
By Mike Nissen
Away from the showroom floor
Over the last two editions of this column, we have toured the shiny end of our industry: the warehouse as the incubator where AI and robotics grow up, and the software layer that is teaching buildings to think ahead. This month, as promised, I want to go somewhere less glamorous – because it’s where most of you actually live.
Here is the uncomfortable arithmetic behind every gleaming automation render. The vast majority of the freight moving through this region tonight will not pass through a purpose-built, eleven-metre-clear, robot-ready fulfilment centre. It will pass through a shed. An ageing, awkward, often fully depreciated building with columns in the wrong places, a floor that ripples like the Tasman Sea in a southerly gale, dock doors designed for trucks two generations smaller, and a lease with three years left on it. Nobody photographs these buildings for a brochure. They are also where the next decade of automation will actually happen – which is why I think of the years ahead as the brownfield decade.
Why the old shed suddenly matters more
For years, the standard answer to an inadequate building was simple: move. Sign a pre-commitment on a greenfield site, specify the slab to robot tolerances, and leave the old shed behind. That escape route is narrowing, and the reason is one of the most interesting collisions in property markets right now.
Industrial land has acquired a new bidder that doesn’t ship a single carton: the data centre. Australia now carries a data-centre investment pipeline estimated at around $155 billion – enough to rank it among the top three destinations in the world – and those developers are reportedly paying 20 to 60 per cent above industrial land values in outer suburban markets, and far more again in infill locations close to power and fibre. Analysts are openly discussing scenarios in which the rent required to justify new warehouse construction sits 30 to 100 per cent above today’s prime rents in major Australian markets. Similar pressures are building across Asia-Pacific, where logistics, manufacturing and digital infrastructure are all chasing the same serviced land.
Translate that out of property-speak and it says something every operations director should pin to the wall: the cheapest, best-located warehouse you will control for the rest of this decade is very likely the one you already occupy. The question is no longer “when do we move?” It is “how do we make this building perform like one we can no longer afford to build?”
What actually works between the columns
The good news is that the technology has been quietly reorganising itself around exactly this problem – and it is no coincidence that the stars of the last two columns are the heroes of this one.
Autonomous mobile robots earn their place in a brownfield precisely because they adapt to the building rather than demanding a new one. They steer around the columns, tolerate the imperfect floor far better than rigid conveyor ever did, and – critically – they can be piloted in one zone, proven, and scaled, without a weekend of demolition. Autonomous case-handling robots climb inside existing racking. Modular shuttle systems and four-way pallet shuttles can be installed aisle by aisle, in sections, while the rest of the operation keeps shipping. Goods-to-person cells can colonise a corner of the floor and expand as confidence grows. Even automated storage can now be retrofitted in stages rather than arriving as a single, bet-the-company megaproject. Add the quiet achievers – drones and camera systems flying inventory counts through existing racking overnight, vision-equipped stations catching errors at pack – and a pattern emerges: the technology that wins in a brownfield is the technology that treats the building as a given, not a variable.
And the cheapest retrofit of all pours no concrete whatsoever. Everything I wrote last month about the warehouse that thinks ahead – predictive execution systems, intelligent order release, orchestration – applies with special force to older buildings, because software doesn’t care where your columns are. I have seen operations lift throughput double digits inside the same four walls purely by letting better software re-sequence the work. In a brownfield, intelligence is the first renovation.
My honest read: the obstacle is the spreadsheet, not the shed
Here is what the case studies gloss over. The hard problems in brownfield automation are mostly economic and contractual, not technical.
Start with the lease. A payback period of five years is respectable for automation; a lease with three years to run makes it impossible arithmetic. Too many good projects die in exactly that gap. The emerging answer is to negotiate the lease and the automation as one conversation – operators trading longer commitments for landlord contributions, and the smarter landlords realising that a tenant who bolts automation into your building is a tenant who stays. An automation-ready building, or one with an automation-invested tenant, is simply a more valuable asset, and the land squeeze makes that truer every quarter.
Then there are incentives. A surprising amount of the region’s freight is handled by third parties whose commercial models were built on charging for manual activity – which means the party best placed to automate is sometimes the party with the least reason to. If you are outsourcing, interrogate the contract before you interrogate the technology: gainshare structures and open-book arrangements do more for automation than any robot demo.
And, as always, the floor matters less than the data. A brownfield site with accurate stock, clean order history and honest process data is a better automation candidate than a brand-new shed flying blind. The constraint I’d check first isn’t slab flatness. It’s whether you can trust your own numbers – because every technology in this article feeds on them.
What this means for you
If you operate from an older building, stop treating it as a waiting room. Map the constraints honestly – clear height, column grid, floor, power, sprinklers – because they will choose your technology shortlist for you, and that’s fine; the shortlist for awkward buildings has never been stronger. Pilot in one zone with a robot fleet you can scale or walk away from. Put the software layer in early, because it improves the business case for everything that follows. And bring your CFO and your landlord into the room at the start, because the decisive variables – lease length, rent trajectory, residual value – live in their spreadsheets, not on the warehouse floor.
If you are a landlord, the message is blunter: the tenants automating your ageing asset are doing you a favour. Meet them halfway.
Until next issue
I said industrial land had found a new bidder, and I meant it literally. Across this region, the data centre industry is competing for the same serviced sites, the same construction crews and – above all – the same electrical capacity as the warehouses that feed our cities, with a pipeline measured in the hundreds of billions. What happens when the shed that stores your stock and the shed that stores your data want the same piece of dirt? That collision – and what it means for where our industry builds next – is next month’s column.
Until then, walk your own floor with fresh eyes. Between those inconvenient columns sits the most undervalued asset in your network – and the brownfield decade belongs to the operators who see it first.
Attend now 