Zoho's AI Spending: The Test a Bootstrapped Company Can't Avoid

Zoho just announced finalists for a hackathon where college students build AI tools to fight wildfire risk and water scarcity in rural Texas and California. The cash prize is $15,000. It sounds like corporate goodwill.

But Zoho is a company that has never raised external venture capital, never gone public, and never treated its customers as a growth metric to be optimized at any cost. The hackathon is not a distraction from the real story. It is a small signal about how the company plans to compete in AI without being crushed by it.

The real story is in the financials.

Zoho crossed ₹12,313 crore in consolidated revenue for fiscal year 2025 — about $1.47 billion, up roughly 17.8% year-over-year. Net profit was ₹3,191 crore, down slightly from ₹3,299 crore the year before. Revenue grew. Profit didn't keep pace.

The gap is telling. Total expenses surged 30.5% to ₹9,216 crore, well ahead of revenue growth. Depreciation and amortization — a proxy for infrastructure spending — jumped from ₹514 crore to ₹741 crore. That is a 44% increase. Employee costs rose 29% year-over-year. Cash reserves nearly tripled to ₹1,878 crore, suggesting the company is funding this expansion from accumulated wealth rather than debt.

In a bootstrapped company that achieved a 25.9% net margin and a Rule of 40 score of roughly 43.7 — far above publicly traded peers like Freshworks and HubSpot — that kind of expense acceleration is unusual. It means something is changing under the hood.

That something is AI.

Every large public software company is spending tens or hundreds of billions on data centers, GPUs, and the infrastructure required to run large language models. The five largest U.S. cloud providers have collectively committed to spending $660 to $690 billion on capital expenditure in 2026. Zoho does not have that scale. It has a $1.5 billion revenue base, about 19,000 employees, and $12.5 billion in estimated private value. It cannot outspend Microsoft or Salesforce.

So what does a company like Zoho do when the entire software industry pivots to AI?

The hackathon gives you a hint. Students were asked to design lightweight, interoperable AI tools for specific rural problems — wildfire risk, water scarcity, agricultural instability. Not foundation models. Not infrastructure. Not a platform that needs to serve every user simultaneously. The deliverable was narrow, contextual, and built for a community that already has urgent, localized needs.

Most people think competing in AI means training bigger models and buying more compute. But look at Zoho.

Zoho already has 55 business applications across CRM, finance, HR, and collaboration, serving 150 million users. Its AI product, called Zia, has existed before the current wave. The company's strategy, at least as the financials suggest, is not to build its own GPT but to weave AI into applications where it solves narrow, specific problems. Let someone else carry the cost of foundation models. Build the layer that uses them well.

This is the same discipline that made Zoho profitable over three decades. The company started in 1996 in Chennai. Sridhar Vembu, the founder, went to Princeton, worked at Sun Microsystems, and returned to India. He moved Zoho's operations into villages in Tamil Nadu, hiring local talent at a time when Indian tech was flowing the other way — from villages to cities, from India to Silicon Valley. The company charged for software early, kept costs low, and never surrendered pricing power to the narrative that growth requires capital raises.

The same pattern is playing out with AI. A bootstrapped company cannot compete on infrastructure. It can compete on application depth, customer context, and pricing that makes sense for the small and midsize businesses that form most of its base. Zoho CRM starts at $14 per user per month. Salesforce starts at $25. The gap widens at higher tiers.

But here is the tension.

The expense surge shows this is not a free lunch. Building AI capabilities — even lightweight ones — requires real investment. The 44% jump in depreciation, the 29% rise in employee costs, the compression in net margins from roughly 32% to 26%: these are not rounding errors. They show a company spending more aggressively than it has in years. For a bootstrapped firm, spending discipline is not just a financial preference. It is the core identity. If you cannot spend without discipline, you no longer have an advantage over the companies that just raise more money.

The question then becomes whether the AI spending compounds or merely accumulates.

If Zoho's AI investments simply keep the company competitive — preventing churn, matching features that Salesforce and Microsoft add — then the spending is a cost of maintaining the moat, not widening it. Revenue keeps growing, margins stay under pressure, and the business becomes more ordinary.

If instead the AI investments unlock something that the big players cannot replicate — deeper integration across 55 applications, pricing that remains accessible, contextual intelligence that comes from understanding a small business owner's actual workflow — then the spending is an investment in widening the moat. And the margin compression is the price of a larger future.

There is no way to know yet. The FY25 numbers show the spending has begun. They do not show what it produces.

What you can test, though, is simple.

Watch what happens to Zoho's net margin over the next two quarters. If it recovers toward the 30% range while revenue keeps growing at 15% or above, the AI investments are probably paying off — the company spent its way to something that compounds. If the margin stays compressed at 25% or below while growth slows, the spending is more likely a defensive cost than an offensive advantage.

The hackathon in Texas and California costs the company very little. The AI spending behind it costs billions of rupees a year. The difference between the two is the difference between a gesture and a strategy. One tells you what Zoho says it believes. The other tells you whether the belief can survive the math.