Google Cloud is winning new AI startup customers at a rapid pace. The unit added coding startups Lovable and Windsurf to its roster this week. Both companies now run their products on Google Cloud and on Gemini models. These deals show how startups can shape cloud growth.
Google Cloud has been growing fast in 2025. The unit reported an annual run rate above $50 billion, and it said it has about $58 billion in contracted revenue to convert over the next two years. Those figures underline why cloud traffic for training and running AI models is rising.
AI startups choose cloud platforms for compute and for model access. Google offers startup credits and access to dedicated GPU clusters. It also gives tight integration with Gemini models and with Vertex AI services. These advantages make it easier for small teams to build products that require heavy computing.
Why it matters
The global cloud market is expanding fast as generative AI takes hold. Industry trackers expect cloud spend to top four hundred billion dollars in 2025. Cloud growth opens paths for vendors that can serve AI workloads well. Google Cloud is trying to turn early startup deals into long-term customers.
Google says it works with nine of the ten leading AI labs and with a large share of generative AI startups worldwide. The company highlighted that many startups that start with cloud credits later scale into heavy users. That creates steady demand for storage compute and specialized AI services.
What startups get
Startups get credits and engineered stacks that reduce time to market. Google for Startups Cloud Program gives substantial cloud credits to qualifying companies. The firm also offers cluster options that bundle Nvidia GPUs for model training and inference. For early-stage firms, these resources lower the cash needed to launch and test models.
Lovable and Windsurf are not the largest customers by spend. The bet from Google is that they will scale. As they grow, they will consume more compute and more managed AI services. That potential growth is why cloud providers court promising small companies as strategic customers.
Costs and trade-offs
Training and operating AI models are expensive. Startups face high cloud bills for training, fine-tuning, and inference. That cost pressure makes credits and efficient model tooling valuable. Google positions its stack to reduce engineering work through managed services and model integrations. Startups balance cost and capability when they pick a cloud partner.
Ecosystem and events
Google hosted its first AI Builder’s Forum this week and announced more than forty new startups building on Google Cloud. The event gathers founders and engineers to share best practices and to spotlight tooling that supports prototypes and scale. Forums and programs help Google stay close to the teams that may become major customers.
What this means for the market
Cloud vendors will compete on price performance and on AI tooling. Providers that offer effective model stacks and that help startups reduce ramp time stand to win long term. For buyers, the choice of cloud will shape where models are trained, hosted, and served. For the industry, the move underscores how AI demand drives broader cloud expansion.
Google Cloud is building momentum by signing fast-rising startups and by packaging generous credits, GPU access, and model integrations. Those steps are fueling demand for cloud infrastructure and are shaping where future AI services will run. Observers will watch whether these early bets turn into major long-term customers.
