Capital & commercial trajectory
Funding rounds and reported revenue
Each point matches a dated itemized record below. Funding points show the disclosed amount for that round, not a running total. Revenue remains a separate series and may represent ARR or run-rate rather than audited revenue.
Valuation, debt or credit, acquisition value, and market cap remain distinct records below and are not combined with funding or revenue.
A structured financial history has not been completed for Zyphra.
The profile will not manufacture a chart from incomplete or unsourced figures. Funding rounds, valuation events, debt facilities, and reported revenue must be dated and linked to a public source first.
Current snapshot
Lifecycle position
Research
Early research & prototypes
Preview
Limited access & evaluation
Growth
Public access & scaling
Current position
Deployment
Production & at scale
Founded
2021
Watch cadence
Weekdays
Last verified
2026-07-23
Focus
Open multimodal foundation models, sovereign AI infrastructure, and scientific AI
Latest move
Released ZUNA1.1 on July 16, 2026, a 380M-parameter Apache-2.0 EEG foundation model trained on about 3.5M channel-hours for reconstruction, denoising, and upsampling.
Category and subcategory context
Competitive context
Recent models
Products
Open weights
Yes
API access
Yes
Watch level
Weekdays
Signals
Infrastructure & distribution
Where Zyphra sits in the AI stack
Model layer
3 tracked model families
Access layer
Managed API + self-hostable weights
Market layer
Open Weight Labs
Public product surface
Stack signals
This section reflects publicly visible product and access signals. Undisclosed GPU, cloud, training, and data-center contracts are not inferred.
Recent coverage of Zyphra
Recent business and product coverage pulled into the profile so you can see what is happening around the company without leaving the page.
Zyphra Releases ZUNA1.1: An Apache 2.0 EEG Foundation Model With Variable-Length Inputs From 0.5 To 30 Seconds
Zyphra Releases ZUNA1.1: An Apache 2.0 EEG Foundation Model With Variable-Length Inputs From 0.5 To 30 Seconds MarkTechPost
Zyphra Release Zamba2-VL: Hybrid Mamba2–Transformer Vision-Language Models That Cut Time-to-First-Token by About an Order of Magnitude
Zyphra Release Zamba2-VL: Hybrid Mamba2–Transformer Vision-Language Models That Cut Time-to-First-Token by About an Order of Magnitude MarkTechPost
Zyphra Releases ZAYA1-8B: A Sub-1B Active Parameter MoE Model That Outperforms Much Larger Rivals
Zyphra Releases ZAYA1-8B: A Sub-1B Active Parameter MoE Model That Outperforms Much Larger Rivals quasa.io
AI model developer Zyphra raising $500 million at $5 billion valuation in funding round that includes commitment from AMD- Forbes
AI model developer Zyphra raising $500 million at $5 billion valuation in funding round that includes commitment from AMD- Forbes TradingView
Zyphra Partners With AMD To Launch An Open AI Platform Powered By A 15MW MI355X GPU Cluster
Zyphra Partners With AMD To Launch An Open AI Platform Powered By A 15MW MI355X GPU Cluster Memeburn
Zyphra Releases ZAYA1-8B Reasoning Model
Zyphra Releases ZAYA1-8B Reasoning Model HPCwire