The business winners and osers of the global AI race in 2025
In case you're still in doubt, real money has been made this year with AI. I was breaking down for one of our industrial customers where the big moves were made, and I was a tad surprised.
Here's the breakdown of the big AI money makers in 2025 (projected):
| Company | Estimated AI Revenue (2025) | Type | AI Business Focus | |
|---|---|---|---|---|
| 1 | NVIDIA | ~$115 B | Tech (Hardware) | AI GPUs, data center systems, software frameworks (CUDA, TensorRT) |
| 2 | Amazon (AWS) | Multi‑$B, est. ~$15–20 B+ | Tech (Cloud) | Cloud AI infrastructure (SageMaker, Bedrock, AI chips like Trainium) |
| 3 | Microsoft | ~$13 B (run-rate) | Tech (Cloud + Software) | Azure OpenAI Service, GitHub Copilot, M365 Copilot |
| 4 | OpenAI | ~$12.7 B | Tech (AI Software) | GPT APIs, ChatGPT Plus/Enterprise, model licensing |
| 5 | Google (Alphabet) | Est. ~$10–12 B+ | Tech (Cloud + Software) | Vertex AI, Gemini models, AI-enhanced ads/search/youtube |
| 6 | Accenture | ~$3.6 B AI consulting bookings | Consulting (IT & Strategy) | Gen-AI implementation, enterprise integration, AI transformation |
| 7 | Anthropic | ~$3–4 B run-rate | Tech (AI Software) | Claude model APIs, Claude Pro, enterprise deployments |
| 8 | AMD | ~$5 B (AI GPU sales) | Tech (Hardware) | Instinct MI300 AI accelerators, server CPUs |
| 9 | BCG | ~$2.7 B (20% of $13.5B revenue) | Consulting (Strategy) | AI strategy and implementation (via BCG X and Gamma) |
| 10 | McKinsey | Est. ~$2–3 B, projected 40% mix | Consulting (Strategy) | AI/ML transformation, operations, proprietary AI tools |
| 11 | IBM | ~$5 B bookings-to-date | Tech + Consulting Hybrid | Watsonx AI software, consulting, model ops |
| 12 | KPMG | Est. ~$1–2 B blended AI | Consulting (Big Four) | AI governance, audit, tax AI integration |
| 13 | PwC | Est. ~$1–2 B blended AI | Consulting (Big Four) | AI assurance, chatbot risk audits, implementation |
| 14 | EY | Est. ~$1–1.5 B AI-related | Consulting (Big Four) | EY.ai platform, foundation models, internal & client-facing bots |
| 15 | Salesforce | ~$1 B ARR (AI + Data Cloud) | Tech (Enterprise Software) | Einstein GPT, AI Cloud, Agentforce |
| 16 | Adobe | ~$250M ARR from Firefly AI | Tech (Creative Software) | Generative image/text (Firefly), Acrobat AI assistant |
| 17 | Bain | < $1 B, no public number | Consulting (Strategy) | AI strategy + OpenAI partnership for enterprise GPTs |
No surprise NVIDIA is largely dominating as a company that integrates the full stack from day 1. Every LLM, every multimodal model, every proprietary AI productivity layer runs through their silicon. This isn’t innovation theatre—it’s plumbing. Foundational, essential, and scaling faster than the market can absorb (which, ironically enough, put their stock in a bind as its valuation is already stratospheric and doesn't seem to be able to go much further). They are not just selling chips, they are the physical infrastructure of the AI economy.
Then you have the Cloud Triumvirate: AWS, Microsoft, and Google. And one can argue they’re not in the business of inventing AI but rather in the business of renting it. Their edge is distribution: cloud platforms, enterprise footprints, developer ecosystems. Their monetization is stealthy, recurring, and deeply embedded in customer operations. You're not buying fancy new tools, you're already subscribed to Microsoft Office. AI becomes another SKU, bundled in consumption models they already dominate.
Logically, OpenAI and Anthropic who are admittedly the poster children of the AI gold rush are just running behind. High-margin, high-growth, yes. But dangerously dependent on a few partners. But for now, they’re printing revenue without even trying to sell as much as they can since they need to win the adoption race with unclear value added when compared to each other. Their positioning is still weird as they are always facing the risk of becoming middleware in someone else’s stack.
And then the real surprise appears: Accenture ranks #6 in the AI money-making game.
But is it a surprise?
Without infrastructures to build or models to invent and run, Accenture has just pivoted without friction to the new tech flavor of the month and sells calming pills to quench incumbents paranoia and fears. While the others were publishing AI reports, Accenture booked $3.6B in real Gen-AI projects. Not projections. Not thought leadership. Actual client work. The firm has become the go-to systems integrator for boardrooms that want AI yesterday. Think SAP in the nineties but without even having to build a software stack.
BCG and McKinsey are catching up fast, layering AI strategy on top of transformation mandates. They’re building capability fast, but still selling PowerPoint more than platforms. Bain, for now, remains oddly quiet. One suspects they’re still drafting their Gen-AI manifesto.
And the Big Four? PwC, EY, Deloitte, KPMG—they’ve invested billions, but it’s all buried in the plumbing. AI shows up in audit tools, tax workflows, and digital twins of compliance. It’s there, but invisible. Which is both a strength (integration) and a weakness (no clear monetization signal). Their AI isn't a product; it's a feature of everything else. Yet, they're not Microsoft.
Lastly, Salesforce and Adobe remind us AI can also be feature-ized. Generative add-ons to SaaS. Firefly images, Einstein GPT prompts, PDF chatbots. Small, smart monetization. But at ~$1B or less, it’s still pocket change in the broader platform wars.
What it the key takeaway for the foreseeable future?
The AI business race is being won by those with the smartest and most aggressive business model. Anthropic or OpenAI, although in the leading pack for now, are exploiting the market in a very weak way, looking more like consumers platforms trying to get some monthly subscriptions rather than proper tech giants organizing the markets around them. And in this game the giant tech incumbents have been very good at repositioning and pivoting around AI; just like some of the big consulting firms.
Whoever has missed the race to AI profit in 2025 is not only signaling they're late, they're signaling they haven't managed to pull out a clear business model to address the market.
So yeah... Apple?
