Tech Startup Fastino Takes a Unique Approach to AI Model Architecture
Tech giants often brag about trillion-parameter AI models that require expensive GPU clusters. However, Fastino, a Palo Alto-based startup, is changing the game with its innovative approach.
Fastino has developed a new AI model architecture that is intentionally small and task-specific. These small models can be trained using low-end gaming GPUs that cost less than $100,000 in total, setting them apart from the industry norm.
The startup’s unique method has caught the attention of investors, with Fastino recently securing $17.5 million in seed funding led by Khosla Ventures. This brings the total funding to nearly $25 million, following a $7 million pre-seed round last November led by Microsoft’s VC arm M12 and Insight Partners.
According to Ash Lewis, Fastino’s CEO and co-founder, their models are faster, more accurate, and cost-effective compared to flagship models. The company offers a suite of small models tailored to specific tasks such as redacting sensitive data or summarizing corporate documents.
While Fastino has not disclosed specific metrics or users, early feedback from customers has been positive. The small size of their models allows for quick responses, with detailed answers generated in milliseconds.
Techcrunch event
Berkeley, CA
|
June 5
BOOK NOW
In a competitive enterprise AI landscape, Fastino faces challenges from other companies like Cohere and Databricks that also specialize in task-specific AI. However, Fastino remains optimistic about its approach, focusing on building a talented AI team that prioritizes innovation over scale.
With early support from investors like Khosla Ventures, Fastino is poised to make a mark in the AI industry. The startup’s hiring strategy targets researchers who think differently about building language models, emphasizing quality over quantity.
