Startup Watch
Nebius buys Inferize for up to $150M to cut cold starts
Nebius paid a reported $100-150 million for Inferize, a 17-person startup founded in January that cuts GPU idle time when AI models start up.
Inference news and analysis
Inference, meaning running a trained model to answer requests: chips built for that job, such as Nvidia's dedicated inference accelerator and OpenAI's own silicon, performance-per-watt claims, and API pricing for hosted models. Each story says whose numbers they are and what they leave out.
New & Next
Sony QSSR brings AI upscaling to the base PS5
Sony's QSSR, built with AMD, brings PS5 Pro-style AI upscaling to the standard PS5. Marvel's Wolverine and Ghost of Yōtei added it on October 1.
AI & LLMs
Laya's typed decisions, explained
A typed-decision model fills in fixed questions with typed answers and a confidence score. Laya's 421M open weights let you verify that yourself.
AI & LLMs
Liquid AI LFM2.5-VL-DSpark speeds up its 3B vision model
Liquid AI's 280M-parameter LFM2.5-VL-DSpark draft model decodes its 3B vision model up to 3.13x faster on an M5 Max, with identical output.
AI & LLMs
Laya is a 421M open-weights answer to Jev
Laya returns typed decisions in 32.8 milliseconds from 421M parameters under Apache 2.0. Its zero-shot accuracy is close to chance.
AI & LLMs
TypeSafe's Jev returns typed decisions, not text
Jev answers in 70 to 500 milliseconds and charges $0.042 per million input tokens, with output free. It cannot generate text at all.
Dev Stack
Fujitsu MONAKA ships 144 Armv9 cores in November
Fujitsu's 144-core MONAKA goes on sale in November, with 2nm cores stacked over 5nm cache and GCC 15 support already upstream.
AI & LLMs
Qwen3.5-9B spends 32KB per token on KV cache
Qwen3.5-9B runs full attention on only 8 of its 32 layers, so its KV cache costs 32KB per token. That ratio decides whether it fits in 8GB.
AI & LLMs
OpenAI says its Jalapeño chip beats Nvidia's GB300 per watt
OpenAI's first published Jalapeño benchmarks claim 1.5 to 1.9 times more work per kilowatt than Nvidia's GB300, at half the power. The numbers are OpenAI's own.
AI & LLMs
Nvidia's Groq 3 LPX inference chip enters full production
Nvidia put its dedicated inference accelerator into full production, splitting agent workloads so GPUs handle context and the new chips handle token generation.
AI & LLMs
DeepSeek warns developers a significant API price rise is coming
DeepSeek warned developers on August 6 that API prices will rise 'significantly' soon, with no figures or dates - while V4-Flash is the cheapest well-known model to run.
The weekly digest
One email every Friday with the week's top stories from all six desks: AI, dev tools, coding, the tech industry, startups and what's next. Free, no spam.