The conversation about Australian semiconductor manufacturing has shifted from “should we” to “how fast can we” over the past 18 months. There’s bipartisan political support, industry advocacy, and genuine security concerns driving the push. But the gap between ambition and industrial capability is enormous, and not enough people are being honest about it.

Let’s separate what’s realistic from what’s wishful thinking.

What’s Actually Happening

Australia currently has no commercial semiconductor fabrication facilities. We’re not talking about being behind—we’re starting from scratch. The closest thing is some R&D capability at universities and CSIRO, but that’s light-years away from production manufacturing.

The federal government’s National Reconstruction Fund has earmarked $1 billion for “critical technologies,” with semiconductors as a priority area. Several state governments have announced their own initiatives—South Australia’s put forward $20 million for a semiconductor research hub, and Queensland’s exploring partnerships with international foundries.

In February, a consortium of Australian companies and universities signed a memorandum of understanding with a Taiwanese equipment manufacturer to explore building a pilot fabrication line in Sydney. It’s very early stage—no committed capital, no site selected, no production timeline. But it’s more concrete than most of the announcements we’ve seen.

There’s also movement on the packaging and testing side. This is the final stage of chip production, where dies are assembled into finished products. It’s less technically demanding than fabrication, and Australia already has some relevant manufacturing capability in precision assembly.

The Economics Are Brutal

A modern semiconductor fabrication plant—a “fab” in industry parlance—costs between $10 billion and $20 billion to build. That’s for a leading-edge facility producing chips at 5nm or smaller process nodes. Even an older-generation fab producing 28nm or 40nm chips runs several billion dollars.

And that’s just construction. Operating costs are massive: ultra-pure water, cleanroom environments, constant equipment maintenance, and highly specialized labor. Chip fabs run 24/7 because shutting down and restarting is prohibitively expensive.

The business model only works at enormous scale. TSMC’s fabs in Taiwan produce hundreds of thousands of wafers per month. You need that volume to amortize the capital costs and compete on price. An Australian fab producing 10,000 wafers monthly—which would be ambitious for a new facility—simply can’t match the economics.

Intel’s building new fabs in Arizona and Ohio with massive US government subsidies, and even they’re struggling with the economics. The CHIPS Act provided $52 billion in funding, and industry estimates suggest that covers maybe 30% of the actual investment needed to rebuild American semiconductor capacity.

Australia’s $1 billion fund doesn’t go far in this context. It might support R&D, workforce development, or pilot facilities. It won’t build a globally competitive commercial fab.

Where Australia Might Actually Compete

The realistic opportunities aren’t in competing with TSMC or Samsung on mainstream production. They’re in specialized niches where Australia has specific advantages or where security concerns justify higher costs.

Compound semiconductors are one option. These are chips made from materials like gallium nitride or silicon carbide, used in power electronics, RF applications, and some defense systems. They’re lower volume than silicon chips, and the manufacturing technology is different—which means the competitive moat built by Asian fabs is less relevant.

Australian researchers have done solid work in compound semiconductors, particularly for solar and communications applications. There’s a clearer path from university research to commercial production here than in mainstream silicon manufacturing.

Radiation-hardened chips for space applications are another niche. Australia’s growing space industry needs components that can survive in orbit, and there’s limited global supply. The volumes are tiny compared to consumer electronics, but margins are higher and security concerns matter.

Trusted fabrication for defense is the most commonly cited justification. The argument is that critical military systems can’t rely on chips manufactured in geopolitically uncertain regions. That’s legitimate, but it’s worth being clear-eyed: defense demand alone won’t support a commercial fab. It’d be government-subsidized production for specific applications, not a profit-driven industry.

There’s also packaging and testing. This doesn’t get the same attention as fabrication, but it’s a real manufacturing capability that Australia could plausibly build. Several Asian packaging facilities are looking at geographic diversification, and Australia’s got a skilled workforce and political stability. The capital requirements are lower than for fabs, and the technology’s more accessible.

The Workforce Problem Nobody’s Solving

Even if money materialized for a fab, there aren’t enough people in Australia who know how to run one. Semiconductor manufacturing is incredibly specialized. You need process engineers who understand thin-film deposition, photolithography technicians, yield optimization specialists, equipment maintenance crews with specific training.

Taiwan has universities churning out thousands of semiconductor engineers annually. South Korea has entire technical colleges focused on chip manufacturing. Australia has maybe a few dozen people with direct fab experience, and most of them left the country years ago when opportunities dried up.

Training a semiconductor workforce takes a decade, minimum. You can’t shortcut this with boot camps or accelerated programs. It’s deep technical knowledge built through years of hands-on experience. Even if Australia started building that educational pipeline tomorrow, you’re looking at 2035 before you’ve got a meaningful talent pool.

Some proposals involve bringing in experienced workers from Taiwan or South Korea. That’s feasible for a small pilot facility, but it’s not a sustainable foundation for a domestic industry. And politically, it’s awkward—the whole point is strategic independence, which is undermined if you’re dependent on foreign expertise.

What Policy Actually Helps

The most useful thing government can do is invest in research infrastructure and workforce development. Not because that’ll lead to a fab opening in three years, but because it builds foundational capability that might pay off in a decade.

Expanding university programs in semiconductor engineering, funding PhD positions, creating industry placement programs—that’s unglamorous work that doesn’t generate headlines, but it’s necessary if Australia’s serious about this.

Supporting compound semiconductor research makes sense. It’s a genuine niche where Australia could develop expertise, and the capital requirements are more realistic. CSIRO and several universities are already doing work here; scaling that up is achievable.

On the defense side, there’s a role for government as an anchor customer. If Australia’s going to build trusted fabrication capability for military applications, the Department of Defence needs to commit to buying domestic chips even when they’re more expensive than offshore alternatives. That means accepting higher procurement costs, which requires political will.

For packaging and testing, the government could facilitate partnerships with international firms looking to diversify supply chains. Tax incentives, infrastructure support, streamlined approvals—create an environment where it makes sense for an Asian packaging company to set up an Australian subsidiary.

The Honest Conversation We’re Not Having

Here’s what politicians don’t say: building a globally competitive semiconductor industry in Australia probably isn’t possible, and trying to do it would consume resources that could be better spent elsewhere.

That doesn’t mean doing nothing. It means being strategic about where Australia can actually add value—specialized niches, research excellence, trusted fabrication for specific defense needs. Those are achievable goals with realistic funding.

What’s not achievable is Australia becoming a meaningful player in mainstream commercial chip production. The economics don’t work, the workforce doesn’t exist, and the competitive dynamics favor established clusters in Asia and (with massive subsidies) parts of the US and Europe.

The risk is that political pressure for visible action leads to wasteful spending on showcase projects that never reach commercial viability. We’ve seen this pattern before in Australian manufacturing policy—big announcements, government co-investment, facilities that operate for a few years burning public money, then quiet closure when political attention moves on.

A measured approach focused on research, workforce development, and carefully selected niches would be more productive. It’s less exciting than ribbon-cutting ceremonies at new fabs, but it’s what the industrial reality demands.

The semiconductor conversation in Australia needs less optimism and more engineering realism. What can we actually build, with the resources and people we have? Start there, and work forward. The alternative is expensive disappointment.