Koniku.

Here is a fact that should bother anyone who has stood in an airport security line: the best explosive detector in the building is usually a dog. We have spent billions of dollars on cameras that see, microphones that hear, and processors that do a trillion sums a second, and yet when the task is "figure out what molecules are floating in this air," the state of the art remains a Labrador on a leash. Koniku's entire premise is that this is embarrassing, and fixable, and that the way to fix it is not to build a better machine but to build a partly living one.

Koniku - the name comes from Yoruba and means, roughly, "immortal" - was founded in 2015 by Oshiorenoya "Osh" Agabi along with co-founders Christopher McAndrew and Christopher Hedvall. The company's product category is a word most people have never had to say out loud: wetware. Not hardware. Not software. Wetware. The idea is to take biological neurons, engineer them to carry specific smell receptors, keep them alive on a chip, and read the electrical chatter they produce when a target molecule shows up. The neurons do the sensing. The silicon does the interpreting. The whole thing, in theory, smells.

If that sounds like a lot of caveats, it is, and the caveats are the point. Building with living cells means your product has to eat, has to be kept alive, and has to be manufactured in a way that no chip fab was designed for. An MIT chemistry professor, surveying the field in 2020, said making this work at scale would take "some technical miracle." This is the kind of quote that either kills a company's fundraising or becomes the epigraph of its eventual case study, and it is not yet clear which. Koniku's response to the skepticism has been, essentially, to keep building and let the demos do the arguing.

The flagship device is the Koniku Kore, usually shortened to Konikore. It is an electronic olfaction system: you program neurons with receptors that bind to a particular substance - benzene, say, or a marker on a cancer cell, or the vapor signature of an explosive - and when that substance arrives, the neurons fire, and the chip reads the pattern. Koniku has claimed the platform can distinguish thousands of compounds, with a headline figure of up to 4,096 smells detected simultaneously. Those numbers are the company's own, and the honest way to hold them is as engineering targets a partner would want independently verified, not as a benchmark someone else has confirmed. But the direction is coherent: match the dog, then beat the dog, then put the dog in a box you can bolt to a doorframe.

What is genuinely unusual about Koniku is how old the idea is relative to the company. The intellectual seed traces back to research at ETH Zurich around 2003 - roughly a decade before there was a company at all. Agabi himself is almost comically over-credentialed for the job: a physicist by training, he collected a PhD from ETH Zurich and a second one, in bioengineering, from Imperial College London, and picked up a public reputation as a scientist willing to say impolitic things. His view that modern science has "devolved into paper-pushing at the expense of real innovation" is the sort of line that gets you disinvited from some conferences and headlining others.

The commercial pitch is that smell is a horizontal capability - the same underlying platform points at wildly different markets depending on which molecule you tune it for. Koniku sells this as a B2B platform, and its partner list reads like a tour of "places where knowing what's in the air is worth money."

The Airbus relationship is the tell. Aerospace companies do not, as a rule, fund science projects - they fund things they think will end up on a jet bridge or a cargo screening line. Koniku's partnership with Airbus dates to 2017, was reported in the neighborhood of $8 million at the outset, and expanded in 2020 toward detecting biological hazards. By 2022, Airbus Americas had joined Koniku's integrator ecosystem, the KTIE, which is the company's attempt to turn a single clever device into a platform other people build products on top of. When an aerospace giant repeatedly re-ups on a company whose core component is alive, that is at least a signal worth noting.

The business model, then, is deep-tech hardware plus platform: pilots and partnerships with enterprise, aviation, defense and healthcare customers, rather than a gadget you buy on Amazon. It is a slow way to build a company and an appropriate one, because the alternative - shipping a living sensor to millions of consumers before you've proven it survives a Tuesday - is how deep-tech companies die.