# 2025 in AI by the numbers: adoption, spend, protocols and price

- **Published:** 2025-12-23
- **Updated:** 2026-10-12
- **Category:** Enterprise Adoption
- **Author:** Sinthetix
- **URL:** https://sinthetix.com/blog/2025-in-ai-by-the-numbers

> A year-end read of sourced figures: how many companies use AI, how few see profit impact, what NVIDIA reported, how MCP grew and how prices moved.

A year of announcements is easy to remember as a blur of model names. The figures underneath are more useful, and they tell a more specific story: adoption is nearly universal, measurable profit impact is not, infrastructure spending keeps climbing, shared standards arrived, and prices fell. Every number below comes from a source we opened, and we flag where it is vendor-reported. The McKinsey figures are taken from IT Brief's coverage of the survey, because we could not open McKinsey's own page.

**Key takeaways**

- Per IT Brief's coverage of McKinsey, 88% of organizations use AI in at least one function, but only 39% report any enterprise-level profit impact [1].
- NVIDIA's data center revenue reached $51.2 billion in one quarter, up 66% year over year [3].
- MCP, the tool-connection standard, reports over 97 million monthly SDK downloads and more than 10,000 active public servers [4].
- Opus-class list prices fell from $15/$75 to $5/$25 per million tokens between May and November [5][6].

## Adoption is broad

IT Brief's coverage of McKinsey's State of AI survey, published in November, reports 88% of organizations using AI in at least one business function, up from 78% a year earlier [1]. The Stanford AI Index put the figure at 55% for 2023 and 78% for 2024 [2], a similar but separately sourced measure.

**Organizations using AI in at least one function**

McKinsey State of AI

Unit: %

| Category | Share of organizations |
| --- | --- |
| 2024 | 78% |
| 2025 | 88% |

*Survey-reported adoption. The 2024 figure is the prior-year comparison cited in coverage of the 2025 survey. Source: [1], IT Brief*

## Value is narrower

The same coverage shows how steeply the numbers fall as you move from use to results. Nearly two thirds of respondents remain in experimentation or pilot stages. Of those surveyed, 62% are at least experimenting with agents and 23% are scaling them in one or two functions. Only 39% say AI affected enterprise-wide earnings (EBIT), most of them attributing less than 5%, and about 6% qualify as "AI high performers" attributing more than 5% of EBIT to AI [1]. These are as reported by IT Brief, not checked against McKinsey's report.

**From use to measurable results**

Share of surveyed organizations

Unit: %

| Category | Share |
| --- | --- |
| Use AI in a function | 88% |
| Experimenting with agents | 62% |
| Any EBIT impact | 39% |
| Scaling agents | 23% |
| High performers | 6% |

*Figures as reported in coverage of McKinsey's survey; categories are not strictly nested. Source: [1], IT Brief*

## Infrastructure spend

NVIDIA reported third-quarter fiscal 2026 revenue of $57.0 billion, up 62% from a year earlier, with data center revenue of $51.2 billion, up 66% [3]. That is one vendor's quarter, not the market, but it indicates how much capital is still flowing into AI compute. The AI Index adds the cost side: the price of GPT-3.5-level inference fell more than 280-fold between November 2022 and October 2024 [2].

## Standards arrived

On December 9 the Linux Foundation announced the Agentic AI Foundation. Anthropic reports MCP has more than 97 million monthly SDK downloads and over 10,000 active public servers, and that its November 25 specification added asynchronous operations, statelessness and server identity [4]. Interoperability is no longer a promise from one vendor; it is a shared project with multiple large backers.

## Prices moved

| Model | Launch | Input / output per M tokens |
| --- | --- | --- |
| Claude Opus 4 [5] | May 2025 | $15 / $75 |
| Claude Opus 4.5 [6] | Nov 2025 | $5 / $25 |
| Gemini 3 Flash [7] | Dec 2025 | $0.50 / $3 |

Anthropic says Opus 4.5 matched Sonnet 4.5's best SWE-bench Verified score at medium effort with 76% fewer output tokens [6]. Google reports 78% on SWE-bench Verified for Gemini 3 Flash [7]. On the open-weight side, DeepSeek's V3.2 report claims GPT-5-class reasoning and a 73.1 SWE-bench Verified score [8]. All are vendor-reported.

## What the numbers do not show

They do not show which projects paid back. Adoption surveys measure use, not return; the 39% figure, as reported, suggests that most organizations cannot yet point to profit impact [1]. That gap, rather than model capability, is where most of next year's work lies.

For businesses, the implication is to enter 2026 with fewer pilots and clearer measures: pick the workflows where cost per task and quality can be tracked, and judge new models and prices against those baselines.

## Sources

1. [McKinsey report shows AI interest but slow scaling (coverage of McKinsey's State of AI in 2025)](https://itbrief.co.uk/story/mckinsey-report-shows-ai-interest-but-slow-scaling), IT Brief, 2025-11-11
2. [The 2025 AI Index Report](https://hai.stanford.edu/ai-index/2025-ai-index-report), Stanford HAI, 2025-04-07
3. [NVIDIA Announces Financial Results for Third Quarter Fiscal 2026](https://nvidianews.nvidia.com/news/nvidia-announces-financial-results-for-third-quarter-fiscal-2026), NVIDIA, 2025-11-19
4. [Donating the Model Context Protocol and establishing the Agentic AI Foundation](https://www.anthropic.com/news/donating-the-model-context-protocol-and-establishing-of-the-agentic-ai-foundation), Anthropic, 2025-12-09
5. [Introducing Claude 4](https://www.anthropic.com/news/claude-4), Anthropic, 2025-05-22
6. [Introducing Claude Opus 4.5](https://www.anthropic.com/news/claude-opus-4-5), Anthropic, 2025-11-24
7. [Gemini 3 Flash: frontier intelligence built for speed](https://blog.google/products/gemini/gemini-3-flash/), Google, 2025-12-17
8. [DeepSeek-V3.2: Pushing the Frontier of Open Large Language Models](https://arxiv.org/abs/2512.02556), DeepSeek-AI (arXiv:2512.02556), 2025-12-02

## How Sinthetix helps

An engineering studio for AI agents, automation, and internal tooling. We build the systems that make operators scale.

[Our services](https://sinthetix.com/services) | [Talk to us](https://sinthetix.com/contact)

Sinthetix is an AI engineering studio. We ship agents, automation, and internal tools for operators who need results, not reports. [About Sinthetix](https://sinthetix.com/about)