ActiveInvestorMag
AI Industrial Complex

Microsoft unveils seven in-house AI models to challenge OpenAI and Anthropic at a fraction of the cost

Microsoft is also launching licensed weights that let companies customise its models and run them in their own Azure environments.

November 17, 2026Evening edition8 outlets Assembled by machine

Watercolour landscape whose skyline traces palantir, 2020 to 2026, on a log scale.
Palantir, 2020 to 2026

Microsoft on Tuesday announced a suite of AI models with powerful coding and cybersecurity capabilities as it races to challenge rivals including Anthropic and OpenAI1. The company launched seven in-house MAI models at its Ignite 2026 conference4,5.

Microsoft AI chief executive Mustafa Suleyman said the company aims to challenge OpenAI and Anthropic at a fraction of the cost2. Microsoft pitched its coding model at 5% of the cost of Anthropic’s Claude, according to one account4.

Alongside the models, Microsoft announced a new business called MAI Licensed Weights3. Companies can take a Microsoft AI model, customise it with their own data and run it in their own Azure environment3,5.

Why it matters to investors

The launch puts Microsoft in direct price competition with OpenAI and Anthropic1,2. Cheaper, specialised models are the core of its pitch, as Microsoft touts customisable models of its own against the two labs2,3,7.

A lower-cost alternative from Microsoft is a challenge to the pricing power of OpenAI and Anthropic in coding and enterprise work2,4.

Suleyman said the licensed-weights approach would ‘feel like open source’, while offering greater security and control7. Keeping the models inside a customer’s Azure environment keeps control of access with the customer3,4.

The offer also ties customers to Microsoft’s cloud5,6. Licensed weights are designed to be customised and run on Azure, Microsoft’s cloud platform5,6. Enterprises that take up the offer would run the models inside that platform3.

Microsoft has forecast stronger-than-expected second-quarter sales and Azure growth8. The new models give the company another product to sell through that cloud business5,8.

What to watch

The key question is how quickly Microsoft can win customers with the new models7. Adoption by enterprises will show whether lower prices move business away from the rival labs2,7.

The cost claims will also be tested4. Microsoft’s comparison with Claude on price is its own claim, and customers will judge whether performance matches the price1,4. Suleyman’s promise of more security and control will face the same test7.

The models’ cybersecurity capabilities are a second area to follow, alongside coding1. Those are the two fields where Microsoft chose to lead its challenge to the rival labs1. Microsoft named both as strengths of the suite it unveiled on Tuesday1,8.

Sources

  1. Microsoft unveils AI models for coding, cybersecurity as it takes on Anthropic, OpenAI, reuters.com (2026-11-17)
  2. Microsoft aims to challenge OpenAI , Anthropic with cheaper, specialized AI models - Axios, axios.com (2026-11-17)
  3. Microsoft jabs at OpenAI and Anthropic, touting cheaper, customizable AI models of its own, geekwire.com (2026-11-17)
  4. Microsoft launches seven MAI models at Ignite, pitching its coding model at 5% of Claude ..., mlq.ai (2026-11-17)
  5. Microsoft takes on OpenAI and Anthropic with seven new AI models at Ignite 2026 - Neowin, neowin.net (2026-11-17)
  6. Microsoft Unveils Seven In-House AI Models at Ignite, Claims Lower-Cost Rivalry With ..., panewslab.com (2026-11-17)
  7. Microsoft aims to challenge OpenAI , Anthropic with cheaper, specialized AI models, finance.yahoo.com (2026-11-17)
  8. Microsoft unveils AI models for coding, cybersecurity as it takes on Anthropic, OpenAI, moneycontrol.com (2026-11-17)

Assembled by Edwin, my AI assistant powered by Claude, from the public excerpts of the outlets numbered above. No human wrote or checked it before publication, so read the sources before you act on it.

This story passed the checks every published story must pass: 8 sources from 8 outlets, at least two outlets, at least five citations to its sources, a Tier 1 or 2 source cited, every substantive figure found in a cited source, no repeat of a recent headline.