Optum published a practical comparison of Apple Silicon vs. NVIDIA GPUs for local LLM inference covering model capacity, Time to First Token (TTFT), output tokens/s, software support, power, noise, and multi-user serving.
Public source
Publisher name
Public post
I commonly read comparisons between Apple and NVIDIA GPUs for serving local LLMs. The posts which favor Apple lean on Apple's unified memory architecture for running lar…
Company
Optum
- Industry
- Hospitals and Health Care
- Location
- Eden Prairie, US
- Company size
- 10,001+ employees
About Optum
We’re evolving health care so everyone can have the opportunity to live their healthiest life. It’s why we put your unique needs at the heart of everything we do, making it easy and affordable to manage health and well-being. We are delivering the right care how and when it’s needed; providing support to make smarter and healthier choices; and making prescription services easier, while helping you save money along the way. It’s everything health care should be. Together, for better health. Optum is part of UnitedHealth Group (NYSE: UNH).
See moreLatest activity
Latest activity from Optum
68 signals
Research & Knowledge
Optum conducted a survey of adults aged 27 to 64 about their experiences before a health care visit to learn what people want from their experience before care begins.
People
Optum is offering applications for the 2027 Optum Advisory Consultant Development Program for recent graduates and upcoming bachelor's or master's graduate candidates.
Presence & Recognition
Optum is coming to Chicago to connect with healthcare leaders and peers to discuss data-driven innovation and AI in managing costs and advancing clinical excellence.
Discover more
Similar signals
Similar public activity from other companies.
Research & Knowledge
Marktechpost AI Media Inc
Marktechpost AI Media Inc published a research paper titled Perplexity Open Sources Lily: A Rust + Metal Inference Engine for Qwen3.6-35B-A3B on Apple Silicon.
Research & Knowledge
General Compute
General Compute completed a review of the landscape of ASICs in AI, highlighting that nine out of fifteen have shipped silicon in volume, with d-Matrix hitting full production in June, FuriosaAI in mass production since January, Positron AI, SambaNova, and Tenstorrent shipping today, and Rebellions' Rebel100 due in the second half.
Research & Knowledge
Apple
Apple discussed vLLM on TPU with TorchTPU and Ray at Ray Summit.
Research & Knowledge
NVIDIA
NVIDIA published Artificial Analysis' 100K context inference benchmark results showing that NVIDIA Groq 3 LPX generated 3,431 output tokens per second on the Gemma 4 31B model at 100K context, nearly 4x the next fastest public endpoint on this benchmark.
Research & Knowledge
Efficient Computer