Adept AI
The business, the turning points, and the decisions behind them.
01 / WHAT IT BUILT
What the company did
Adept AI was a San Francisco startup founded in 2022 by ex-Google Brain and OpenAI researchers David Luan, Ashish Vaswani and Niki Parmar. It set out to build AI agents that could perform actions on any software tool using natural language: read the screen, click the buttons, fill the fields. It shipped the ACT-1 model and published Fuyu multimodal models, raising over $415 million at a valuation around $1 billion from backers including Nvidia, Atlassian, Workday and Greylock.
02 / Adept AI FAILURE REASONS
Failure factors supported by the current evidence
Adept concluded that pursuing foundation models and an enterprise-agent product at the same time would require substantial fundraising effort. It narrowed the company to agentic solutions while founders and part of the team moved to Amazon under a technology-licensing arrangement.
Maintaining the original dual strategy would have required substantial continuing fundraising for foundation-model development.
The company reduced its original scope to concentrate on agentic products.
03 / EVENT HISTORY
Events in the structured record
- Strategy narrowed
Adept ended its original dual-track strategy and focused the remaining company on agentic solutions.
- Team and technology agreement
Adept's co-founders and some employees joined Amazon, which licensed Adept technology and models.
04 / CONTEMPORARY REBUILD ASSESSMENT
Difficulty of rebuilding this project today
Material for this section is being gathered and checked. It will appear here once ready.
Overall assessment: Not completed
05 / CAUSAL CHAIN
How the factors connect
- 01
AI agents that use software via natural language
- 02
$415M raised at ~$1B valuation
- 03
Amazon licensed tech and hired co-founders in June 2024
- 04
Left as a shell licensing entity; independent thesis ended