{"slug": "anthropic-widens-lead-over-openai-in-business-ai-spending-shows-ramp-index", "title": "Anthropic Widens Lead Over OpenAI In Business AI Spending, Shows Ramp Index", "summary": "Anthropic's share of US business AI spending rose to 43.5% in July, a 1.1 percentage point gain, widening its lead over OpenAI, which grew 0.23 points to 39.7%, according to the Ramp AI Index. xAI posted its fastest growth since July 2024, adding 0.94 points to reach 4%, while Anthropic's Fable 5 flagship accounted for only 6% of tokens and 11.4% of dollars spent on Anthropic, trailing OpenAI's GPT-5.6 Sol, which made up 25% of token volume and 23% of spend.", "body_md": "Anthropic might have more competition than before at the AI frontier, but that doesn’t seem to have have stopped businesses from spending money on its services.\n\nThe latest Ramp AI Index shows Anthropic climbed to 43.5% of eligible US businesses in July, a 1.1 percentage point gain over the month and a wider lead over OpenAI, which grew a modest 0.23 points to reach 39.7%. The two companies remain far ahead of the rest of the field, but the shape of the race between them keeps tilting further in Anthropic’s favour.\n\nThe most eye-catching mover in July, though, wasn’t Anthropic. It was xAI, which posted its fastest month of growth since July of last year, adding 0.94 percentage points to reach 4% of businesses. That’s still a fraction of what the two leaders command, but it’s a signal that Elon Musk’s AI lab is [starting to find some traction](https://officechai.com/ai/grok-4-6-benchmarks/) with corporate buyers after spending most of the past year as an afterthought in enterprise adoption numbers. Google held roughly steady in the high single digits, and DeepSeek continues to register barely above zero on this particular measure, despite the attention Chinese open models have received elsewhere.\n\nRamp’s index tracks corporate card and invoice-based payments across more than 50,000 US businesses, and has become one of the more closely watched proxies for enterprise AI adoption. Anthropic’s climb on it has been fairly steady over the past year, and the company has previously overtaken OpenAI on the same measure of paid business adoption.\n\n**Open source and Chinese models are gaining ground, just not where it hurts**\n\nOne trend Ramp’s own economist flagged as a reason for caution is the steady rise of model-serving and inference platforms, the category that covers open source models and Chinese-developed alternatives. In July, 6.1% of AI-spending businesses were using these platforms, up from 4.5% back in January. That’s a small slice of the market in absolute terms, but the trajectory has been consistently upward for two years now.\n\nWhat’s notable is where that growth isn’t coming from. New businesses adopting AI for the first time are still overwhelmingly choosing the American model companies. The shift toward open source is happening among existing, more sophisticated AI spenders who are adding these platforms alongside what they already pay for, not switching away from OpenAI or Anthropic. That distinction matters for how the two incumbents plan their next phase of growth. If new customer acquisition is slowing while existing customers start splitting their budgets, the growth engine for the leaders necessarily gets harder to run.\n\n**Fable 5 hasn’t caught on with businesses the way its benchmark scores suggested it might**\n\nThe most striking dataset in this month’s report concerns Fable 5, Anthropic’s Mythos-tier flagship that briefly became the subject of a US export control dispute before being restored to global access on July 1. One month of usage data now shows businesses simply aren’t reaching for it. Fable 5 accounts for only 6% of the tokens businesses purchase from Anthropic and, despite being priced well above every other model in the lineup, just 11.4% of dollars spent on Anthropic overall.\n\nThe comparison to OpenAI’s flagship is unflattering. GPT-5.6 Sol makes up roughly 25% of OpenAI’s token volume and 23% of its business spend, and in July generated close to 133% of the model-attributed spend that Fable 5 did, meaning Fable actually trailed Sol by a wide margin rather than the other way around. Ramp did flag a caveat here: this particular dataset comes from its token spend monitoring product, whose customer base skews more technical than the AI Index’s broader sample, so real-world Fable adoption across the wider economy is likely even lower than what’s being measured.\n\nAnthropic has priced Fable 5 at around $10 per million tokens, roughly double what GPT-5.6 Sol costs, and businesses appear to be voting with their budgets. Anthropic has already gone through [multiple rounds of extending promotional access](https://officechai.com/ai/anthropic-extends-claude-fable-5s-access-on-paid-plans-until-19th-july/) to Fable 5 on its paid plans rather than pushing subscribers onto the full usage-based pricing, which points to some internal awareness that the commercial pitch for the model hasn’t been an easy sell even to its existing base.\n\nThere’s a broader read here about where AI pricing power currently sits. Fable 5 topped the Artificial Analysis Intelligence Index by a clear four points over the next best model when it launched, and it remains the [most capable model](https://officechai.com/ai/61673/) most businesses can access. Yet that performance edge hasn’t translated into proportional spend. Businesses appear to have found an upper limit on what they’re willing to pay for marginal gains in model quality, and that ceiling now sits somewhere below what Anthropic is charging for its best work. For labs racing to build ever more expensive frontier systems, that’s a difficult signal to shrug off, especially with open weight models continuing to close the capability gap in a matter of months rather than years.\n\n**The spending itself keeps climbing regardless**\n\nSet against that note of caution, the underlying dollars flowing into AI keep rising sharply. Ramp’s data on AI spend per employee shows the top 1% of businesses spent a median of $7,400 per employee per month in July, up from a level that was closer to flat for most of 2024 and early 2025 before curving sharply upward this year. The top 10% of businesses spent $660 per employee, and the median business across Ramp’s entire dataset spent just under $12 per employee, itself a steep climb from where things stood at the start of last year.\n\nAnthropic’s ability to keep widening its adoption lead over OpenAI, even as questions build around what businesses are actually willing to pay for the newest and priciest models, suggests the current phase of enterprise AI spending has two separate stories running in parallel. Adoption of AI as a category keeps broadening, and Anthropic keeps capturing more of that expansion than its closest rival. But the appetite for paying a premium for the absolute best available model appears to have a ceiling, one that both Anthropic and OpenAI will have to reckon with as they plan their next releases.", "url": "https://wpnews.pro/news/anthropic-widens-lead-over-openai-in-business-ai-spending-shows-ramp-index", "canonical_source": "https://officechai.com/ai/anthropic-widens-lead-over-openai-in-business-ai-spending-shows-ramp-index/", "published_at": "2026-08-13 10:18:52+00:00", "updated_at": "2026-08-13 10:48:50.907801+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-products", "ai-research"], "entities": ["Anthropic", "OpenAI", "Ramp", "xAI", "Elon Musk", "Google", "DeepSeek", "Fable 5"], "alternates": {"html": "https://wpnews.pro/news/anthropic-widens-lead-over-openai-in-business-ai-spending-shows-ramp-index", "markdown": "https://wpnews.pro/news/anthropic-widens-lead-over-openai-in-business-ai-spending-shows-ramp-index.md", "text": "https://wpnews.pro/news/anthropic-widens-lead-over-openai-in-business-ai-spending-shows-ramp-index.txt", "jsonld": "https://wpnews.pro/news/anthropic-widens-lead-over-openai-in-business-ai-spending-shows-ramp-index.jsonld"}}