Artificial intelligence (AI) is moving from a set of tools that help people work faster to systems that can capably perform entire workflows. For long-term investors, that shift is changing the way value is created, where new companies are formed and how existing businesses defend their position.
Australian venture capital firm Square Peg’s Partner and Head of Distribution Leila Lee argues that the next phase of AI will be defined less by broad enthusiasm for the technology and more by disciplined backing of founders building durable companies in large markets. “We are at the start of what may be the most consequential technology era yet,” she says. “The returns will come from being long-term and disciplined, not from chasing the hype.”
Square Peg has spent 14 years backing technology founders across the Asia Pacific region, including early investments in companies such as Canva, Airwallex and Supabase. The firm’s current focus reflects a broader shift in venture capital: AI-native companies are no longer just selling software budgets, but increasingly taking on the much larger cost base associated with services and labour.
That creates a larger opportunity set than previous technology cycles, but also a more difficult environment for investors. Entry valuations for fast-growing AI businesses have risen sharply, while failure rates remain high. In this setting, the challenge is not simply to identify exposure to AI, but to distinguish companies with lasting economic advantages from those benefiting from short-term momentum.
For Lee, the defining question is whether a company is using AI to create a deeper product advantage, stronger customer adoption or access to a market that was previously too difficult or costly to serve.
While Square Peg invests across the technology stack, much of its recent early-stage capital has been directed to the application layer. These are companies using AI to move software from assisting work to performing specific tasks, including business tax, customer support, sales intelligence and advertising.
The appeal of these businesses is that they can address budgets that traditional software companies have not always been able to reach. If AI can complete a workflow rather than simply improve it, the market opportunity expands from software spend to labour and services spend.
Vertical, AI-native software is one area where this is becoming particularly important. In specialised industries, durable advantages may come from proprietary data, deep integration into existing workflows, domain-specific judgement and low tolerance for error. In those settings, general-purpose language models may not be enough on their own.
Supabase illustrates how quickly the role of a company can change in an AI cycle. Square Peg first backed the business in 2022, when annualised revenue was less than USD 1m. Founded as an open-source alternative to Google’s Firebase, Supabase had built a strong following among developers and was becoming a common way to set up a database.
Since then, the business has benefited from a shift in how software is built. AI agents creating software need databases to be launched quickly and reliably, and Supabase has become part of that workflow. According to Square Peg, database launches on the platform have grown significantly over the past year, with a growing share now created by AI tools rather than human developers.
The example is notable because AI appears to expand the company’s moat rather than erode it. Developer adoption, open-source distribution and product depth were already important. The rise of AI-assisted software creation has made those attributes more valuable.
AI is not only creating new companies. It is also forcing existing businesses to reassess their competitive position. Lee defines an incumbent simply as any business launched before ChatGPT, which means the term can apply as much to a young software company as to a century-old enterprise.
Square Peg’s framework starts with two questions. First, is AI a structural tailwind or headwind for the business? Second, is the company behaving like an AI native? The first question is largely determined by the market and the company’s existing moat. The second is a choice, shaped by capability, urgency and mindset.
While companies cannot choose their structural exposure, they can control the quality of their response. For investors, that makes management behaviour and organisational pace increasingly important when assessing how AI will affect existing businesses.
For private investors, AI presents both an opportunity and a discipline test. The scale of potential value creation is significant, but early-stage investing remains volatile. Outcomes are uneven, and a small number of companies typically account for a large share of returns.
Lee’s view is that valuation matters but is unlikely to be the main driver of long-term performance if the companies selected are not exceptional. In venture investing, the quality of the business and founding team can matter more than small differences in entry price.
That requires breadth, patience and access to founders early in their journey. It also requires a willingness to look beyond the most visible AI themes. While foundational models and infrastructure continue to attract attention, the application layer may be where many investors see AI’s effect most directly, as software begins to absorb more complex work.
The global AI conversation remains heavily centred on the US, but Square Peg sees an important opportunity across the Asia Pacific. Australia has already produced globally recognised software companies, while Singapore is emerging as a more important technology hub.
For much of the past decade, Southeast Asian technology was often viewed through the lens of regional platforms serving regional markets. That is changing as more companies from the region build for global customers from the start. Supabase is one example, while newer companies such as Octen AI and PixAI point to a broader pipeline.
The next phase of AI will not only be about the largest model providers or the most highly funded companies. It will also be shaped by specialist applications, infrastructure businesses and founders using AI to solve very specific problems in large markets.
For long-term investors, the question is not whether AI will matter. It is how to participate in the value it creates while remaining selective, patient and realistic about the risks.
Receive ongoing access to LGT Wealth Management’s insights and observations - a curated stream of thought leadership, market perspectives, and strategic updates designed to inform sophisticated investors.
