The Country That Invented WiFi and Couldn't Sell It

We keep treating this like an invention problem. It’s a commercialisation problem — and the latest tax proposal aims straight at the people who solve it.

Inventing the thing was never our problem

The WiFi in the laptop you’re reading this on was invented in a Sydney lab. Radio-astronomers at the CSIRO, chasing exploding black holes, worked out how to send a clean wireless signal through a room full of echoes. It went into the global 802.11 standard. Most of the connected world runs on it.

Australia made about $430 million from that invention. Not by building a product. Not by founding a company anyone has heard of. We made it by getting a patent granted in 1996 and then spending the better part of a decade suing Intel, Dell, Microsoft, HP and Belkin through American courts until they paid up. The patent expired in 2013. The cheques stopped. That’s the story we put in the brochure.

Now look at the rest of the highlight reel. Polymer banknotes: the research started in the late 1960s, the first plastic note reached a wallet in 1988, the full set landed in the 1990s. Roughly two decades from lab to till. Aerogard took a similar walk from a CSIRO entomologist’s bench to a can you could buy. The vaccine work, the radio telescope, the diet book, the longwall mining automation — some of it took a generation to reach a market, and a chunk of it never reached one at all because it was never meant to.

That last part deserves its own line, because it flatters the list. Myxomatosis, the biological control of the prickly pear, the Parkes dish that helped carry the pictures of the moon landing — some of the most celebrated work the CSIRO ever did was never going to be a product. It was public-good research, done for the country, with no market at the end of it by design. Strip those out and the commercial highlight reel gets shorter and lonelier than the brochure lets on.

Run your eye down that list and the pattern is hard to miss. The inventing was the easy part. We are genuinely world-class at the inventing. What we have never been able to do, reliably, at any scale that moves the economy, is take the thing we invented and turn it into a business that sells.

This matters right now because there’s a tax change on the table that treats the problem as if it were the opposite. As if Australia is drowning in commercial ambition and the job of policy is to skim a little more off the top. We aren’t, and it isn’t. We’re the country that invented WiFi and couldn’t sell it, and we’re about to make selling it harder.

Our most famous success came out of a courtroom

Spend a second on the WiFi number, because it’s the one everyone reaches for, and it’s the most damning entry on the list.

The single biggest commercial return the national science agency ever produced came from litigation. From a courtroom, years after the industry it enabled had been built by other people. We out-built nobody. We shipped no device. We held a piece of paper, watched the rest of the world build a multi-billion-dollar industry on top of our idea, and sent lawyers to collect rent before the clock ran out.

That’s not a criticism of the CSIRO. Defending the patent was the right call, and the people who did it were sharp. It’s a criticism of the model. When your proudest commercial success is an enforcement action, you’ve told on yourself. You’ve admitted that the part where someone takes the idea, raises the money, hires the team, builds the thing, finds the customer and survives long enough to scale (the part that turns an invention into wealth) happened somewhere else. To someone else. In another country, under another flag, funded by other people’s capital.

The $430 million looks like a win. Measured against what the 802.11 industry was worth, it’s a tip. We invented the thing and captured a rounding error of the value, because capturing the value was never our strength. The capturing is the bit we keep losing.

Invention is the cheap part. The risk lives in commercialisation

Here’s the bit policy keeps getting wrong, decade after decade, government after government. It treats invention and commercialisation as the same activity with the same incentives. They’re nothing alike.

Invention is a cost. You fund a lab, you pay clever people, you wait, and sometimes something comes out. The risk is real but it’s bounded and it sits in an institution that won’t go bankrupt if the experiment fails. That’s why governments are comfortable funding it. A failed research grant is a line item. Nobody loses their house.

Commercialisation is the opposite shape of risk. Someone has to look at an unproven idea and put their own years, their own money, their own reputation behind it, knowing the most likely outcome is that it dies and takes the lot with it. Most startups fail. That isn’t a bug in the system, it’s the physics of it. The whole model only works because, occasionally, one of them doesn’t — and that one pays for all the others and then some.

For that arithmetic to function, the person carrying the risk has to be looking at a payoff big enough to justify the near-certainty of loss. That’s the entire deal. Founders and the investors behind them take a wildly asymmetric bet (small chance of a large win, large chance of nothing) and the size of the possible win is the only thing holding the equation together. Shrink the win and the bet stops being worth taking.

This is the group that crosses the gap we can’t cross. Not the science agency. Not the grant program. Not the minister with the innovation statement. The founder who remortgages, and the investor who writes a cheque they can afford to lose, are the only actors in the entire system whose job is specifically to turn an Australian idea into an Australian company that sells. They are the commercialisation layer. They are the thing we’ve been missing for fifty years.

So when a country with our track record sits down to design tax policy, the first question ought to be: does this make that bet more attractive or less? Because we need many more of those bets than we’re getting.

Every government for forty years has starved the same thing

This is where it would be easy, and cheap, to turn it into an Albanese problem. It isn’t one. He happens to be holding the latest version of it, and the budget has his name on it, but the habit is bipartisan and it’s about forty years deep.

Look at the record, the one that doesn’t care which side is in. Australian R&D spending has been sliding as a share of the economy for more than a decade — from around 2.24 per cent of GDP in 2008 down to roughly 1.68 per cent, while the OECD average sits up near 2.71 per cent. Government R&D as a share of GDP recently hit its lowest level on record. That slide happened under both colours of government. Nobody owns it because everybody does.

Then there’s the R&D Tax Incentive, the main lever we use to nudge companies toward research. It was introduced in 2011, replacing the scheme before it, and it has been reviewed, trimmed, redesigned and re-trimmed almost continuously ever since — including, again, in the 2026–27 budget, with changes landing in 2028. Pick any five-year window in the last twenty and you’ll find a founder who built a plan around the rules as they were, only to watch the rules move. That instability is its own tax. Capital hates a moving target, and we’ve made the target move on a near-permanent basis, regardless of who’s in.

And underneath all of it sits the innovation-statement cycle. Every few years a government discovers innovation, announces a glossy package, gets a news cycle out of it, and then the attention drifts to the next thing before anything compounds. We’ve had ideas booms and digital economies and advanced-manufacturing pushes. What we’ve never had is forty years of a stable, boring, predictable deal for the people who take the commercial risk. The story keeps changing. The one constant is that the commercialisation layer gets treated as a rounding error.

So no, this belongs to no single party. It’s a national habit, forty years in the making, and the CGT proposal is just the newest brick in a wall we’ve all been building.

The tax change aims straight at the behaviour we need more of

Here’s the proposal that started the fight. The government is looking at scrapping the 50 per cent capital gains tax discount (the rule that halves the tax on a gain when you’ve held an asset longer than a year) and replacing it with a cost-base indexation model. On top of changes to negative gearing. The detail will move around, the way these things do, but that’s the core of it.

Now, you can mount a perfectly reasonable argument that the 50 per cent discount is too generous for property speculators flipping their fourth investment unit. Plenty of economists have. The problem is what a blunt change to CGT does to the founder and the early investor, who are sitting in the exact same tax bucket as the property speculator but doing something the country desperately needs.

Think about the early-stage investor. They put money into ten startups. Seven die. Two limp along. One, if they’re lucky, becomes something. The entire return on that portfolio is concentrated in the long-held gain on the one that worked, and the maths only stands up because that gain is taxed in a way that respects the years of risk it took to get there. Weaken the treatment of that gain and you haven’t touched the seven failures — those still cost what they cost. You’ve reached in and shaved the single payoff the whole strategy depends on. The asymmetric bet gets less asymmetric in the wrong direction.

Same story for the founder who took shares instead of salary for five years, ate the risk, and finally has an exit in sight. The long-held capital gain on their equity is the entire compensation for half a decade of underpaying themselves and betting their time on a coin that mostly lands tails.

The clearest signal that the design is pointed at the wrong target is that the government is already talking about carve-outs for startups. Read that move carefully. You only need a carve-out when your main instrument hits something you didn’t mean to hit. The need to exempt the founders is an admission that the policy, as drawn, lands on exactly the people you’d least want to discourage. Better to design an instrument that doesn’t need the patch. A carve-out bolted on under political pressure is the least stable kind of rule there is — it’s the first thing trimmed in the next budget, which brings us right back to the moving-target problem.

Risk capital reallocates the moment the deal gets worse

There’s a comfortable assumption buried in revenue policy: that if you tax a thing a bit more, you get the same amount of the thing plus more tax. For most settled activities that’s roughly true. For risk capital it isn’t, because the capital is mobile and the behaviour is optional.

Money doesn’t sit still and absorb a worse deal. It goes looking for a better one. Make the after-tax return on backing a risky Australian startup less attractive and the capital reallocates. Some of it goes into property, which is safer, more familiar, and where we’ve spent decades teaching it to go. Some of it goes offshore, into jurisdictions that have figured out it’s worth bending over backwards to keep founders and their early backers onshore. Some of it just stays in cash and does nothing useful at all.

Be honest about how strong that property pull already is. We’ve built one of the most efficient machines in the world for funnelling private capital into housing: negative gearing, the existing CGT discount, an entire advice industry pointed at the investment property. A startup raising its first round competes against more than other startups for that money. It competes against a tax-advantaged, bank-financed, culturally default bet on a third rental. Make the startup side of that contest less rewarding and you don’t need a forecast to know which way the money leans. It leans the way it always has.

And the cost of that reallocation doesn’t show up as a line in the budget. It shows up as an absence. As the company that never got its second round and folded. As the founder who looked at the numbers, did the maths on the downside, and took the safe corporate job instead of starting the thing. As the engineer who builds the next 802.11 in San Francisco because that’s where the capital that rewards the risk lives.

You can’t see any of that in a revenue forecast. The Treasury model can tell you, more or less, what scrapping the discount raises. It cannot tell you the value of the companies that now never get funded, because they don’t exist to be counted. That’s the trap with this kind of policy. The benefit is visible and bankable. The cost is invisible and lands on the people with the least power to complain — the founders who never start, in a country that already can’t afford fewer of them.

Risk, and the work of building, relocate to wherever the deal is better. Right now we are busy making sure that’s somewhere else.

The number worth running

I’d put one question to anyone weighing this up, on either side of politics, because the framing matters more than the party.

For every dollar of additional capital gains tax this change is forecast to raise from early-stage investment and founder equity, what’s the expected value of the Australian company that now never gets funded?

If you can’t answer that (and you can’t, because the model doesn’t hold the companies that don’t get built) then you’re trading a number you can see for a number you’ve chosen not to. That’s a fine trade for a property portfolio. It’s a terrible one for the one layer of the economy we’ve spent fifty years failing to grow.

We are very good at inventing things. We have a national lab full of proof. The gap has always been everything that happens after the invention: the founding, the funding, the building, the selling, the surviving long enough to scale. That gap is crossed by people taking an asymmetric bet on an Australian idea, and the size of the possible win is the only thing that makes the bet worth taking.

You don’t have to love every founder or every venture fund to see the shape of this. A country with our record should be doing everything it can to make that bet more attractive, more often, with rules that hold still long enough for someone to plan around them. Instead we’re reaching for the one part of the deal that makes the whole thing work, and trimming it, while telling ourselves we’ll patch the damage with a carve-out later.

We invented the wireless standard the world runs on and captured a tip for it. If we want a different ending next time, the work isn’t in the lab. It’s in giving the people who do the selling a reason to do it here.