Overview
Citi analysts state that demand for open‑weight and open‑source artificial‑intelligence models is rising sharply as regulators impose restrictions on access to frontier proprietary systems.
Performance Gap
The gap between proprietary and open models has narrowed notably after Z.ai introduced its GLM‑5.2 model, improving open‑source offerings on both intelligence and cost metrics.
Competitive Landscape
Citi’s model rankings place open models from Z.ai, DeepSeek, MiniMax (HK:0100) and Moonshot AI increasingly close to proprietary counterparts in intelligence and cost evaluations.
Growth Indicators
- Cohere has tripled its internal annual recurring‑revenue projection for 2027.
- OpenRouter’s share of open‑source tokens processed grew to 65 % in June, up from 34 % in January.
- Fireworks’ open‑source token volume doubled between April and June, reaching 30 trillion tokens.
Funding and Deal Activity
Investors are backing cost‑effective AI deployment: Baseten completed a $1.5 billion funding round, Upscale AI raised $190 million, and Qualcomm announced a planned acquisition of Modular.
Infrastructure and Talent Constraints
More than 300 data‑center bans or moratoriums have been enacted by U.S. localities since 2023, with over 275 introduced since the start of this year. Talent shortages and intense competition for AI researchers and engineers, together with opposition to new data‑center construction, are identified as key bottlenecks that could limit capacity expansion.
Outlook
Citi expects these constraints to favor established AI leaders that possess deep talent pools and extensive computing resources, even as open‑source models continue to capture a larger market share.