Market Intel

Anthropic Takes 73% of New Enterprise AI Spend, OpenAI Drops to 27%

Ramp data shows Anthropic captured 73% of new enterprise AI spending in the past ten weeks, up from a 50-50 split with OpenAI. The shift signals vendor lock-in is happening now, with enterprises committing to Claude for coding and data analysis workflows. For sales teams evaluating AI tools, the window to switch providers is closing fast.

OnTargetIsh Team · · 2 min read
Anthropic Takes 73% of New Enterprise AI Spend, OpenAI Drops to 27%

Anthropic Takes 73% of New Enterprise AI Spend, OpenAI Drops to 27%

Ramp spending data puts a number on what enterprise buyers have been doing quietly: 73% of new AI spend is going to Anthropic. Ten weeks ago, it was 50-50 with OpenAI. That is not a trend. That is a market shift.

The marginal buyer matters more than total spend. OpenAI still has more enterprise dollars overall, but the companies making AI decisions right now are choosing Claude. Once they have spent weeks tuning workflows, training models on their context, and QA-ing outputs, they are not switching back. The soft costs of switching are too high, even when token pricing looks cheaper elsewhere.

Anthropic now holds 32-40% of enterprise LLM market share, up from 12% in 2023. OpenAI dropped from 50% to 25-27% in the same period. The gap is widest in coding (Anthropic at 42-54% share) and data analysis, where Claude 3.5 and 3.7 Sonnet have become the default for enterprises building AI agents and automation workflows.

What This Means for Sales Teams

If your team is evaluating AI tools for prospecting, outreach, or workflow automation, pay attention to what enterprises are locking into. Tools built on Claude (like some AI SDR platforms and sales intelligence products) are seeing stronger enterprise adoption than ChatGPT-based alternatives.

Token costs are noise for most sales applications. If your AI tool is delivering $200k in pipeline value and costing $2k/month in API spend, you are not switching providers to save $500. The switching cost is the re-training, the lost context, the workflow disruption.

Enterprise AI spend hit $37B in 2025, up 180% year-over-year, averaging $7M per company. That budget is consolidating into three providers: Anthropic, OpenAI, and Google (88% of API usage). Vendors are locking in now. If you are evaluating tools, ask which LLM they are built on and whether they support multi-model fallback. The answer tells you whether they are hedging or committed.

OpenAI still owns consumer: 2.5B daily ChatGPT prompts. But the enterprise coding market, the highest-value application for AI spend, is locking in on Claude. That gap will not close by waiting.

Related News

Market Intel

Australia's AI edge is data, not models, says National AI Centre

Lee Hickin, National AI Centre exec director, says Australia should focus on AI applications in data-rich sectors like agriculture and healthcare, not competing on frontier models. Worth noting for sales teams: enterprise AI adoption is already at 68%, but the play is sector-specific tools, not generic platforms.

Mar 25, 2026 · 2 min
Market Intel

Google pauses $20B Australia data centre plan over tax dispute

Google has put a reported $20 billion data centre investment on hold after the Australian government indicated the infrastructure would create a permanent establishment, triggering local tax obligations. The move comes as competitors like AWS commit billions to ANZ cloud infrastructure.

Mar 24, 2026 · 2 min
Market Intel

Tech sector hits $248B, 9% of GDP, but job growth misses target

Tech Council of Australia reports the sector now contributes $248.5 billion to the economy, making it Australia's third-largest industry. Worth noting: nearly half that value comes from indirect tech adoption, not the tech industry itself. Job growth is lagging the council's 2030 targets.

Mar 24, 2026 · 2 min