Samaya, a financial artificial intelligence startup that raised 43 million dollars in 2025, provided forecasts for third-quarter results at 50 large companies before their earnings releases. Its estimates proved more accurate than Bloomberg’s analyst consensus for 40 companies, after analysts won the first four comparisons. The startup notably forecast revenue of 2.30 billion dollars for Palantir, compared with 2.18 billion dollars from analysts, and earnings per share of 0.69 dollar for Permian Resources, versus 0.625 dollar. It performed less well on Yum Brands, Amazon and Google, particularly by underestimating the degree to which revenue would fall short of expectations. The experiment covers only one earnings period and a limited sample, but raises questions about Wall Street consensus biases, investor-relations pressure and the future of artificial-intelligence-based earnings forecasts.
Source: Read the original article


