
Argmax is a technology company providing an on-device AI platform that enables developers to run foundation models locally on users' devices. Their flagship product, the Argmax SDK, supports private, real-time, and cost-predictable AI applications by leveraging device NPUs instead of cloud or CPU/GPU resources. Key products include WhisperKit Pro for speech-to-text, SpeakerKit Pro for speaker recognition, and DiffusionKit Pro for advanced diffusion models on Apple Silicon. Argmax operates a subscription-based business model with free open-source options and paid Pro and Enterprise plans, targeting developers and enterprises seeking scalable, private AI solutions. The company was founded by former Apple ML research and inference engine leaders and is headquartered in Palo Alto, USA.

Argmax is a technology company providing an on-device AI platform that enables developers to run foundation models locally on users' devices. Their flagship product, the Argmax SDK, supports private, real-time, and cost-predictable AI applications by leveraging device NPUs instead of cloud or CPU/GPU resources. Key products include WhisperKit Pro for speech-to-text, SpeakerKit Pro for speaker recognition, and DiffusionKit Pro for advanced diffusion models on Apple Silicon. Argmax operates a subscription-based business model with free open-source options and paid Pro and Enterprise plans, targeting developers and enterprises seeking scalable, private AI solutions. The company was founded by former Apple ML research and inference engine leaders and is headquartered in Palo Alto, USA.
Core product: Argmax SDK — on-device AI platform for running foundation models locally
Technical focus: Model compression and on-device inference using device NPUs (Apple Silicon support noted)
Notable products: WhisperKit Pro (speech-to-text), SpeakerKit Pro, DiffusionKit Pro
Business & licensing: Open-source options plus subscription Pro and Enterprise plans
Funding signal: Seed round (Nov 1, 2024) led by Salesforce Ventures; participation from General Catalyst, Michele Catasta, Julien Chaumond, Amjad Masad
On-device AI inference and model compression to reduce reliance on cloud GPUs/CPUs and improve privacy/latency.
Software
8000000.00
Seed round with five investors including Salesforce Ventures (lead), General Catalyst, Michele Catasta, Julien Chaumond, and Amjad Masad.
“Seed round led by Salesforce Ventures with participation from General Catalyst and individual investors Michele Catasta, Julien Chaumond, and Amjad Masad.”