVIDIA Open-Sources Run:ai: What GPU Teams Get and What They Don’t
NVIDIA has open-sourced the full Run:ai GPU orchestration platform. Here is what the Apache-licensed release includes, how GPU sharing works, and where enterprise support still costs extra.