When Kevin Scott, Microsoft’s chief technology officer, takes the Disrupt stage later this month, he will do more than tout new features: he will sketch a vision of how Microsoft intends to be the infrastructure and safety layer for a rapidly expanding AI economy. Scott’s appearance at TechCrunch Disrupt 2025 is being framed inside the company’s landmark partnership with OpenAI and a string of product integrations that have remade both consumer and enterprise software.

“AI is not a single product — it’s a new substrate for software,” Scott told TechCrunch in a preview interview. “Our job is to make that substrate accessible, useful and trustworthy for every developer and every organization.” He is expected to detail how Microsoft combines cloud compute, developer tools and model access into what executives describe as a platform play: not only selling Azure compute but embedding AI into Office, Dynamics, and GitHub to lock in enterprise value.

The partnership with OpenAI — which has evolved from an early investment and licensing pact into deep technical collaboration — is central to that thesis. Microsoft has integrated large language and multimodal models into customer-facing services, offered hosted model access through Azure, and calibrated its go‑to‑market to pair Azure with software suites that reach millions of users. For startups, the upside is immediate: access to state‑of‑the‑art models and deployment infrastructure that would otherwise be prohibitively expensive.

But the relationship raises familiar tensions. Startups and venture investors have praised the lower barrier to building AI applications while warning of dependency and concentration. “When the platform controls the models and distribution, it can accelerate innovation — and crowd out it at the same time,” said an early‑stage investor who follows enterprise AI. Scott has acknowledged that dynamic, framing Microsoft’s role as enabling partners while guarding against anti‑competitive behaviors.

Security and governance will figure heavily in his remarks. Microsoft has publicly committed resources to model safety, red teaming and detection tools meant to curb misuse and deepfakes. “We have to engineer for trust,” Scott said. “If the foundations are unsafe, nobody benefits.” He is expected to articulate how Microsoft balances product velocity with safeguards, and how it is working with policymakers on standards and oversight.

For founders on the Disrupt floor, the message will be both opportunity and caution. Access to models and go‑to‑market channels can speed product development and customer acquisition, but reliance on one platform invites strategic risk. Venture activity has shifted accordingly: investors funnel funding to niche, vertical AI startups that promise defensibility even in a world dominated by a few model providers.

Scott’s Disrupt appearance comes at a moment when regulators are sharpening their focus, competition among cloud providers is intensifying, and enterprises are wrestling with operational questions about data, compliance and staff retraining. Microsoft’s bet is that the company’s scale and product portfolio position it to shape that transition — provided it can convince partners that the platform is an enabler rather than a gatekeeper.

As the industry grapples with both opportunity and risk, Scott’s pitch will test whether Microsoft can convert technological advantage into an ecosystem that preserves room for independent startups, rigorous safety practices and a competitive market. The answers will matter to enterprises choosing their AI stack, to startups deciding where to build, and to policymakers weighing how much power a handful of companies should hold over the next generation of computing.