When people hear the word technology, what do they imagine?
For many, it means artificial intelligence, humanoid robots, self-driving cars, smartphones or enormous data centers packed with expensive chips. Venture capital has poured billions into software capable of generating text, images, video and code, while technology companies race to build increasingly powerful AI models.
Yet some of the most valuable innovation may be happening somewhere considerably less glamorous.
It could be inside a party balloon.
It could be in the coating on a frying pan.
It could be a new material capable of removing dangerous chemicals from drinking water or making batteries, semiconductors and industrial manufacturing cleaner.
That is the argument behind a growing interest in what might simply be called useful technology: innovation designed not merely to impress people on a screen but to solve stubborn problems in the physical world. A recent Financial Times column highlighted this contrast, pointing to companies including Bioloon and CuspAI as examples of technology being applied to materials people actually touch, use and eventually throw away.
The idea sounds almost old-fashioned.
But perhaps that is exactly why it matters.
A Party Balloon Can Be a Serious Technology Problem
A balloon hardly looks like advanced technology.
Someone blows it up, hangs it at a birthday party and throws it away a few hours later. Yet that simplicity hides a materials-science problem that has existed for generations.
Many conventional latex balloons are made using natural rubber, which would ordinarily be capable of degrading. Manufacturers, however, need the rubber to be strong and elastic enough to function as a balloon.
That normally requires vulcanisation.
The process creates cross-links within the rubber, traditionally using sulphur and other additives. It gives the material the strength and flexibility people expect, but it can also make the resulting balloon much harder to break down naturally.
Researchers at Imperial College London have been working on an alternative.
Their Bioloon project developed a modified latex material that avoids conventional sulphur vulcanisation while retaining the performance needed from a balloon. Imperial says the resulting material can be inflated repeatedly, works with existing balloon-manufacturing approaches and breaks down much more rapidly after disposal.
More information about the research is available through Imperial College London’s Bioloon project.
What makes the development interesting is not that humanity desperately needs better party decorations.
It is what happened when someone looked at a completely ordinary object and asked whether its underlying material could be redesigned.
The Problem Started With a Bin Bag of Used Balloons
Bioloon did not begin because a major technology company decided balloons represented the next trillion-dollar market.
Charlotte Melia and Christie Molgaard were working in the events industry and regularly encountered large quantities of discarded balloons after parties. Packaging sometimes claimed that latex balloons could biodegrade relatively quickly.
They became skeptical.
After watching used balloons fail to disappear as expected, they approached researchers at Imperial College London. That curiosity eventually led to years of research involving chemical engineering and materials science.
The result is a useful lesson about innovation.
A problem does not have to be enormous individually to matter at scale.
One discarded balloon appears insignificant. Billions of single-use latex products are something else entirely.
And balloons may only be the beginning.
Researchers involved in the project have suggested that similar material technology could eventually be relevant to products including gloves and other disposable latex goods.
Suddenly, an invention that began with children’s parties starts looking like a platform for changing an entire category of materials.
That is useful technology.
Frying Pans Reveal an Even Bigger Materials Problem
The same principle can be applied to something sitting in millions of kitchens.
The nonstick frying pan is one of the most successful examples of materials engineering becoming invisible.
Someone cracks an egg into the pan, cooks it and watches it slide across the surface. They rarely think about the chemistry responsible for that convenience.
But nonstick performance has historically relied heavily on fluorinated chemistry, including materials belonging to the broad family known as PFAS.
PFAS are often called “forever chemicals” because many of them are extraordinarily persistent in the environment. They have been used far beyond cookware, appearing in applications ranging from textiles and packaging to electronics and industrial coatings.
That creates an enormous substitution problem.
A company cannot simply decide that a chemical is undesirable and remove it.
They need another material that performs the same function.
A replacement coating still needs to resist heat, repel substances, survive manufacturing and remain economically practical. A replacement used in semiconductor production may require completely different properties from one used in clothing or cookware.
Finding those alternatives can take years.
This is where another fashionable technology—artificial intelligence—could become genuinely useful.
AI Could Become More Important in the Laboratory Than in the Chat Window
CuspAI is trying to use generative AI for materials discovery rather than simply content generation.
Instead of asking an AI system to write an email or create an illustration, a scientist can define properties they need from a material and use computational models to explore structures that might deliver those properties.
The company calls the approach inverse design.
Researchers begin with the desired outcome and use AI to help search for materials that could produce it.
CuspAI has been working with Finnish chemicals company Kemira on materials capable of removing PFAS from water. According to reporting on the project, its systems explored an enormous potential chemical search space to identify promising candidates for further testing.
The technology has attracted substantial investment. CuspAI raised hundreds of millions of dollars in 2026 and reached a reported valuation of $2.6 billion, with applications extending into semiconductors, aerospace, automotive technology and environmental materials.
That does not mean an AI model can simply type out a molecule and declare the pollution problem solved.
Predicted materials still have to be synthesized, tested, manufactured, validated and made economically viable.
But AI can potentially change where researchers begin.
Instead of manually exploring a tiny fraction of an almost unimaginable chemical universe, they can use computation to narrow the search dramatically.
That is a far less visible application of AI than a chatbot.
It could also prove far more consequential.
Useful Technology Often Looks Boring Until It Works
There is a recurring problem in the way technological progress is discussed.
Attention tends to follow spectacle.
A humanoid robot walking across a stage generates millions of views. A new consumer gadget receives launch events, reviews and social-media debates. A generative AI model producing a photorealistic video can become global news overnight.
A new polymer generally does not.
Yet civilization depends heavily on materials innovations that most people never notice.
Better cement changes construction.
Better catalysts change chemical manufacturing.
Better membranes can improve water treatment.
Better battery materials can increase energy storage.
Better coatings can reduce corrosion or eliminate hazardous chemistry.
Better semiconductor materials can change how much computing performance is possible from a given amount of energy.
Imperial College London alone has spun out companies working on biodegradable dyes, recycling textile colorants, recovering high-purity vanadium from industrial waste and producing advanced two-dimensional materials for energy storage and sustainable construction.
None is likely to become a household name as quickly as the next social-media platform.
Their technologies could still quietly affect millions of lives.
Silicon Valley’s Software Obsession Has a Cost
Software became attractive to investors for understandable reasons.
It can scale quickly.
A company can create one digital product and distribute it globally without constructing a factory for every new customer. Margins can be extraordinary, and successful platforms can grow at speeds almost impossible for conventional manufacturing businesses.
Generative AI intensified that fascination.
Huge amounts of investment have flowed toward foundation-model developers and businesses built around them. The Financial Times argues that this concentration risks narrowing the technology industry’s imagination, particularly when physical-world innovation often requires slower research, laboratories, specialized equipment and long periods before revenue arrives.
A materials startup cannot necessarily release an unfinished polymer and patch it next Tuesday.
If the product goes into medical equipment, drinking-water treatment, aircraft, batteries or industrial machinery, it must work reliably in the real world.
That makes physical innovation harder.
It may also make successful breakthroughs much more defensible.
The Best AI Story May Eventually Be About Things, Not Words
None of this requires dismissing generative AI.
The more interesting possibility is that the current AI boom eventually escapes the screen.
If artificial intelligence can help scientists discover materials faster, optimize industrial processes, predict failures, design new molecules and reduce the number of laboratory experiments required to reach a useful result, then its impact becomes physical.
A person may never know AI was involved.
They simply buy a pan whose coating contains fewer problematic chemicals.
They use a medical glove made from a more sustainable material.
Their drinking water passes through a filter better able to capture persistent contaminants.
Their electric vehicle carries a battery produced with improved materials.
Their home is built using concrete with a lower environmental footprint.
That version of AI is less theatrical.
It may also be closer to what technological progress has historically looked like.
Perhaps Technology Should Be Judged by What It Fixes
A biodegradable balloon will not transform civilization on its own.
Neither will a new frying-pan coating.
That is almost beside the point.
The more useful question is whether researchers are developing tools capable of solving thousands of similarly ordinary problems.
Modern economies are filled with materials designed decades ago because they were cheap, durable or effective. Some consume too much energy. Others create waste that persists for generations. Some use chemicals researchers would now prefer to replace.
Redesigning them will not happen through one giant invention.
It will happen product by product and material by material.
That makes Bioloon an unexpectedly good symbol of what technology can be.
Someone looked at a pile of discarded party decorations and did not ask how software could make the party more engaging.
They asked why the balloons were still there.
Scientists then tried to change the chemistry.
CuspAI represents the next stage of the same philosophy: taking powerful computational technology and pointing it toward physical problems rather than keeping it trapped inside screens.
Technology does not always need to feel futuristic to be transformative.
Sometimes the most impressive invention is simply one that makes an everyday object work better without leaving such a damaging legacy behind.
Perhaps the next era of technology should therefore be judged less by how convincingly machines can imitate humans and more by how effectively they help humans redesign the physical world.
From balloons to frying pans, that could turn out to be where some of the most useful innovation has been hiding all along.
