Behind OpenAI’s completion of a massive $110 billion funding round is strategic meaning beyond mere numbers. What emerges from the announcement in late February is a sweeping land-grab in the AI era currently unfolding between Amazon and Microsoft.

In this round, with Amazon investing $50 billion and NVIDIA and another major player each investing $30 billion, one detail worth noting is the order of Sam Altman’s thank-you messages. The sequence—Amazon, Microsoft, NVIDIA, and SoftBank—plainly reflects today’s power dynamics. Microsoft, once the absolute partner, has ended up trailing behind Amazon, the largest new investor.

However, what truly matters is not the funding size. As industry analysts point out, the two technical terms hidden within this announcement are the fault line that separates AI’s present from its future: the stateless API and the stateful runtime environment.

The stateless API is currently the mainstream. When industries such as finance, retail, and healthcare adopt AI, they almost exclusively use this format. It’s optimized for standalone tasks—answering questions, summarizing documents, enhancing search. Companies can incorporate AI capabilities without making major changes to existing systems. That’s why it spread so quickly.

That said, as model performance levels off and computational costs decline, stateless APIs become increasingly commoditized. As a result, profit margins are destined to compress.

In contrast, the stateful runtime environment is an entirely different plane. This is not just an expansion of functionality—it means a shift in the business model itself. AI autonomously executes tasks, keeps context, and coordinates across multiple tools to operate long-term—functioning as digital labor. Company budgets expand from mere API-call costs into automation, process management, and even labor cost reduction.

From 2026 to 2027, it’s expected that nearly all companies’ roadmaps will shift from one-time API calls to “autonomous agent workflows.” If that happens, the market size for the stateful runtime environment will far exceed current projections.

Comparing Microsoft and Amazon’s partnership agreements makes the strategic differences clear. Microsoft secured a $250 billion contract and exclusive cloud provider rights. All stateless API traffic goes through Azure—no matter who the customer is, the billing ultimately flows back to Azure. It’s a highly predictable cash flow, but it carries the risk of shrinking profit margins.

Amazon, meanwhile, secured the foundational hosting rights for the AI agent era through $50 billion in on-the-ground investment and an expansion of its contract to $100 billion. Computing power, storage, and workflow management—all are consolidated in AWS’s execution environment. In other words, while holding onto current cash flows, it is also betting on future productivity structures.

Through this decentralized strategy, OpenAI has greatly strengthened its negotiating power. It used to rely on Microsoft’s infrastructure and was in a weaker position as Microsoft’s 27% equity holder. But by holding several powerful partners at the same time, OpenAI has carved out a path to grow without tying its growth to any single company.

Neither Microsoft nor Amazon can afford to let go of OpenAI in this moment. As long as neither can walk away from the table, the negotiation leverage naturally returns to OpenAI.