{"id":721,"date":"2026-11-08T07:06:21","date_gmt":"2026-11-07T23:06:21","guid":{"rendered":"https:\/\/aidashxp.com\/ramp-ai-spend-stalling-token-price-war\/"},"modified":"2026-11-08T07:07:08","modified_gmt":"2026-11-07T23:07:08","slug":"ramp-ai-spend-stalling-token-price-war","status":"publish","type":"post","link":"https:\/\/aidashxp.com\/en\/ramp-ai-spend-stalling-token-price-war\/","title":{"rendered":"AI usage rises but spending doesn\u2019t: Enterprise AI expenditure grew only 1.5% from August to September as price wars collapse the growth curve"},"content":{"rendered":"<p class=\"wp-block-paragraph\">Enterprises are using AI more than ever\u2014but spending on AI hasn\u2019t meaningfully increased. This is an enterprise spend management platform. <strong>Ramp<\/strong> That\u2019s the conclusion from the latest report, and the most robust data evidence yet on the AI bubble this year.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Here are the specific numbers:<strong>Token spending rose by just 1.5% combined in August and September<\/strong>In July, this figure rose 42% month-on-month. Meanwhile, usage volume continued climbing. Spending rose while costs stayed flat\u2014the gap between them was absorbed by price wars and smarter user behavior.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Three data points reveal exactly where the money is going<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Ramp segments enterprises by \u201caverage monthly AI spend per employee\u201d\u2014a more informative metric than total spend:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Top 1%<\/strong>At its summer peak, this group spent nearly $10,000 per employee per month, dropping below that threshold in August and September. According to Ara Kharazian, Ramp\u2019s Chief Economist, even this small cohort slowing down<strong>is enough to stall overall revenue growth<\/strong>\u2014even as usage volume continues rising.<\/li>\n<li><strong>Top 10%<\/strong>$600\u2013$1,000 per employee per month, still growing.<\/li>\n<li><strong>Median enterprise<\/strong>$12\u2013$15 per employee per month. This figure is critical\u2014most companies remain two or three orders of magnitude away from being \u201cheavy AI users.\u201d<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">A rebound occurred in October: token spending rose 8.9% month-on-month from September 23 to October 23. Yet Kharazian\u2019s stance is clear\u2014headwinds remain strong, and a single rebound does not signal a trend reversal.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">It\u2019s not that no one is using AI\u2014it\u2019s that prices have collapsed<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">The most commonly misinterpreted aspect of this slowdown<strong>is not demand-related<\/strong>Usage volume is rising, and the number of companies adopting AI is rising\u2014only the price per token is falling.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">In September, OpenAI halved API pricing for GPT-6 Sol and GPT-6 Luna. Anthropic followed suit with steep cuts to Fable 5.1 and Opus 5.5\u2014its recent filing acknowledged this move reduced revenue growth by approximately <strong>$600 million<\/strong>Both firms are also locking in enterprise customers via discounts, effectively trading revenue targets for retention.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Supply-side pressure is even fiercer: Meta and Nvidia are pushing low-cost enterprise models and open-source models, while open-source model vendors in China and the U.S. have driven prices down to just a few cents per million tokens. Several products we previously reviewed are direct outcomes of this price war\u2014<a href=\"https:\/\/aidashxp.com\/en\/tencent-hy4-review\/\">Tencent Hunyuan Hy4 at $0.42 per million tokens<\/a>,<a href=\"https:\/\/aidashxp.com\/en\/particle-1-review\/\">Particle-1 at just $0.25<\/a>Both are fractions of the price for flagship capabilities.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Another new product worth noting is a lightweight tool called <strong>Jev<\/strong> that selects the most suitable output from several preset options; it has been climbing steadily on Ramp\u2019s spending leaderboard. Kharazian\u2019s explanation is straightforward: \u201cIn many AI use cases, you only need a yes\/no decision. So if you run an already inexpensive model on Jev, cheap becomes even cheaper.\u201d<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Enterprises are cutting costs themselves\u2014and doing so with increasing sophistication<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">The other reason is user maturity. Kharazian puts it plainly: \u201cWhen you become highly proficient with AI, you learn how to control it better.\u201d This manifests in three concrete actions\u2014now widely adopted practices:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Default disabling expensive models for long-running tasks<\/strong>Downgrading backend batch processing and scheduled jobs\u2014tasks that don\u2019t require top-tier intelligence<\/li>\n<li><strong>Setting budget caps for employees<\/strong>Capping allocations per person or per project, rather than allowing unlimited usage<\/li>\n<li><strong>Routing upstream<\/strong>Sending simple requests to lightweight models, reserving flagship models only for complex ones<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">Notion co-founder Akshay Kothari\u2019s internal data aligns completely with Ramp\u2019s observations; his summary is one sentence: \u201cTokens surge, costs plunge.\u201d<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">What this means for AI companies: The Uber playbook no longer works<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Big tech\u2019s classic user-subsidy strategy\u2014low prices first to capture market share, then price hikes after competition clears out\u2014worked for Uber and Lyft. But Kharazian argues this script fails in AI, for one reason only:<strong>Competition is too intense<\/strong>If you cut prices, Meta counters with open-source models; if you cut further, Chinese open-source models undercut you.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This is an unwelcome signal for Anthropic, which is preparing for IPO. If investors still value AI companies on the assumption of \u201csustained hypergrowth,\u201d Ramp\u2019s data undermines that premise. More troubling is structural risk: Kharazian points out that nearly all AI spending comes from venture-backed, high-growth tech companies\u2014meaning<strong>once market conditions shift, this entire cohort of customers will contract simultaneously<\/strong>creating a chain reaction.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">There\u2019s another ceiling: traditional industries like doctor\u2019s offices and car dealerships genuinely need AI, but they\u2019re unlikely to reach $10,000 per person per month in spending. \u201cThe era of hypergrowth may be over,\u201d says Kharazian. \u201cNow we know the revenue we can extract from the top 1% may be capped.\u201d<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">What this means for individual users and teams<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">The conclusion has flipped\u2014and it favors buyers:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Now is the window for negotiation.<\/strong>When renewing enterprise AI service contracts, buyers can directly leverage competitor quotes; vendors\u2019 discount headroom is significantly larger than it was a year ago.<\/li>\n<li><strong>Stop defaulting to flagship models.<\/strong>Tier tasks by difficulty and route them accordingly\u2014costs for the same work can drop by over 50%. See reference: <a href=\"https:\/\/aidashxp.com\/en\/glm-5-3-flashx-review\/\">GLM 5.3 FlashX<\/a> Rely on these high-speed, low-cost models as your primary stack; reserve flagships only for requests that truly demand advanced reasoning.<\/li>\n<li><strong>Open-source solutions are regaining competitiveness.<\/strong><a href=\"https:\/\/aidashxp.com\/en\/qwen38-2-4t-a95b-review\/\">Qwen3.8 2.4T<\/a>,<a href=\"https:\/\/aidashxp.com\/en\/ernie-5-5-flash-base-review\/\">Wenxin 5.5 Flash Base<\/a> The cost structure of such open-source weights remains entirely unaffected by API price hikes.<\/li>\n<li><strong>Treat \u201cAI spend per person\u201d as a management metric.<\/strong>Ramp\u2019s data essentially gives you a ruler: Is your team spending at the median ($12\u2013$15 per person per month) or in the top 10% ($600\u2013$1,000)? Overspending doesn\u2019t necessarily signal efficiency\u2014it may simply reflect a lack of routing.<\/li>\n<\/ul>\n\n\n\n<h2 class=\"wp-block-heading\">Frequently Asked Questions (FAQ)<\/h2>\n\n\n\n<h3 class=\"wp-block-heading\">Has enterprise AI spending truly plateaued?<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">Slowing spend growth does not mean usage has stalled. Ramp\u2019s data shows token spend rose only 1.5% between August and September, then rebounded 8.9% in October\u2014while call volume continued rising steadily. The underlying driver is a sharp decline in unit pricing, decoupling the \u201cusage growth\u201d and \u201cspend growth\u201d curves.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Why are AI prices falling so rapidly?<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">A three-way squeeze: OpenAI and Anthropic are slashing prices (API price adjustments at roughly 50% levels since September), open-source models have driven marginal costs down to mere cents, and enterprises deploying routing and budget controls are actively filtering out expensive requests. With oversupply and increasingly sophisticated buyers, prices can only go down.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">Does this affect valuations of AI companies?<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">Yes. Anthropic itself disclosed that its price cuts reduced revenue growth by approximately $600 million, confirming that the price war hits revenue directly. If IPO pricing still assumes hypergrowth, sustained flattening of spend growth poses significant valuation correction risk.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\">How should our team reduce AI costs?<\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">Three steps: tier tasks by difficulty, route simple requests to cheaper models, and set per-user budget caps. Most teams cut over half their costs with just the first two steps\u2014without any noticeable quality degradation\u2014because expensive models spend most of their time handling tasks that don\u2019t warrant their cost.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Summarize<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">\u201cToken prices rise, but costs stay flat\u201d\u2014this signals a disrupted growth story for model providers, yet delivers the best possible news for users. AI is shifting from a token-priced luxury to an on-demand utility like electricity or water\u2014only when prices fall can true large-scale adoption begin.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Want to pick a more cost-effective model combination? View our <a href=\"https:\/\/aidashxp.com\/en\/ai-models\/\">AI Model Library<\/a> and <a href=\"https:\/\/aidashxp.com\/en\/compare-tools\/\">Tool Comparison Engine<\/a>or continue reading:<a href=\"https:\/\/aidashxp.com\/en\/glm-5-3-flashx-review\/\">GLM 5.3 FlashX benchmarking<\/a> \u00b7 <a href=\"https:\/\/aidashxp.com\/en\/particle-1-review\/\">Document your initial hypothesis and metric selection with a timestamp. Experiments show this does not fully prevent influence, but at least lets you see afterward what changed in your thinking<\/a> \u00b7 <a href=\"https:\/\/aidashxp.com\/en\/deepseek-15b-funding\/\">Lock in your own judgment before reviewing the Agent\u2019s output<\/a>.<\/p>","protected":false},"excerpt":{"rendered":"<p>\u4f01\u4e1a\u7528 AI \u7528\u5f97\u6bd4\u4ee5\u5f80\u4efb\u4f55\u65f6\u5019\u90fd\u591a\uff0c\u4f46 [&hellip;]<\/p>\n","protected":false},"author":0,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"site-sidebar-layout":"default","site-content-layout":"","ast-site-content-layout":"default","site-content-style":"default","site-sidebar-style":"default","ast-global-header-display":"","ast-banner-title-visibility":"","ast-main-header-display":"","ast-hfb-above-header-display":"","ast-hfb-below-header-display":"","ast-hfb-mobile-header-display":"","site-post-title":"","ast-breadcrumbs-content":"","ast-featured-img":"","footer-sml-layout":"","ast-disable-related-posts":"","theme-transparent-header-meta":"","adv-header-id-meta":"","stick-header-meta":"","header-above-stick-meta":"","header-main-stick-meta":"","header-below-stick-meta":"","astra-migrate-meta-layouts":"default","ast-page-background-enabled":"default","ast-page-background-meta":{"desktop":{"background-color":"var(--ast-global-color-5)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"ast-content-background-meta":{"desktop":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"tablet":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""},"mobile":{"background-color":"var(--ast-global-color-4)","background-image":"","background-repeat":"repeat","background-position":"center center","background-size":"auto","background-attachment":"scroll","background-type":"","background-media":"","overlay-type":"","overlay-color":"","overlay-opacity":"","overlay-gradient":""}},"footnotes":""},"categories":[46],"tags":[],"class_list":["post-721","post","type-post","status-publish","format-standard","hentry","category-ai-news"],"_links":{"self":[{"href":"https:\/\/aidashxp.com\/en\/wp-json\/wp\/v2\/posts\/721","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/aidashxp.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/aidashxp.com\/en\/wp-json\/wp\/v2\/types\/post"}],"replies":[{"embeddable":true,"href":"https:\/\/aidashxp.com\/en\/wp-json\/wp\/v2\/comments?post=721"}],"version-history":[{"count":1,"href":"https:\/\/aidashxp.com\/en\/wp-json\/wp\/v2\/posts\/721\/revisions"}],"predecessor-version":[{"id":724,"href":"https:\/\/aidashxp.com\/en\/wp-json\/wp\/v2\/posts\/721\/revisions\/724"}],"wp:attachment":[{"href":"https:\/\/aidashxp.com\/en\/wp-json\/wp\/v2\/media?parent=721"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/aidashxp.com\/en\/wp-json\/wp\/v2\/categories?post=721"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/aidashxp.com\/en\/wp-json\/wp\/v2\/tags?post=721"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}