Rising AI’s Costs Put SaaS Companies Between A Rock And A Hard Place

Sridhar Vembu, Zoho co-founder, has warned that the economics of building software in an AI-driven market are posing challenges for SaaS companies. He noted that server prices have risen significantly, and AI usage is driving up operating costs through higher token expenditures. In response, Zoho plans to cut its marketing budget, freeze headcount for 18 months, and pause salary hikes this year.

AI is making software development easier, but running a software business is becoming more expensive. While AI coding tools and agents are challenging traditional subscription and per-seat pricing models, AI-driven productivity gains in software development are currently ranging from 25% to 50%. However, testing, verification, and other stages of product development still require human involvement.

AI costs impact SaaS revenue, and every AI query incurs computing power usage, increasing operating expenses. As AI infrastructure demand grows, hardware prices rise, making it costlier for companies to run their software. According to ICONIQ’s 2026 data, AI products have an average gross margin of around 52%, compared to the 80% benchmark for traditional SaaS.

For established SaaS companies integrating AI features, gross margins can drop by 12-17 percentage points depending on their efficiency in managing AI computing costs.

Some SaaS companies are accepting higher costs, expecting AI features to attract more users, enhance their products, and create future revenue opportunities. Others are adding capabilities without proportionately increasing subscription prices. For instance, Fireflies, a meeting note-taking product, has maintained its subscription price for five to six years while expanding into other tools like email summarisation and meeting preparation.

The pressure to manage AI expenditure is evident in Indian companies' financial results. Groww, for example, reported a 28% year-on-year rise in total expenses to ₹556 Cr in Q2 FY27, driven by inflation in compute costs due to AI adoption. The company expects these inflation-driven costs to become part of its operational baseline going forward.

Companies are adopting strategies like fine-tuning smaller language models for specific tasks, optimising infrastructure, and using expensive models only when necessary. Perfios, a financial technology firm, is fine-tuning smaller language models and optimising GPU architecture to make specific workloads cheaper to run. Freshworks acquired Torii, a software management platform, to enhance its Freshservice product with AI spend governance, tracking software applications, licences, and AI consumption to attribute spending to the people and teams responsible for it.

Perfios, meanwhile, tracks profitability at both the product and customer levels, focusing on optimising costs and directing spending towards areas that deliver better business outcomes.