While major tech companies race to build trillion-parameter AI models that demand massive GPU clusters, Fastino is taking a radically different path. The Palo Alto startup has designed a new AI architecture that trades size for speed, precision, and affordability.
This lean approach is catching the attention of top investors. Fastino has just closed a $17.5 million seed round, led by Khosla Ventures, the same firm that backed OpenAI in its earliest days. This new investment brings Fastino’s total funding to nearly $25 million.
The company had previously raised $7 million in a pre-seed round led by Microsoft’s M12 and Insight Partners last November.
Small Models, Big Results: Fastino’s Task-Specific AI
Fastino’s core innovation lies in its deliberately small and task-specific AI models. These models can be trained using low-end gaming GPUs worth less than $100,000, according to co-founder and CEO Ash Lewis. That’s a stark contrast to the multimillion-dollar training infrastructure required by the likes of OpenAI or Google DeepMind.
“Our models are faster, more accurate, and cost a fraction to train while outperforming flagship models on specific tasks,” Lewis told TechCrunch.
Rather than building one massive model to handle everything, Fastino offers a suite of narrow, purpose-built models. Examples include tools that automatically redact sensitive data or summarize long corporate documents. Each model is optimized for a specific enterprise use case.
Fastino isn’t publicly sharing its client list yet, but early feedback has been enthusiastic. In one demo, Lewis showcased a model delivering a detailed response in milliseconds — outputting the entire answer as a single token. This lightning-fast performance could make Fastino’s models appealing to industries where speed and accuracy are mission-critical.
Fastino’s Edge in a Crowded Enterprise AI Market
The enterprise AI market is fiercely competitive, with players like Cohere, Databricks, Anthropic, and Mistral all offering solutions tailored to business users. However, Fastino’s focus on minimalist models — built for specific jobs rather than general use — could give it a unique edge.
Many in the AI community believe the future of enterprise generative AI lies in smaller, more efficient models. These systems are easier to train, faster to deploy, and simpler to secure. Fastino’s approach aligns perfectly with that vision.
What sets Fastino apart isn’t just its model size, but its contrarian philosophy. “Our hiring strategy is very much focused on researchers that maybe have a contrarian thought process to how language models are being built right now,” said Lewis.
Rather than chasing benchmarks or chasing size records, Fastino is building a team focused on real-world utility and operational speed. The company is actively recruiting top AI researchers who value efficiency over scale.
Backed by a major vote of confidence from Khosla Ventures, Fastino now plans to grow its team and refine its enterprise offerings. The goal is to deliver AI tools that work faster, cost less, and outperform bulky general-purpose models in targeted tasks.
Though it’s still early days, Fastino’s vision of compact, high-performance models could make a significant impact — especially as more enterprises look for affordable, scalable, and secure AI solutions that don’t require deep infrastructure investment.
As the AI landscape shifts from “bigger is better” to “smarter is better,” Fastino is positioning itself as a pioneer of the next generation of AI tools for business.
