TokenPost topic
AI inference
49 TokenPost stories are tagged AI inference. The newest appear first.
Latest AI inference stories
- Prime Intellect to Acquire Moondream to Accelerate AI Inference
The deal would add Moondream’s open-source vision models and inference software to Prime Intellect’s infrastructure stack. Financial terms and the closing date were not disclosed.
· Business
- Intel Outlines Crescent Island GPU for AI Inference; Sampling Expected in 2H 2026
The 350-watt, air-cooled PCIe card uses Xe3P architecture and includes 32 Xe cores, 256 XMX engines and 160 GB of LPDDR5X memory.
· Technology
- CoreWeave Forge Shows 15 Times Faster Model Checkpoint Loading in Test
The platform connects inference, training, evaluation and agent development, while a preview feature loads updated checkpoints into live deployments without redeployment.
· Technology
- Xeal Proposes Laitent Network to Convert Idle EV Capacity Into AI Compute
The proposed network would place pods with up to 48 Nvidia GPUs at EV-charging properties and target metropolitan response times below 20 milliseconds.
· Technology
- Forum Markets Forms U.S. AI Venture With 11-Megawatt Plan
The majority-owned venture with Edge Node Inc. will deploy GPU capacity at powered data centers and mobile-tower sites, with 2027 revenue guidance of $125 million to $175 million.
· Business
- Nebius Integrates Inferize Technology Into Token Factory
The acquisition adds engineering work focused on reducing delays when AI models start and scale. The deal’s financial terms were not disclosed.
· Technology
- CoreWeave Puts NVIDIA Vera Rubin NVL72 Into Production for Cloud Customers
Cognition’s early tests found up to 4.8 times the total token throughput of NVIDIA’s GB200 NVL72 in software-engineering AI inference tasks.
· Technology
- Modal Labs Nears $750 Million Round at $15.75 Billion Valuation
The proposed financing would more than triple the AI infrastructure company’s valuation four months after its previous fundraise.
· Business
- Pearl Links AI Inference to Mining Through Proof-of-Useful-Work
The Layer-1 network uses matrix multiplication for mining and AI workloads, while a Together AI endpoint applies PRL emissions to discount inference costs by more than 25%.
· Technology