The Investment in Cloud Infrastructure Keep growing at an important pace. So, In the first three months of 2025 it registered an interannual increase of 21%reaching 90.9 billion dollars, according to the data of the Channels consultant.
It is above all the large companies that have finally recognized that the deployment of AI applications requires a reinforcement of migration to the cloud, which is deriving in large -scale investments both in cloud infrastructure and in the one destined for ia.
On the other hand, to accelerate the business adoption of large -scale AI, the main cloud suppliers are redoubled efforts to optimize their infrastructure. Especially through the development of owner chips, in order to reduce the cost of the use of AI. Also to improve inference efficiency.
In the first quarter of this year, the classification of the three main cloud providers remains the same as in previous: AWS, Microsoft Azure and Google Cloud. Among the three, 65% of world investment in Cloud, with a 24% year -on -year increase in cloud -related spending.
But the growth status is different for each one. Microsoft Azure and Google Cloud had growth rates above 30%, although Google Cloud has dropped a little with respect to the previous quarter. Meanwhile, AWS has only grown by 17%, 19% less than in the last quarter of 2024.
This fall is apparently due to the difficulties of suppliers, who have limited their ability to quickly meet the demand caused by the growth of AI. This has not been a problem for the rest of hyperscales to continue to invest in infrastructure to increase their capacity and prepare for long -term growth.
In general, the Cloud Services market grew solidly between January and March, thanks to the fact that companies focused on two strategic priorities: accelerating migration to the cloud and exploring the possibility of adoption of generative. In the first case, the acceleration was due either as the cloud of more work loads, or the revitalization of the transitions at pause from local.
The takeoff of the generative AI, which depends largely on the cloud infrastructure, has reinforced the cloud strategies of companies, and has accelerated migration times. This has led the main cloud providers, as we have mentioned, to invest more in optimized infrastructure for AI. Hyperscales, in addition, have begun to use proprietary chips, such as Trainium and TPU. Also families of instances developed on purpose, in order to reduce the total costs of AI and improve in inference.
The three main suppliers and their advance in cloud infrastructure
AWS remained market leader in the first quarter of 2025, with 32% of its quota global, and an interannual increase in income of 17%. Its division of AI continues to grow above 100% per year, although it is still in the early stages of its development. Last March, AWS launched a Price reduction strategy to promote the adoption of its Trainium AI chips compared to NVIDIA -based solutions, with a higher cost. To do this, the best value for money of the Trainium 2, between 30% and 40%.
The company has also accelerated the growth of its Bedrock Service, to which the Claude 3.7 Sonnet of Anthropic and calls 4 of Meta has added. In addition, it has become the first cloud supplier to manage the complete Deepseek R1 and Mixtral Mistral. In addition, to underline its long -term commitment to Cloud infrastructure, AWS has announced a capital investment of about 4,000 million dollars to launch a new cloud region in Chile before the end of 2026.
Microsoft Azurethe second main cloud supplier in the first quarter of 2025, has 23% of the market share, and has registered an interannual growth of 33%. According to Microsoft, 16% of that growth is due to AI, which has led to the fastest growing quarter since the second of 2024 for Azure.
Azure AI Foundry, the Microsoft platform for the development and management of applications and AI agents, already has developer users of more than 70,000 companies. The platform processed more than 100 billion tokens between January and March, five times more than during the same quarter of 2024.
The Redmonds have also focused on the reduction of the cost of adoption of AI, which has led them to an improvement close to 30% in terms of their yield with AI with constant energy consumption and a reduction greater than 50% in cost per token. As part of its infrastructure expansion worldwide, the company has opened data centers in a dozen countries in the world between January and March.
As to Google Cloudthe third major cloud supplier in the world, had a 10% market share in the first quarter of 2025 and registered in that period an interannual growth of 31%. Its revenues in the period were 92.4 billion dollars, an amount slightly lower than that of the previous quarter, especially due to supplies problems. Above all, in computeic capacity, which limited its ability to meet the demand of its customers.
Since the beginning of the year, the active use of Google Ai Studio and Gemini’s API has grown approximately 200%, reflecting an increase in the adoption of AI solutions by developers. In addition, Google has opened a new cloud region in Sweden, number 42 worldwide, and has compromised 7,000 million dollars for the expansion of its Iowa data center (United States), with the aim of supporting the growth of its computer and space for cloud and AI work loads.
