Victoria’s startup agency is putting money behind the very earliest stage of company building, funding a new run of pre-accelerator programs designed to catch AI and deep tech founders before they have a product, a pitch deck or, in many cases, a co-founder.
LaunchVic, the Victorian Government body set up in 2016 to grow the state’s startup economy, says the programs will help aspiring entrepreneurs turn early ideas into businesses that can actually stand up to investor scrutiny. The move reflects a growing recognition across Australia’s innovation policy circles that the gap is no longer at the accelerator stage, where dozens of programs already compete for promising teams, but well before it, at the point where a researcher or engineer with a genuinely novel idea has no obvious first step.
What is on offer
Pre-accelerators sit a rung below the more familiar accelerator model. Where an accelerator typically takes a formed startup, invests a small amount and puts it through an intensive few months aimed at raising a seed round, a pre-accelerator works with individuals and loose teams who are still testing whether they have a business at all. The emphasis is on validating a problem, finding customers, sharpening a value proposition and, crucially for deep tech, working out whether the underlying science can be commercialised without years of burn.
According to LaunchVic’s announcement, the funding will support a suite of programs rather than a single provider, spreading the bet across multiple operators working with founders in artificial intelligence and deep tech fields such as advanced manufacturing, quantum, medical technology, robotics and clean energy. That portfolio approach has become a hallmark of the agency, which prefers to fund the organisations that run programs rather than back individual companies directly.
The logic is straightforward. Deep tech ventures are harder to start than a consumer app. They demand technical founders who can straddle a lab and a business model, they take longer to reach revenue, and they often stall not because the science fails but because the person carrying it has never learned how to build a company around it. By intervening earlier, LaunchVic is wagering that a larger and more diverse group of Victorians will make it far enough to attract private capital later.
Two ways to read the play
Supporters of this kind of intervention argue it addresses a real market failure. Private investors rarely fund a founder who has nothing more than a hypothesis, which means the most experimental ideas never reach the point where they can be assessed on their merits. Public money at the pre-seed and pre-company stage, on this view, is exactly what a government agency should be doing, because no rational venture fund will do it. It also widens the funnel to people who are locked out of the traditional startup networks, including academics, migrants, regional founders and women, who remain badly under-represented in Australian deep tech.
The sceptical reading is harder to dismiss than it once was. Critics of publicly funded innovation programs point out that a program running does not mean a company scaling, and that the sector has a long history of activity metrics standing in for outcomes. Counting founders who complete a workshop is easy. Counting the ones who go on to raise a round, hire staff and pay tax is the number that matters, and it takes years to land. There is also a persistent worry that a crowded support ecosystem can create founders who are very good at applying for programs and less good at finding customers. LaunchVic has generally been more disciplined than most on measurement, publishing follow-up data on the companies its programs touch, and the credibility of this latest round will rest on whether it can show the same.
Why it matters for Australia
Victoria does not operate in isolation, and the timing of this push is not accidental. Every state is now competing for the same scarce pool of technical talent and the same investor attention, at a moment when artificial intelligence has vacuumed up much of the oxygen in the funding market. New South Wales has its own investment vehicles, Queensland has been aggressive with its startup and advanced manufacturing programs, and South Australia has leaned hard into defence and space. A Victorian pipeline aimed squarely at AI and deep tech founders is, in part, a bid to make sure the next generation of technical companies chooses Melbourne over Sydney, or over San Francisco.
The national stakes are real. Australia has world-class research institutions and a chronic weakness at the point where that research turns into companies, a problem successive federal reviews have flagged without solving. Deep tech is where the country’s comparative advantage in science could translate into industries that outlast the resources boom, but only if founders can get past the earliest and riskiest phase. Programs like this one are a small piece of that puzzle, and their value will be judged less by how many people they enrol than by whether Australia ends up with more fundable, globally competitive companies emerging from its universities and labs.
What comes next
The immediate question is delivery. LaunchVic will need to select the organisations that run the programs, and prospective founders will want to know the eligibility criteria, the level of support on offer and whether there is any capital attached or purely mentorship and structure. Expect the agency to open applications through its chosen delivery partners in the coming months, with the first cohorts of founders moving through not long after.
The longer question is whether this cohort produces companies that raise money and grow. That answer will not arrive for two or three years, which is precisely why early-stage program funding is so easy to announce and so hard to prove out. For now, Victoria has restated a clear position: it wants to be the place where an AI or deep tech idea gets its first real backing, and it is prepared to spend public money to sit at the front of that queue.
Sources: LaunchVic
