Enterprise coding spending reached $34 billion this year, and two companies now split 70 percent of that market, though one of them has pulled far ahead and keeps widening its lead

Technology
November 20, 2026 · 3:30 PM5 min read

A developer opens a terminal window, types a one line prompt, and watches a wall of working code appear before the coffee finishes brewing.

Multiply that moment by thousands of engineering teams across the country.

Finance departments that once argued over software licenses now argue over which chatbot writes the cleanest code.

Inside that argument, one company has quietly pulled ahead of the rest.

The gap is now big enough that it shows up in corporate budgets, not just coding forums.

How did one company end up capturing most of the spending?

The short answer is habit and trust. Once an engineering team adopts a coding assistant and builds its review process, test suite and internal tooling around that one model, switching costs climb fast.

Enterprise buyers rarely swap a core development tool the way a consumer swaps a phone app. That stickiness compounds.

A model that answers coding questions well six months ago keeps getting chosen again, because the people who approve software budgets would rather extend a contract than retrain a team. So an early lead in raw coding performance hardens into a lead in renewal rates, and renewal rates are what show up in market share reports.

Anthropic built that early lead with successive versions of its Sonnet and Opus models, which have repeatedly ranked at or near the top of coding benchmarks such as SWE-bench Verified since 2024. That kind of sustained benchmark strength is unusual in a field where new models arrive almost monthly.

Picture the budget line that used to not exist

Two years ago, most companies had no separate line item for AI that writes software. Now it is one of the fastest growing expenses inside any technology department, and the jump is not subtle.

According to new figures from Menlo Ventures, coding ($34.0B, up 7.7x) and enterprise AI chatbots ($25.0B, up 3.0x) make up most of the entire $77.3B market for enterprise AI applications this year.

Coding, in other words, is no longer a side feature bolted onto a chatbot; it is the single largest category of enterprise AI spending in the country. Inside that coding category, two companies are capturing almost everything.

Anthropic’s Claude Code and OpenAI’s Codex combine for $24.0B, 70% of the $34.0B coding market, up from about $1B last year. That is a market that went from roughly the size of a single mid sized startup round to tens of billions in a single year, and most of that money is landing on just two products.

What the hardest numbers actually show

These figures come from Menlo Ventures’ latest State of Enterprise AI report, which combines a survey of nearly 400 U.S. enterprise decision-makers overseeing more than $4 billion in AI spending with a bottom-up analysis of the market rather than simply polling opinions about which tool people prefer.

The firm is also an investor in Anthropic, a detail the report itself discloses. The widening shows up outside coding too.

In enterprise LLM API spend, Anthropic rose from 40% to 55%, OpenAI from 27% to 31%, and Google fell from 21% to 9% , according to the same report. So the shift is not confined to developer tools; it runs through the entire enterprise AI spending stack.

A separate measure in the same report shows just how concentrated the whole industry has become. Anthropic, OpenAI, and Google together take 69% of deployment spend ($122.7B, up from 50%) , and Anthropic and OpenAI alone now account for the large majority of enterprise API spend.

Three labs, and increasingly just two of them, are absorbing nearly every dollar that used to flow across a wider field of vendors, a concentration echoed in how tightly agentic coding workflows now cluster around a handful of providers.

Where this advantage runs thin

None of this means the race is settled. Even as its enterprise coding position hardens, OpenAI still commands a far larger consumer audience outside the office, and that consumer reach funds its own aggressive push into coding tools.

Open source alternatives are also chipping away at the edges. Independent coding companies grew 2.9x to $10.1 billion , led by Cursor, Cognition and Factory, and app builders such as Lovable, and free, downloadable coding agents have picked up real developer interest even among teams that keep paying for a leading commercial tool, because engineers like having backup options available even if unused daily.

And the category itself keeps shifting shape. Coding grew up 7.7x in a single year, which means today’s leaders are being measured against a market that barely existed in its current form, not a mature, settled one.

A single strong release from a rival lab could move these numbers again before the next report arrives, the same way security habits around everyday tools, like the password managers hidden inside office laptop browser extensions, can shift once a cheaper or safer alternative appears.

What it means for the rest of the industry

For companies writing the checks, the practical lesson is less about picking a winner and more about watching how fast winners can change. A market that moved from roughly a billion dollars to tens of billions in a single year is not one where any position is permanent.

Even so, the current numbers describe a real, measurable shift, not a marketing claim. Coding has become the largest single category of enterprise AI spending in the country, and one company is capturing most of the growth inside it.

That leaves plenty open for next year’s report, including whether open alternatives erode pricing power, whether a rival model closes the performance gap, and whether today’s combined market share holds steady or splits wider as more vendors chase the same budget line.

Hugo Rojas
Editor
Hugo Rojas is the editor of GCN. With a Master of Science in Engineering, he specializes in technology, data, and science, and brings a human-centered perspective informed by psychology.