Small AI models are taking center stage, and Fastino is leading the charge. The Palo Alto startup just secured $17.5 million in seed funding to expand its suite of compact, task-specific models. The round was led by Khosla Ventures, known for backing OpenAI early on.
Fastino’s models don’t need a warehouse full of GPUs. They run on gaming hardware worth under $100,000 in total. Yet, they still beat larger systems on targeted enterprise tasks. These small AI models are built for speed, accuracy, and low cost—making them a practical choice for businesses.
Unlike the giant general-purpose models, Fastino creates lean tools that do one job well. Whether it’s redacting sensitive data or processing forms, each model is trained for a specific use case. That focus allows companies to get better results without massive infrastructure costs.
Ash Lewis, the company’s CEO and co-founder, says Fastino’s models are faster and more accurate than big-name alternatives. More importantly, they cost less to train and deploy. “We believe smaller AI models can outperform the giants by doing less—but doing it better,” he explains.
Fastino first raised $7 million in a pre-seed round last year, backed by M12 (Microsoft’s venture arm) and Insight Partners. With nearly $25 million now raised in total, the team is scaling fast. Demand is growing as more companies look for agile solutions that don’t break the bank.
This new wave of small AI models is gaining momentum. Businesses want focused tools, not bloated systems. And startups like Fastino are proving that you don’t need size to succeed in AI—you just need precision and purpose.
