Microsoft Is Building Its Own AI Empire — Beyond OpenAI
In a nutshell
The most important AI story of 2026 is not Google I/O. It is not Meta's Muse Spark. It is not OpenAI's advertising platform.
It is Microsoft — quietly, systematically — building the infrastructure to no longer need OpenAI.
The MAI Models — Microsoft's Own AI Stack
Microsoft AI — the tech giant's research lab — announced the release of three foundational AI models that can generate text, voice and images. The models were developed by Microsoft's MAI Superintelligence team, an AI research team led by Mustafa Suleyman, the CEO of Microsoft AI, that was formed and announced in November 2025.
Three models. Three modalities. One clear message.
MAI-Transcribe-1 starts at $0.36 per hour. MAI-Voice-1 starts at $22 per 1 million characters. MAI-Image-2 starts at $5 for 1 million tokens for text input and $33 for 1 million tokens for image output. In an increasingly crowded LLM market, MAI hopes a selling point is that these models are cheaper than those from Google and OpenAI.
Mustafa Suleyman wrote: "At Microsoft AI, we're building Humanist AI. We have a distinct view when creating our AI models — putting humans at the center, optimizing for how people actually communicate, training for practical use. You'll see more models from us soon in Foundry and directly in Microsoft products and experiences."
"More models soon." That is not a product announcement. That is a roadmap signal.
Why This Changes the Microsoft-OpenAI Relationship
Microsoft's $13 billion investment in OpenAI bought it something specific: access to GPT models to power Azure AI services and Copilot. That arrangement made sense when Microsoft had no frontier model capability of its own.
Microsoft is developing its own AI models to reduce reliance on OpenAI. By offering multiple model options including its own, Microsoft can better meet the diverse procurement and compliance requirements of its vast enterprise customer base — strengthening its position as the indispensable platform for modern work.
The logic is straightforward. If Microsoft can offer its own models — cheaper than OpenAI, integrated into Azure, compliant with enterprise procurement requirements — it reduces the revenue share it pays OpenAI for every API call, increases its own margins and eliminates a strategic dependency that has made investors nervous since the relationship began.
OpenAI is already shifting toward Amazon. Amazon committed $50 billion to OpenAI in February 2026. Microsoft's MAI models are the strategic response: if OpenAI is diversifying its cloud partnerships, Microsoft is diversifying its model dependencies.
Microsoft 365 Copilot — The July 1 Pricing Change
Microsoft announced it will expand the availability of AI, security and management capabilities across its Microsoft 365 offerings in 2026, alongside commercial pricing updates effective July 1, 2026. The changes apply across both commercial and government suites.
The 2026 Microsoft 365 updates embed Copilot more deeply into Word, Excel, PowerPoint, Outlook and Teams — transforming it from a passive assistant into a role-adaptive agent. New capabilities include Copilot Studio for building custom workflow bots, AI-generated reporting in Excel and personalized agents that learn corporate context.
July 1, 2026 is six weeks away. Every enterprise Microsoft customer is currently reviewing what the pricing changes mean for their annual budgets. The move to compute-based pricing — which we analysed in our Google I/O Day 2 coverage — is now arriving at Microsoft too. The flat-rate AI subscription era is ending across the entire industry simultaneously.
The Global AI Diffusion Report — The Number That Matters
Global AI adoption continued to rise in the first quarter of 2026. During the quarter, AI usage increased by 1.5 percentage points from 16.3% to 17.8% of the world's working age population. The United States moved from 24th to 21st in national rankings with a 31.3% usage rate. The UAE leads global AI diffusion at 70.1%.
Software developer employment reached approximately 2.2 million, rising 8.5% year over year and marking a record high for the profession. Early data for Q1 2026 shows that software developer employment in March 2026 was about 4% higher than in March 2025.
That last finding contradicts the dominant narrative that AI is eliminating developer jobs. At least for now, the data suggests the opposite: AI coding tools are increasing demand for software developers, not reducing it. The jobs being eliminated are in other cognitive work categories — the ones we documented in our Meta layoffs coverage this week.
Microsoft's Security AI — The Underreported Advantage
Microsoft Research's Vega — announced May 21 — lets users prove who they are while protecting their privacy, using zero-knowledge proof technology integrated with Microsoft's AI identity stack.
Microsoft announced a major step forward in AI-powered cyber defence: a new multi-model agentic scanning harness codenamed MDASH — a cybersecurity AI system that autonomously scans for vulnerabilities across enterprise environments.
Security AI is Microsoft's most underappreciated competitive advantage. While Google competes on Search and consumer AI, while Meta competes on social and advertising AI, Microsoft is embedding AI into the security layer of enterprise infrastructure — where switching costs are highest, procurement cycles are longest and customer relationships are most durable.
What This Means for GAFAM
Microsoft's MAI models, July 1 pricing changes and Copilot deepening across Microsoft 365 represent the most coherent enterprise AI strategy in the GAFAM landscape. Google has consumer scale. Apple has device distribution. Meta has social data. Amazon has infrastructure. Microsoft has enterprise lock-in — and it is tightening that lock with every Copilot update, every MAI model release and every pricing change that makes switching more expensive.
Microsoft is engineering the central nervous system for enterprise operations, aiming to embed AI deeply into core workflows. The upcoming 2026 release wave will further unify workflows, automating complex processes across sales, HR and supply chain operations.
The company that controls enterprise workflows controls enterprise AI budgets. Microsoft understood this before any of its GAFAM peers — and it is executing on that understanding with a precision that its competitors are struggling to match.
The European Perspective
Microsoft's MAI models — priced below OpenAI and Google — arrive at precisely the moment when European enterprises are most sensitive to AI pricing. The July 1 Copilot pricing changes will land on the desks of European IT procurement teams during summer budget reviews. European CIOs who have built AI cost models on current Microsoft 365 pricing are advised to request detailed breakdowns of the July 1 changes immediately — before annual budgets are locked. The MAI models also raise a compliance question: when Microsoft replaces OpenAI's GPT models with its own MAI models inside Copilot, does that constitute a material change to the AI systems European enterprises have already assessed under GDPR and the EU AI Act? The answer is almost certainly yes — and the legal implications for enterprise procurement contracts signed before the MAI transition began are significant. gafam.ai will be watching.
We are not first. We are right.