At its peak, SaaS (software as a service) valuations (average EV/revenue multiple) had hit 30x in 2021 but post the hike in interest rates, concerns over weak growth and the advent of generative AI, the multiples have been oscillating around 5-7x. Harshjit Sethi, Managing Director at Peak XV Partners (formerly Sequoia Capital India & SEA), a leading venture capital and growth investing firm, shares his perspective on the possible trends that will shape the estimated $300 billion industry, valued at over $3 trillion. Sethi highlights the growing importance of profitability in a high-cost capital environment and explains how SaaS companies must adapt — not only by leveraging AI but by rethinking business models to stay competitive.

Post-Covid, we saw SaaS company valuations soar as everything moved to the cloud. Now, with the emergence of Generative AI, while there’s potential for SaaS to accomplish far more than before, the market seems to be signalling that AI might disrupt SaaS instead of enhancing its value. What’s your perspective on this?

During the Covid period and the subsequent shift to digital, SaaS multiples skyrocketed because these companies, with gross margins around 80%, were expected to generate substantial cash flow over time. If we look at the largest SaaS companies, they do generate a reasonable amount of free cash flow, around 20–30% of their revenue. But as interest rates rose, the higher discounting of cash flows naturally resulted in lower multiples, and that’s largely what we’ve seen play out.

Now, when you factor in AI, as you mentioned, there’s an additional layer of complexity. SaaS has traditionally built workflow tools that allow us to work more efficiently by standardising processes. For instance, an HR team onboarding a new employee goes through specific steps that a SaaS product organises. But with AI, rather than just managing workflows, we can perform the work itself. Similarly, instead of creating a tool to assist a customer support agent, AI might handle the customer query directly.

This fundamentally changes the value proposition, as companies may need fewer employees if AI can handle tasks, potentially disrupting the per-seat pricing model of SaaS. So, in my view, some SaaS companies will be more vulnerable to disruption, while others might continue to adapt and evolve.

The speed at which these companies evolve will be a key factor, as most SaaS leaders recognise that Generative AI is the next big wave. If you ask any SaaS company’s management, they’ll likely confirm they’re working on integrating AI, though their effectiveness will vary. Some SaaS products and industries are easier to automate, while others that are more complex and harder to replicate with AI might be the ones that may survive.

So, certainly, the SaaS landscape is different from what it was three or four years ago, especially with AI’s rise. This shift benefits younger companies, as disruption creates opportunities for them to challenge established players. Otherwise, big companies with their resources and reach would continue dominating. That’s the part that excites us — it opens up space for innovation and new entrants.

If we talk about billion-dollar revenue companies in SaaS of Indian origin, apart from Zoho and a few others, most are still striving to reach that mark. Zoho, ironically, stands out as the complete antithesis of what a funded company typically looks like — it’s bootstrapped yet generates billions. Meanwhile, many well-funded companies, where some have raised over half a billion, have yet to achieve comparable ARR. Given this backdrop and the high levels of disruption in the industry, do you think reaching a billion-dollar revenue will be feasible for most? Or will we end up with a proliferation of “bonsai” SaaS companies that will eventually consolidate?

The key consideration here is that if a SaaS company can transition from merely selling workflow tools to execute tasks, it not only caps software spending but also reduces costs associated with human capital — essentially, employee expenses. This shift dramatically expands the potential market for SaaS companies. The question is whether they can evolve from being purely workflow-focused to incorporating AI-driven solutions.

For those that manage this transition successfully, the market becomes significantly larger, allowing for faster growth. From what we’ve observed, AI companies are scaling revenue faster in their early stages compared to SaaS companies at a similar point. So, if they capture the AI opportunity, some of these companies could grow even faster than the broader market. However, for those unable to make this leap, challenges will undoubtedly arise, potentially leading to consolidation or other forms of realignment in the sector.

With the new value proposition AI brings, does the pricing model for SaaS also need to change? How much can a SaaS provider realistically charge? Many companies have already shifted to usage-based pricing, so it seems increasingly challenging to justify prices and drive revenue growth. What’s your perspective?

There’s potential to charge more. Previously, SaaS companies relied on seat-based pricing, where growth depended either on the client company hiring more employees or on cross-selling additional products. But now, by saying, “I’ll handle the work for you,” AI-driven SaaS enables companies to theoretically need fewer employees. This means they don’t have to expand their workforce in line with business growth, creating efficiency gains.

A portion of those gains should benefit the customer, and some should benefit the SaaS provider. If a company can shift from seat-based to usage-based pricing effectively, there’s a strong case for charging more. The main challenge, though, is that you’ll need to disrupt your existing pricing model to make this transition.

For instance, if you’re selling to the head of customer support, you’re currently charging per seat based on team size. With AI, you’re essentially proposing a shift — suggesting they’ll need fewer support staff and that you’ll price, based on usage instead. This can be a more difficult conversation, as it challenges traditional pricing assumptions.

However, if AI enables companies to reduce hiring needs for agents, lowers training costs, and lessens the need for software to support those agents, then, theoretically, you should be able to charge more since you’re delivering increased value. 

For SaaS companies, does the shift to AI-driven models imply more layoffs, especially as we’ve already seen significant reductions even before Generative AI’s full impact? How would this affect SaaS margin, given the current high gross margins? Do you foresee a phase of cost-cutting, with optimisation on the engineering or sales side? How do you see these trade-offs playing out?

The potential for optimisation isn’t unique to SaaS — it’s something that could affect all types of companies. For example, software, fintech, and tech companies in general may find ways to optimise headcount. If the engineering teams are using tools such as Copilot or various code-generating platforms, they might not need as many software engineers. This shift can occur across industries, not just SaaS.

Similarly, AI-driven tools are emerging to handle tasks such as prospecting or business development outreach, automating aspects of the BDR [business development representative] function that involve reaching out to potential clients and converting them. As this automation grows, B2B companies may need fewer staff for these roles, leading to efficiency gains that extend beyond just SaaS.

Regarding gross margins, a transition from software-based to AI-driven models could indeed change the cost structure. For traditional software companies, primary costs include cloud expenses for storage and compute. With AI, there’s an additional cost related to the AI model itself, which could impact margins. My hypothesis is that, while we haven’t yet seen many scaled application AI companies, their gross margins might end up lower than typical SaaS margins — perhaps around 50-60% rather than 75-80%.

However, this would still be relatively high, given that these companies provide a highly specialised service. If they continue to deliver substantial value, they should be able to capture that in their pricing.

There’s a growing, albeit not yet mainstream, narrative suggesting that with AI, companies might no longer need SaaS providers. Some customers are considering doing things in-house with AI, reducing their dependency on SaaS. For instance, one company recently decided to replace Salesforce, reasoning that they could handle those functions internally with just a few coders and generative AI. Doesn’t this pose a threat to the SaaS industry?

That perspective, while intriguing, is a bit simplistic. Take Salesforce, for example — it’s been around for over two decades and has invested tens of billions of dollars in building out its capabilities. It’s a complex, robust platform that addresses the needs of very large, mature organizations. For companies such as Bajaj Finance, one of Salesforce’s largest global customers, the platform manages vast workflows and intricate organisational processes.

While a young or smaller company might find a simplified in-house AI solution sufficient, it’s a different story for larger, more established companies with complex needs. To assume that a small team of coders could replicate everything that an enterprise-grade solution such as Salesforce has developed over decades would be a significant stretch.

But what if Bajaj Finance does that? Because Bajaj Finance is very strong on the tech side, and they’re using Salesforce. Hypothetically, while this may not fully align with their core as a financial services company, what about mid-sized competitors?

The question I’d pose for Bajaj Finance, if we fast-forward a few years, is about the potential for automation. Currently, they have agents who handle collections and reach out to customers for personal or home loans. If that process becomes fully automated, they may not need human agents for these tasks. For example, we have a portfolio company Sarvam AI, which has built multilingual voice agents capable of reaching out to customers, answering questions, gathering information such as employment type, first or second home status, and more. If an AI-driven voice agent can handle all these interactions and directly input data into a system such as Salesforce, Bajaj Finance might need fewer agents, impacting SaaS providers whose pricing models rely on per-agent fees.

Salesforce, Zendesk, Freshdesk, and similar companies are undoubtedly aware of this shift. The question in my mind is: how many of these companies will be able to quickly evolve their products to adapt to this new world? And, on the other hand, how many will be slower to respond because their customers — especially large organizations — tend to take time making decisions, and nothing will happen overnight. Some may move too slowly, which could open up opportunities for new companies to enter and capture market share.

Within your own portfolio, what conversations do you have with SaaS company founders? Are there any specific concerns you’re hearing from them?

If you’re asking whether SaaS companies are currently experiencing significant sales pushback or pricing pressure due to AI, that’s not really the case yet. Right now, AI companies are still relatively small in terms of revenue. Even if you combined the revenue of all application-focused AI companies globally, it would likely only reach a few hundred million dollars, which is minor compared to the tens of billions generated by SaaS. So, at this stage, AI companies aren’t significantly impacting SaaS sales.

The main conversation right now, especially at the board level, is about ensuring these companies are ready for the future with AI. We’re asking, “How are you evolving your product to stay relevant? Are you investing enough in AI to avoid potential disruption down the line?” Many of our SaaS companies are exploring how they might redesign their product from scratch in an AI-driven world and leverage their existing strengths to adapt. So, the focus is heavily on product innovation rather than immediate sales concerns. We’re still quite a way from AI directly affecting sales in a noticeable way.

Another factor with SaaS is that, just as Indian IT services, it has benefited from the cost advantage of being based in India — leveraging lower costs here while selling to overseas markets. But as we’ve seen with IT services, there’s now a need to be closer to the customer, with more sales teams and offices on-site. In today’s market, where every incremental sale is challenging, this means adopting higher-cost structures. Do you think this inherent cost advantage will eventually diminish?

While there’s often talk about India’s cost advantage in SaaS, I don’t think it’s played as big a role as it might seem.

But hasn’t this cost advantage played out to some extent in SaaS, especially as a source of gross margins?

While gross margins do reflect some cost efficiencies, they don’t typically account for R&D costs, which include product development. The expectation was that R&D and selling costs would be lower in India, contributing to higher margins. Zoho is a good example of a company that’s made this model work, but it’s not something many companies have fully achieved.

What companies have realised is that customers don’t primarily choose SaaS solutions because they’re cheaper; they choose them for the product’s competitive edge. While a few customers may focus on price, in general, a compelling product is more important than lower costs. This has become the mindset for most venture-backed software companies with access to capital.

Consequently, for many SaaS companies, especially those targeting mid-market or enterprise clients, founders and key teams often relocate to the U.S. to build out on-site sales and customer success teams. Larger clients expect in-person interactions and a closer presence, which is challenging to achieve solely through remote means. You can’t typically close a million-dollar deal over the phone; clients want to see and trust who they’re dealing with.

For smaller business (SMB) segments, a remote model might work. But as you move up to mid-market and enterprise clients, having a presence in the U.S. or the target region becomes necessary. So, while the cost advantage has played a role to some extent, it hasn’t been the driving factor in SaaS success. The companies that have truly excelled are those with globally competitive products.

SaaS has traditionally focused on the U.S. market. However, companies like Salesforce reached close to a billion dollars in revenue in India in FY23. Zoho is also starting to look inward. Are we at an inflection point where the Indian market could become just as lucrative? If Salesforce can reach this level in India, are SaaS companies missing an opportunity by focusing so heavily on the U.S. market, or is it a matter of some products not having a viable market in India?

I believe it’s somewhat market specific. Personally, I see a lot of potential in the Indian market. For example, I’m on the board of Darwinbox, an HR software company that serves major Indian enterprises like SBI and Vedanta. They’ve done exceptionally well in India and are now a unicorn. Similarly, Sarvam, which is making waves in the Generative AI space, is focused on Indian customers. So, there is a significant market opportunity here.

That said, the U.S. economy is around $30 trillion, compared to India’s $4 trillion. Software spending in India is still considerably lower, so certain niche markets that may be substantial in the U.S. are still relatively small in India. But for broad applications such as HR software, which serves nearly every company with employees, there is strong potential in India as well.

In short, while not every SaaS product has a large market here, we are bullish on India’s software opportunity. There’s enough potential in specific sectors to make it worthwhile.

Follow us on Facebook, X, YouTube, Instagram and WhatsApp to never miss an update from Fortune India. To buy a copy, visit Amazon.