Global AI computing capacity is expanding rapidly, doubling every seven months. Since 2022, the total compute available from AI chips has increased by about 3.3 times per year. NVIDIA dominates this space, accounting for over 60% of the total compute capacity, while Google and Amazon fill much of the remaining share. These insights are drawn from revenue data, financial disclosures, and analyst reports, although the coverage varies by manufacturer. For instance, Nvidia and Google's data starts from 2022, while other manufacturers' data won't begin until 2024. The analysis employs a log-linear regression model to track sales of AI chips measured in H100-equivalents. The model estimates substantial growth in computing capacity, with a 90% confidence interval suggesting the real growth could range from 2.7 to 4.1 times per year. This growth is significant, given that the doubling time is pegged at seven months, indicating a strong trend in AI chip sales. There are limitations to the data. While Nvidia's figures date back to early 2022 and Google's to late 2022, other chip manufacturers only contribute data starting in early 2024. However, these manufacturers represent less than 10% of total production as of early 2024, minimizing their impact on the overall growth rate. The report also highlights that while the data tracks chip sales rather than actual deployments, the rapid sales growth means depreciation effects are minor. Overall, this growth trajectory indicates a robust and expanding market for AI computing.