The company closed a Series B funding round earlier this summer at a $4 billion valuation. This scaling occurs as laboratories prioritize high-quality data to improve model performance, with analyst Ali Sarinik estimating that as much as 10% of US GDP could soon be dedicated to training AI models.
