LLM Companies vs SaaS Giants: Who’s Winning in AI Monetisation?

OpenAI’s revenue reportedly surged to $3.7 billion last year. Its compute bill, however, was estimated at $5 billion. Meanwhile, Microsoft’s AI revenue run rate climbed past $13 billion in early 2025, quietly built into its Office and Azure stack.

A new report by Bondcap lays out the monetisation story in generative AI, venture-fuelled LLM players chasing speed and scale, versus SaaS leaders adding AI into existing workflows.

With over $95 billion in capital raised by leading LLM companies and SaaS giants rapidly repackaging AI across their portfolio, the question of who is winning remains open.

LLM Companies are Catching Up Fast, But Profits, Not So Much

The leading AI-first companies are generating impressive top-line numbers while burning cash at unprecedented rates. OpenAI, Anthropic, xAI, and Perplexity collectively generate over $11 billion in annualised revenue. Yet, despite this growth, all remain deep in the red.

The growth trajectories are remarkable. Perplexity hit $120 million annual recurring revenue (ARR) in under two years. Anthropic crossed $2 billion after growing 20 times in 18 months. Moreover, valuations tell a different story—OpenAI was valued at 33 times of its revenue, Anthropic at 31 times, and Perplexity at a staggering 75 times.

These companies make money through monthly subscriptions ranging from free to $250 and pay-per-use pricing that can reach $40 per million tokens. Business customers have been eager adopters. ChatGPT Enterprise reached 2 million business users by March 2025 and is now used by 80% of Fortune 500 companies. In China, DeepSeek’s mobile app attracted 54 million monthly active users just four months after launching.

As cited in the report, data by payment processing giant Stripe shows that the top 100 AI companies reached $5 million in yearly revenue faster than older software companies, taking just 24 months compared to the usual timeline. Despite this growth, making money remains tough, as the cost of running AI systems keeps growing faster than revenue.

SaaS has the Userbase and a Cautious Approach

The SaaS giants haven’t sat still either. Microsoft’s AI revenue reached a $13 billion annual run rate by January 2025, an increase of 175% year-on-year. This includes GitHub Copilot, which is used by 77,000 organisations, and Microsoft 365 Copilot, which saw daily usage double and engagement intensity rise over 60% quarter-on-quarter.

Salesforce's Agentforce added $900 million in ARR from AI and data cloud products within 90 days of launch in late 2024. Canva’s Magic Studio tools were used over 16 billion times by May 2025. Adobe Firefly topped 20 billion asset generations. Atlassian Intelligence reached one million monthly users, while Zoom’s AI Companion was active in 3.7 million accounts by the end of 2024, marking a 68% quarter-on-quarter jump.

However, the monetisation paths here are less headline-grabbing. These companies aren’t introducing separate AI products, they’re integrating AI into their existing services. For most users, AI is just a feature toggle within familiar interfaces. 

This approach is also financially cautious. Microsoft, Amazon, Alphabet, and Meta all saw CapEx grow between 50% and 63% last year, driven largely by AI infrastructure investments. Yet, their models remain closer to profitability than their LLM counterparts. These firms are making room for AI, not replacing their human workforce with it.

So, Who is Winning?

What makes the current comparison tricky is that LLM firms and SaaS giants aren’t playing the same game. The former are trying to become platforms, horizontal, foundational services like the internet or mobile operating system, while the latter are using AI to make existing platforms more useful.

While LLM players pursue rapid user growth and API integrations, SaaS incumbents are focused on expanding customer lifetime value through AI enhancements. One conclusion could be that one model chases new revenue while the other strengthens old pipelines.

Valuations reflect this divide. LLM startups command multiples that stretch belief. Investors are buying the promise of platform dominance. SaaS, in contrast, is growing more slowly but operating within better-understood boundaries.

LLM companies are outperforming SaaS in the speed of growth and user acquisition. However, the economic engine beneath that speed is still under construction. AI usage, for all its scale, hasn’t yet proven its ability to drive sustainable margin at the same rate.

So are LLM companies outpacing SaaS in AI monetisation? In terms of speed to revenue milestones, yes. But this is not an overall outlook of the answer.

In the long run, both may win. However, the LLM track demands deeper capital, higher risk, and patience that stretches beyond quarterly earnings. The speed of progress, while impressive, is insignificant without financial stability. The critical measure of success will be profit margins, not just rapid development.