Cloud AI Partnerships Move From On-Demand to Multi-Billion Commitments
Cloud AI partnerships are large, long-term infrastructure agreements in which companies commit billions to a single cloud provider to secure compute, storage, and AI model access for mission-critical workloads. These enterprise infrastructure deals mark a shift away from purely pay-as-you-go cloud usage toward strategic alliances that lock in capacity, pricing, and technical roadmaps for years. Two recent examples highlight how central AI has become to cloud strategy. Pinterest has signed its largest-ever infrastructure agreement with Amazon Web Services, while coding startup Lovable is multiplying its Google Cloud footprint fivefold under an expanded multiyear partnership. Both arrangements are designed to support fast-growing AI workloads, from recommendation systems to code-generation agents, and they signal how providers are turning AI infrastructure investment into stable, predictable revenue streams.
Pinterest’s USD 4 Billion AWS Bet on AI Discovery
Pinterest has agreed a planned USD 4 billion (approx. RM18.4 billion) commitment to AWS through 2031, an enterprise infrastructure deal that will fund its next phase of AI-powered growth. The visual discovery platform, which serves more than 600 million monthly users, will deepen use of AWS Trainium and Graviton chips to run large language and vision-language models behind visual search, recommendations, and its Pinterest Assistant. The company is also modernizing its stack by moving from traditional EC2-based setups to Kubernetes on Amazon EKS to improve reliability and developer productivity. “Pinterest is heavily investing in AI to make discovery more personal, visual, and actionable for the hundreds of millions of people who use our platform every month,” said CTO Matt Madrigal. The long-term AWS Pinterest agreement trades short-term cloud flexibility for guaranteed capacity, hardware options, and closer technical alignment.
Lovable’s Fivefold Google Cloud Expansion Tightens AI Dependencies
Lovable, a fast-growing “vibe-coding” startup, is scaling its Google Cloud infrastructure fivefold under a new multiyear agreement that sharply expands its AI footprint. The deal gives Lovable broader access to Anthropic’s Claude models and Google’s Gemini models for coding tasks, tying its growth to Google’s USD 10 billion (approx. RM46 billion) investment in Anthropic and potential follow-on funding linked to performance. Lovable has already crossed USD 400 million (approx. RM1.84 billion) in annualised revenue with 146 employees and claims more than half of Fortune 500 companies as users, making its workloads attractive to Google. Under the expanded partnership, Lovable’s agent will appear in Google Cloud’s Gemini Enterprise Agent Gallery and integrate with Wiz, Google’s USD 32 billion (approx. RM147.2 billion) security acquisition, to detect and fix vulnerabilities in human and AI-generated code in real time.
Why Enterprises Are Accepting Vendor Lock-In for AI Scale
Both the AWS Pinterest agreement and Lovable’s Google Cloud expansion show enterprises are increasingly willing to accept deeper vendor lock-in to secure AI capacity. Instead of purely elastic, pay-as-you-go cloud consumption, these cloud AI partnerships commit customers to specific hardware, model ecosystems, and security stacks over many years. In return, companies secure priority access to cutting-edge accelerators like Trainium, vertically integrated services such as Gemini Enterprise Agent Gallery, and closer integration with security tools like Wiz. For providers, these long-term AI infrastructure investment deals convert volatile AI experimentation into forecastable revenue that justifies massive data center and chip spending. For customers, the trade-off is less about short-term price and more about guaranteed scale, performance, and co-development opportunities as AI moves from pilot projects into the center of product roadmaps.
