Ramp transaction data shows Anthropic overtook OpenAI in its share of paying US business customers by late 2025
Story Overview
Ramp's anonymized transaction records from tens of thousands of US companies reveal Anthropic edging past OpenAI in the share of businesses making actual paid AI purchases, with the crossover accelerating through late 2025 and confirmed in April 2026 data where Anthropic reached 34.4 percent versus OpenAI's 32.3 percent.
Coding workflows justified the spend
Businesses appear to have favored Anthropic once tools delivered measurable productivity lifts in software development, turning experimental usage into recurring paid subscriptions that Ramp's card data could track.
Mid-market sample leaves enterprise gaps
Ramp's platform skews toward growth-stage firms, so the reversal may undercount large enterprises that still route OpenAI payments outside corporate cards or rely on free tiers the dataset misses entirely.
Many users praised Anthropic's disciplined focus on coding and enterprise tools as the key reason for its rapid overtake of OpenAI in US corporate AI adoption.
No Digg Deeper questions have been answered for this story yet.
Most Activity
Holiday vacation winter 2025 was a huge inflection point for Claude code & Anthropic
The perfect timing window of having the best model + harness at the time, when corporate decision makers had time to play on side projects, won Claude trillions $$$ in future enterprise tokens sales.
I expect model labs to push huge releases in December from now on, the same way retailers & video game launches try to get things out around that time for the Christmas cycle.
Anthropic's business case needs to be studied. At the end of 2025 and the beginning of 2026, there was an incredible increase in usage in the business/enterprise sector, which made them number 1.
@kimmonismus A very interesting topic and worth exploring further.
@kimmonismus The enterprise focus from the start and PR stunts made them famous, which led to a snowball effect given how good Opus 4.5 and Opus 4.6 were in comparison to the competition.
@kimmonismus Have a better model.
Done, there's your study for you.
@kimmonismus This is the more durable win.
Consumer AI usage swings with hype cycles. Enterprise deployment decisions take months and don’t get reversed easily once integrated into workflows.
@kimmonismus Enterprise usage ... bundles of thousands of equivalent individual subscribers in one transaction...
@kimmonismus Claude Code and focus on coding were key. Win software - win big tech - win every startup - win Fortune 500
@aatxcus Enterprise is key to win the AI race. That’s why OpenAI is trying to get there with codex
@kimmonismus Huge number of devs tried Claude Code for the first time over the holidays on personal projects, realized agents work now, and came back to work agent-pilled on Claude Code.
@kimmonismus i believe this is going to continue too
chatgpt by many is seen as “the boomer AI” while claude is cool, sounds like you, and has tools actually used in business every day
openai will catch up no doubt, but until then, claude will run rampage
Anthropic were very disciplined. Didn't waste time with image generation or video generation or any gimmicks. Went straight for coding and focused entirely on that and it paid off. Opus 4.5 really was a phase transition for them though you can see the effect of people trying Claude Code after it was released and going "Oh wow this is actually good now"
Compare to OpenAIs massive subscriber consumer base and then take the corps that will overlap and not allow personal AI and then the marketing is diff cause you are targeting CIO and COO while exciting the CFO and CEO ...
But at the end of the day it comes down to [stability, cost] and [accuracy, security] usually in that order with cost over stability if necessary...
Fact checking my own post @grok pointed out that I had my corporate prioritization wrong but then threw me a bone and acquiesced that IRL corps may prioritize cost above all else ... hmmm go figure ... nahhh ... lol
"Enterprises do make cost trade offs."
So my generalization holds reasonably well ...
Security, privacy, compliance, and trust are currently the main bottlenecks but you watch cost and ROI are the true driving force ... as soon as it can be implemented without tanking the business ...
@NickADobos I buy the timing, but the harness was the unlock: real repo, failing test, clean diff, repeat. The side project was just where the habit formed.
@CharuruCha14310 That’s probably a key element, true. But Anthropic did it smarter in many ways
@kimmonismus Anthropic on motion right now, huh
@kimmonismus Which is why they are now fumbling it
@kimmonismus actually they've played all their cards right until the Fable fumble. Hopefully they get things in order 🤞
The were the only ones supporting in a chat app (claude app) MCP the agent layer that enabled business to communicate with any API, DB or service, by november 2025 reached a critical point. meanwhile all the others Chatgpt, Xai, Gemeni were in a sidequest of making videos, memes, art and maximizing AI websearch, nonoe of those apply to enterprise.
@kimmonismus How come enterprise adoption seems to be the complete opposite of community consensus across most platforms (Reddit, x, forums)? I try not to stay in a bubble, but right as Anthropic’s line blows up is when I and most people I know across sectors dumped Claude for Codex.