How AI Is Killing the SaaS Playbook and Reshaping Venture Capital

The SaaS Model's Unraveling and the Rise of AI-Native Competition

For more than a decade, the venture capital playbook for software companies was straightforward: a SaaS business that reached US$20 million in annual revenue was almost certain to deliver a profitable exit. That assumption has disintegrated. The catalyst is the rapid advance of artificial intelligence, which can now replicate functionality that once took years of development in a matter of months. A small team with API access can build a competitive product against established incumbents, eroding the defensive moats that justified high SaaS valuations.

The shift is starkest in numbers from the US venture market. In 2025, American VCs invested US$340 billion, nearing the 2021 peak, but half of that capital went into just 0.05 per cent of all deals. Some 480 transactions, representing only 2.9 per cent of deal count, absorbed 70 per cent of the money. Meanwhile, sub-US$100 million deal volume hit a 14-year low in the fourth quarter. Capital is concentrating heavily in a small group of perceived AI winners, while the rest of the startup ecosystem is starved.

Australia is not immune. While the domestic VC market is far smaller—US$22 billion over the past decade versus the United States’ US$1.92 trillion—the same pattern is emerging. In 2025, 61 per cent of all Australian venture dollars went to companies that could plausibly claim an AI connection. Like their American counterparts, Australian backers are pouring money into the hot AI theme, but the local market can neither write cheques on the scale of the largest US rounds nor produce its own frontier-model companies.

The case of AI coding assistant Cursor shows how quickly the ground can shift. The company grew from a US$400 million annualised run-rate in 2024 to US$4 billion in 2026 and was widely seen as the dominant player. Then Anthropic released its Claude Cowork product, and within months Cursor’s lead evaporated. When Cursor tried to raise at a US$50 billion valuation in the second quarter of 2026, the market refused, even though revenue was still growing rapidly. It ultimately sold into Elon Musk’s SpaceX/Grok group for stock at US$60 billion. For investors, the lesson is that a tenfold revenue increase over two years may no longer be enough to command a premium valuation if an AI rival can close the gap almost overnight.

What the Shift Means for Investors, Startups, and Australia's VC Landscape

Why the Old SaaS Multiples Are Unsustainable

Investors who were once comfortable paying 15 to 25 times revenue for a SaaS business growing at 35 per cent annually are now rethinking those assumptions. The core concern is that the protective moat—a combination of switching costs, scalable technology, and customer traction—has become much shallower. Large language models and AI-native applications can enter vertical use cases with speed and efficiency that traditional software firms cannot match. Unless a company can credibly articulate how it will withstand AI-driven competition, mid-stage non-AI startups are finding it increasingly difficult to attract capital, regardless of their growth and retention metrics.

The Barbell Dynamic in US and Australian Markets

The US venture scene has adopted a barbell approach: massive, valuation-insensitive bets on a handful of companies everyone has already decided are winners, and tiny, speculative cheques spread across hundreds of hopefuls. The middle—where disciplined, analysis-driven investing in promising but not yet dominant companies once lived—is being squeezed out. Australia exhibits its own, smaller-scale version of this barbell, with large deals at one end and small ones at the other. The difference is that even the large end of the local barbell doesn’t reach global scale. Australian generalist VCs, many of whom pitch themselves as ‘can do everything’ funds, are caught in the middle and risk irrelevance if they cannot differentiate.

Who Gains and Who Loses

The biggest winners are the top-tier US firms that have the capital to dominate late-stage AI rounds and the brand to attract the best founders globally. Non-traditional capital—sovereign wealth funds, superannuation funds, and hedge funds—are also pushing into late-stage venture, writing growth-equity cheques into already anointed winners, a task that requires little specialist skill. The losers are mid-sized generalist funds and the thousands of well-run but non-AI startups that are being overlooked. For Australia, the risk is that a landscape dominated by generalist firms will miss the opportunity to back deep tech, hardware-software combinations, and other long-duration bets that do not fit the current AI hype cycle but which align with the country’s R&D strengths.

The Case for Specialist Funds

Historical data persistently shows that larger venture funds return less, not more, than smaller ones. This suggests that as the current AI frenzy cools, investor appetite should eventually swing back toward smaller, deeply specialist managers—those focused on drug discovery, frontier materials, or hardware-centric startups with genuine intellectual property moats. For Australia, a nation that punches above its weight in many areas of global research, the dearth of such specialist funds is a structural weakness that Daniel Petre, co-founder of AirTree Ventures, argues must be addressed. The institutional backers of Australian VC—superannuation funds and government-linked investors—could play a decisive role by choosing to allocate to smaller, domain-expert teams rather than continuing to pour money into generalist vehicles that are increasingly trying to surf the AI barbell.

A Path Forward for Australian Investors and Founders

• Australian institutional funders should actively seek out specialist VC managers with proven domain expertise in areas such as deep tech, quantum computing, or drug discovery. These funds are better positioned to commercialise the country’s R&D strengths and, according to historical performance data, tend to outperform larger generalist competitors.
• Generalist VC firms must develop a credible, differentiated AI thesis. Without a clear story around how their portfolio companies can either withstand or leverage AI disruption, they risk capital flight to both the mega-round winners and the specialist boutique funds.
• Mid-stage SaaS startups need to articulate an AI moat now. Companies that cannot explain why AI will not eventually erode their advantage will find it increasingly difficult to raise growth capital, even with strong current metrics. Leadership teams should prepare for a tougher fundraising environment and consider strategic partnerships or earlier exits if necessary.
• Investors in existing SaaS-heavy portfolios should stress-test holdings against the threat of rapid AI replication. A basket that looked diversified a year ago may now carry concentrated vulnerability that demands active reappraisal.