Microsoft and Google are taking Anthropic and OpenAI to AI coding models

In the growing productive AI market, Anthropic has moved to the front of the field, mainly thanks to Claude Code, its AI coding assistant. Realizing where the money was, OpenAI shifted its focus from the consumer market to the enterprise, where its Codex offering went up against Claude Code.

Now, Google again Microsoft they’re making a concerted effort to get into the game, using their large balance sheets and scalable cloud businesses to try to attract developers.

At a developer conference last month, Google put a lot of emphasis on agent AI, including the unveiling of Antigravity 2.0, which the company says can “program multiple agents to perform tasks in parallel, such as having one agent code a website while another generates product assets.” And at the beginning of the blog post announcing Gemini 3.5 Flash, Google said that the new model offers “operational boundaries for agents and code.”

Microsoft’s unveiling comes this week at its annual Build conference in San Francisco. When the company intends to announce the type of code that will be available in Copilot, it emphasizes its low price compared to others, said a person familiar with the programs who asked not to be identified in order to discuss internal matters.

The Information reported on these plans last week. Microsoft declined to comment.

As businesses quickly adopt coding assistants, market research firm Mordor Intelligence predicted that the market for AI coding tools will grow 26% annually, from $9.3 billion this year to nearly $30 billion by 2031. The numbers are huge, but Google and Microsoft aren’t just looking for new sources of revenue. They want developers to use their tools, run workloads in their cloud and rely on their models so the models themselves can do more training and become bigger and smarter in the process.

“It’s very important that these companies compete in this market,” said Gil Luria, an analyst covering software and other technology stocks at DA Davidson.

It’s a sentiment shared by Tomasz Tunguz, founder of Theory Ventures, who described AI coding as the most attractive market for generative AI models. He estimates that AI may eventually represent 30% to 60% of research and development costs.

In addition to Anthropic and OpenAI, companies like Cursor are specifically targeting the coding opportunity. In May, Cursor signed an agreement with SpaceX that gives Elon Musk’s company the right to acquire the startup for $60 billion.

The tools became so powerful that Wall Street began bailing out software stocks — although the battered sector rebounded in May — out of concern that entire companies could be wiped out because their products might be coded for vibe. Anthropic on Thursday announced the closing of a financing round valued at $965 billion, surpassing OpenAI, which on Monday said it had privately filed for an IPO.

“AI was moving in a more consumer-focused direction, but Anthropic also focused more on coding because they understood that coding was where the frontier of AI would be,” Luria said. “When everyone was distracted by images and videos, Anthropic knew that coding would improve the performance of models and help with everyone’s jobs, and now they are focusing on coding.”

And last week, Anthropic unveiled an upgrade to Claude Opus, its most advanced model for complex tasks like coding. With Opus 4.8, Anthropic announced an automatic core of 1 million tokens, increasing the amount of working memory it offers for complex coding tasks in a single session.

Google, meanwhile, is resetting the Gemini token quota for its Antigravity product following its developer conference last week, raising model rate limits after developers complained that they used up the initial quotas too quickly.

And Microsoft is starting to charge its Copilot AI coding assistant based on usage to keep pace with rising costs.

“Any time you ask any of these tools, ‘I built this thing,’ they burn tokens,” said Ken Parmelee, an analyst at technology industry researcher Forrester. “This is a new entry drug. It’s a way to connect people to their other products.” A sign is about three-quarters of a word.

At a data analysis software company Snowflakeprogrammers rely heavily on the company’s CoCo development tool, and Claude Code, says CEO Sridhar Ramaswamy. He said it’s not uncommon for an “incredibly productive engineer” to spend $50,000 a year on it.

In addition to Antigravity, Google has Gemini Code Assist that can work within a variety of code editors. Last year, Google signed a $2.4 billion licensing deal for Windsurf technology and hired the AI ​​coding startup’s CEO, Varun Mohan, along with key researchers.

Google CEO Sundar Pichai acknowledged his company’s coding challenge in a recent interview with the “Hard Fork” podcast.

“When it comes to agent code for the use of tools, and the instructions that follow, long-horizon jobs, I think we’re a bit behind at the moment,” Photosi said.

At Google I/O, the company is marketing itself as an affordable coding option, announcing a subscription tier for AI developers at $100 per month.

“They can offer cheap tools, because they know that if you’re in their ecosystem, you’re going to pay to use their memory, their integration,” Parmelee said. He added that the intense competition gives Google an elbow room, calling it “a turbulent market right now.”

“The more users they have on the platform, and the more use cases, the better the models will be, which means the results will be better,” Parmelee said.

Although DeepMind serves as Google’s AI lab, Luria is quick to point out that Google has a big business providing cloud infrastructure and services.

“They are in the business of providing businesses with tools,” said Luria.

Even with Anthropic jumping into a big lead and OpenAI racing to catch up, users of coding tools tend to explore more options, and there’s very little vendor lock-in. Database software maker MongoDBfor example, it released three AI productivity tools, including Anthropic’s Claude Code, to its developers, CEO CJ Desai said.

“In different situations we use different things,” he said. “So if we find that one solves all, I like to combine.”

MongoDB buys AI products one year at a time rather than opting for longer deals, Desai said.

“If Gemini comes up with something better, or the 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 very small fraction of our expensive population.” But if costs rise too much, he said, the company will impose restrictions.

The pressure is on for Microsoft to stay in the game. The software giant has direct access to millions of developers through its GitHub code repository. And with GitHub Copilot, developers can tap models from Anthropic, Google and OpenAI.

Whether Microsoft will simply offer all high-end coding models or whether it will also be the primary model provider is an open question.

On Microsoft’s January earnings call, CFO Amy Hood talked about the company’s need to balance its AI infrastructure needs across Azure, products like GitHub Copilot and research that could lead to product innovation.

“It creates opportunities to understand where it can change if it shows things about Microsoft,” said Michael Turrin, an analyst at Wells Fargo who recommends buying the stock.

Developers in general will gravitate towards the newest and most capable AI models, Turrin said.

“Microsoft is going to have to provide some use cases beyond the cost of that audience,” he said.

When GitHub Copilot launched in 2021, relying exclusively on OpenAI, it was a pioneer in the emerging field of AI coding. But it lost momentum with the rapid emergence of the Curse Code and Claude.

Cursor, an AI-powered code editor, has 300 employees and is one of the fastest-growing cloud software companies in history, going from $4 million to $2 billion in annual revenue in just 18 months. Of the potential SpaceX deal, “People are like, I’m not sure what to make of this, but it’s still a great tool,” Parmelee said.

Rob Sanfilippo, an analyst at the consulting firm Directions on Microsoft, said that Microsoft “may be able to really differentiate by building their models on what they learned about the capabilities of their competitors’ models, and in some cases they lose a lot of their initial share.”

Sanfilippo said this could be part of what CEO Satya Nadella “plans to fix the course of AI at Microsoft.”

WATCH: AI tokens or humans? A new debate is reshaping business budgets