Microsoft and Google take on Anthropic and OpenAI in AI coding models
Microsoft CEO Satya Nadella speaks onstage during the Microsoft AI Tour in Munich, Germany, on February 25, 2026.
Sven Hoppe | Picture Alliance | Getty Images
In the rapidly evolving generative AI market, Anthropic has stayed ahead of the field, thanks in large part to AI coding assistant Claude Code. Seeing where the money is, OpenAI has shifted much of its focus from the consumer market to the enterprise market, where its Codex offering rivals Claude Code.
Now, Google And Microsoft They’re pushing hard to get into the game, using their massive balance sheets and vast cloud businesses to attract developers.
At last month’s developer conference, Google put a big emphasis on agency AI, including the unveiling of Antigravity 2.0; The company said it can “manage multiple agents to execute tasks in parallel, such as one agent coding a website and another creating brand assets.” And at the beginning of something blog post Announcing Gemini 3.5 Flash, Google said that the new model offers “superior performance for agencies and coding.”
Microsoft’s debut will take place this week at the annual Build conference in San Francisco. The company plans to announce an encoding model that will be available on Copilot, highlighting its lower price compared to alternatives, said a person familiar with the plans who asked not to be named to discuss internal matters.
Information Information about the plans was given last week. Microsoft declined to comment.
As businesses rapidly adopt coding assistants, market research firm Mordor Intelligence He predicted that the market for AI code tools will grow 26% annually, from $9.3 billion this year to nearly $30 billion in 2031. The numbers are big, but Google and Microsoft aren’t just looking for new revenue streams. They want developers to use their tools, run workloads in their cloud, and trust their models so the models themselves can do more training and get bigger and smarter in the process.
“It’s absolutely critical for these companies to be able to compete in this market,” said Gil Luria, an analyst who covers software and other technology stocks at DA Davidson.
This view is shared by Tomasz Tunguz, founder of Theory Ventures, who describes AI coding as the most attractive market for generative AI models. He estimates that AI could eventually represent 30% to 60% of research and development spending.
In addition to Anthropic and OpenAI, companies like Cursor have also specifically targeted the coding opportunity. In May, Cursor signed a deal with SpaceX that gave Elon Musk’s company the right to buy the startup for $60 billion.
The tools became so powerful that Wall Street began bailing out software stocks — although the beaten sector rebounded in May — out of concern that entire companies could be wiped out because their products could be jitter-coded. Anthropic announced Thursday that its funding round had closed at a $965 billion valuation, beating out OpenAI, and said it had confidentially filed for an IPO on Monday.
“AI was moving in a very consumer-focused direction, but Anthropic also became laser-focused on coding because they had the understanding that coding was where the AI frontier would be,” Luria said. “While everyone was distracted by images and videos, Anthropic knew that coding would improve the performance of the models and help everyone with their tasks, and now they are all turning to coding.”
Also last week, Anthropic announced an upgrade to the Claude Opus, its most advanced model for complex tasks like coding. With Opus 4.8, Anthropic announced a new default context of 1 million tokens, increasing the amount of working memory it offers for complex encoding tasks in a single conversation.
Meanwhile, Google reset the token quota for Gemini’s Antigravity product following last week’s developer conference and raised model rate limits after developers followed suit. reported He said they exhausted their initial quota very quickly.
To be more in line with rising costs, Microsoft is starting to charge based on usage for its Copilot AI coding assistant.
“When you say to any of these tools, ‘Build this thing for me,’ they burn coins,” said Ken Parmelee, an analyst at tech industry researcher Forrester. “This is the new gateway drug. It’s a way to connect people to their other products.” One token is equal to approximately three-quarters of a word.
At data analytics software company SnowflakeCEO Sridhar Ramaswamy said programmers mainly rely on the company’s own CoCo development tool and Claude Code. He said it wasn’t unusual for an “extraordinarily productive engineer” to spend $50,000 a year on it.
In addition to Antigravity, Google has Gemini Code Assist, which can run in a variety of code editors. Last year, Google signed a $2.4 billion licensing deal for Windsurf technology and hired Varun Mohan, CEO of the AI coding startup, and key researchers.
Alphabet CEO Sundar Pichai during the Google I/O Developers Conference in Mountain View, California, on May 19, 2026.
David Paul Morris | Bloomberg | Getty Images
Google CEO Sundar Pichai recently acknowledged his company’s coding challenges report With the “Hard Fork” podcast.
“I think we’re a little bit behind right now when it comes to tool-mediated coding and instruction following, long-term tasks,” Pichai said.
At Google I/O, the company marketed itself as a more affordable option for coders, announcing its AI developer subscription tier for $100 per month.
“They can afford to make the tools cheaper because they know that if you are in their ecosystem, you will pay to use their memory and integrations,” Parmelee said. He added that the tense competition gives Google an opportunity to intervene, calling it “a very turbulent market right now.”
“The more users and the more use cases there are on the platform, the better quality the models will be, which means the better the outputs will be,” Parmelee said.
While DeepMind serves as Google’s artificial intelligence laboratory, Luria is quick to point out that Google has a massive business that offers cloud infrastructure and services.
“They are in the business of providing tools to businesses,” Luria said.
Customer choice
Even as Anthropic jumps out to a big lead and OpenAI races to catch up, users of coding tools regularly experiment with multiple options and there is little vendor lock-in. Database software manufacturer MongoDBAnthropic, for example, has introduced three generative AI tools to its engineers, including Claude Code, CEO CJ Desai said.
“We use different things for different use cases,” he said. “So if we find that someone has figured it all out, of course I’d like to reinforce it.”
MongoDB is acquiring AI products one at a time rather than in longer deals, Desai said.
“If Gemini came up with something better, or Codex is better, then I want to be able to use that and not make a long-term commitment, right?” he said.
Desai said AI coding tools currently represent “a tiny fraction of our human cost.” But he said the company could impose limits if the cost increases significantly.
There’s pressure on Microsoft to stay in the game. The software giant has direct access to millions of developers through its coding repository GitHub. Through GitHub Copilot, developers can leverage models from Anthropic, Google and OpenAI.
Microsoft CEO Satya Nadella (right) watches OpenAI CEO Sam Altman speak at the OpenAI DevDay event in San Francisco on November 6, 2023.
Justin Sullivan | Getty Images
It’s an open question whether Microsoft will offer all the best coding models, or whether it will also be a provider of a major model.
On Microsoft’s January earnings call, finance chief Amy Hood talked about the company’s need to balance AI infrastructure needs across Azure, products like GitHub Copilot, and research that could drive product innovation.
“Build presents opportunities for perception to change if they can show something more specific to Microsoft,” said Michael Turrin, an analyst who recommends buying shares in Wells Fargo.
Developers will often gravitate towards the newest and most capable AI models, Turrin said.
“Microsoft will need to offer some specific use cases for this audience beyond cost,” he said.
When GitHub Copilot launched in 2021 based solely on OpenAI, it was a pioneer in the emerging field of AI coding. However, it lost momentum with the rapid emergence of Cursor and Claude Code.
cursorThe AI-powered code editor employs 300 people and is one of the fastest-growing cloud software companies in history, growing from $4 million in annual revenue to $2 billion in just 18 months. “People are saying, ‘I don’t know what to do with this, but it’s still a great tool,'” Parmelee says of the possible SpaceX deal.
Rob Sanfilippo, an analyst at Microsoft consulting firm Directions, said Microsoft “can really differentiate by building its own models with what it learns about the capabilities of the models they compete with and, in some cases, losing a lot of their early share.”
Sanfilippo said this could be part of what CEO Satya Nadella is “planning to correct course on AI at Microsoft.”
WRISTWATCH: AI Tokens or people? New debate reshaping corporate budgets
