The seed market for commerce technology in APAC looks materially different in September 2024 than it did in 2022. The change isn't primarily in valuations — those have normalised from the peaks, but not to the degree that some people predicted when the rate cycle turned. The more significant change is in the quality distribution of companies seeking first institutional capital. The cohort of founders coming to seed in 2024 has, on average, more operational depth, cleaner thinking about the commercial model, and a more honest accounting of the go-to-market challenge than the equivalent cohort in the 2021–2022 vintage. Some of that is a selection effect — the founders who would have rushed to market in a permissive capital environment waited, and the extra time showed up as product and customer development. Some of it is the AI capability shift, which has allowed small founding teams to build and ship more substantive product with less capital than was possible two years ago.
What hasn't changed is the bar for what constitutes a genuine insight. The funding market reset removed some of the noise — the companies that were raising on vibes and surface-level AI positioning have largely not materialised into repeat pitches — but it didn't make the genuine insights more common. A founder who has identified a real structural problem in APAC commerce infrastructure, has a non-obvious theory about why the existing solutions fail, and has the operator background to understand what solving it actually requires is still rare. That rarity is the investment opportunity, and it doesn't change with the funding cycle.
One shift that we've tracked carefully is the changing relationship between AI capability and seed-stage product ambition. In 2022, a seed-stage commerce AI company was typically proposing to build a specific, bounded ML model — a recommendation system, a pricing model, a demand forecasting system — and the technical ambition was bounded by what a small team could ship before they ran out of runway. In 2024, the AI capability available to a two or three-person team has expanded enough that the technical scope of what's plausible at seed has shifted. A team can now propose a more ambitious product and demonstrate more of it before the first institutional cheque — which is good for diligence quality but also means the signal-to-noise problem in evaluating AI capability claims is harder than it was. Anybody can demo a product built on foundation model APIs. The question is what the proprietary layer is and how defensible it becomes as the category matures.
The geography of the APAC commerce seed market has also shifted in ways worth noting. In 2021 and 2022, the preponderance of interesting commerce infrastructure companies we saw were based in Sydney or Melbourne, with a smaller number in Singapore. In 2024, Singapore's position as a launch market has strengthened for founders targeting Southeast Asian commerce corridors, and we're seeing more founders who started in Australia and are structuring their companies with Singapore as the APAC headquarters from the early stages. That structure makes sense for companies whose primary market opportunity sits in the Indonesia-Malaysia-Thailand corridor — but it introduces fundraising complexity because the Australian seed ecosystem and the Singapore seed ecosystem have different risk appetites and different LP bases, and a company structured in Singapore with Australian founders is sometimes viewed as ambiguous by both.
Our conviction about the category has grown since we started Banksia. The underlying thesis — that APAC commerce infrastructure is being rebuilt from scratch on AI-native architecture — has played out more clearly than we anticipated in the early years, and the portfolio companies that have reached the point of demonstrable commercial traction have done so in ways that confirm the structural nature of the opportunity rather than explaining it away as market timing or execution flukes. The seed stage in APAC commerce is still early. Most of the best companies in this category over the next decade haven't been founded yet. That's a good place to be as an investor with a specific thesis and a track record of backing the right founders early.