GKRootWire
AI Stripe's OpenRouter Buy Is About Payments, Not the SingularityGadgets Amazon Sets Sights on 500 Neighborhoods for Drone Delivery by 2026Security Kansas Police Department Pulls the Plug on Flock License Plate CamerasAI ChatGPT Goes Down Hard as Logins and Signups BreakDev Tools New Algorithm Speeds Up Day-of-Week CalculationsDev Tools Why 'Turns' Might Beat Radians for Angle Math in CodeAI Stripe's OpenRouter Buy Is About Payments, Not the SingularityGadgets Amazon Sets Sights on 500 Neighborhoods for Drone Delivery by 2026Security Kansas Police Department Pulls the Plug on Flock License Plate CamerasAI ChatGPT Goes Down Hard as Logins and Signups BreakDev Tools New Algorithm Speeds Up Day-of-Week CalculationsDev Tools Why 'Turns' Might Beat Radians for Angle Math in Code
Cloud & Sysadmin

Groq Raises $350M as It Shifts From AI Chips to Neocloud Hosting

The inference-chip startup is now valued at $3.5 billion as it builds out Nvidia-powered data centers instead of just selling its own silicon.

Groq, once known primarily for its custom LPU chips designed to accelerate AI inference, has raised $350 million in new funding at a $3.5 billion valuation. The money will fuel the company's shift toward becoming a "neocloud" — a specialized cloud provider that rents out GPU and AI compute capacity rather than only selling chips.

Notably, Groq is expanding its data center footprint using Nvidia hardware, a shift from its original pitch of offering an alternative to Nvidia's dominance in AI accelerators. This suggests the company sees more near-term opportunity in renting out compute than in convincing customers to switch chip architectures.

The move puts Groq in more direct competition with other neoclouds like CoreWeave and Lambda, all racing to meet surging demand for AI inference capacity.

Why it matters: This pivot signals that even chip-first AI startups are finding it more lucrative to sell compute-as-a-service than hardware alone, especially as inference workloads explode. It also reflects the reality that beating Nvidia at its own game is harder than renting its chips profitably.

Sources: TechCrunch