🤖 This Is the Week Robots Learned Touch Plus Two More Signals You Must Watch
Three early deeptech breakthroughs from the 100+ signals DeepRadar scans each week.
DeepRadar Signals | This Week’s Top 3 Breakthroughs
📅 Tuesday, November 18, 2025
Your 3-minute briefing to the future of deeptech.
My Note | The Radar That Never Sleeps
Last week I shared the first public presentation of DeepRadar™, the AI engine I’ve been building quietly for a long time and it made me reflect on why I created it in the first place.
Honestly, I just wanted a simple way to spot the important deeptech shifts early… without going through endless PDFs, hype posts, or vague headlines pretending to be insights.
So I built a system that could help me (and you) stay ahead.
DeepRadar™ is the AI-powered early-warning engine I’ve been training on patents, lab papers, hidden experiments, stealth startup moves, and research signals most people never see.
It’s not a chatbot.
It’s not a newsfeed.
It’s my quiet researcher that never sleeps, scanning the weird corners of science and engineering where the future usually shows up first.
Its job is to surface the small things that later turn into big things.
My job is to take those raw signals and turn them into insights you can actually use.
Over the last two weeks, DeepRadar™ processed 140+ early signals.
Three of them stood out immediately and not because they were loud, but because they revealed something deeper:
Technology is gaining capabilities that once belonged only to humans.
Touch.
Adaptation.
Intelligence built directly into materials.
Here are the signals that matter this week.
I’m also opening Alpha Signals for the first time this week. It’s my private layer of DeepRadar where I share one early signal every Friday including the FAST Score (my framework for evaluating a signal’s future weight, investment potential, and real-world scope) and the thinking behind how I read these breakthroughs.
If you want a clearer view of what’s coming, you’re welcome to join me there.
Signal #1 | Robots Are Learning Human Touch
What’s happening
A research team in China engineered a three-layer humanoid finger made of rigid “bones,” flexible joints, and a soft skin layer (a biomechanical stack very close to the human hand). It can pick up a strawberry without bruising it yet also stabilize a heavy object with real strength. This is the closest we’ve ever been to solving robotic dexterity.
Why it matters
Movement is solved. Vision is solved.
But touch has always been the missing piece.
Human-level dexterity unlocks real robotic work: kitchens, clinics, eldercare, warehouses, manufacturing floors — anywhere precision and gentleness matter.
This is the moment robots stop acting like machines and start operating like teammates.
For investors
• Bet on the components controlling dexterity: tactile sensors, compliant materials, robotic hand modules, micro-actuator systems. These will underpin every humanoid and service robot coming to market.
• Look for companies building precision manipulation tech. If their hardware improves accuracy or reduces breakage, they’ll be indispensable to every robotics OEM.
For founders
• Build robot upgrades, not full robots. Everyone is racing to build humanoids — almost no one is building the hands, sensors, and compliant materials those robots rely on.
• Position around precision, zero-damage handling, and consistent performance. These three qualities are what operators in logistics, healthcare, and hospitality will pay for first.
For builders
• Deepen your skills in soft robotics, compliant mechanisms, tactile sensing, and manipulation. These will become high-demand competencies as robots move into real-world environments.
• Learn how to retrofit biomimetic designs into current robotic architectures. Engineers who can do this will be rare and valuable.
Future Prediction
By 2027, robots that can handle objects as gently and precisely as humans will start showing up in everyday workflows. Once that barrier drops, it becomes obvious where robots make sense, and adoption moves much faster than most people expect.
If I had to bet…
If I had to bet, by 2027 we’ll see robots with human-level touch working in real environments, not just pilot projects. Places like kitchens, clinics, warehouses, and eldercare settings will start using machines that can handle delicate or irregular objects without damaging them.
And I honestly believe the biggest winners won’t be the companies building full humanoids. It’ll be the teams building the hands, the soft materials, the tactile skins, and the tiny actuators. Once robots can interact with the physical world with the same nuance we do, a lot of the hesitation around adopting them disappears.
Signal #2 | Biology Just Rewired Carbon Capture
What’s happening
Danish researchers added the natural enzyme carbonic anhydrase to a standard direct-air-capture column. The enzyme accelerates CO₂ conversion 3× faster without requiring any extra energy. Same machines, dramatically more carbon captured.
Why it matters
Carbon removal has always struggled with slow kinetics and high energy costs.
Enzymes flip the equation.
Biology is stepping in as a force multiplier for climate tech: increasing efficiency without increasing complexity or operating cost.
For investors
• Watch enzyme manufacturers, biocatalysis platforms, and hybrid sorbent–enzyme companies. These will define the margins of the next wave of climate tech.
• Bet on teams that combine strong IP with drop-in industrial integration. Adoption moves fast when OPEX drops.
For founders
• Lead with cost/ton reduction, not the idealism of climate tech. Industrial buyers move on economics, not emotion.
• Build solutions that fit directly into existing DAC infrastructure. No new hardware → faster deployment → faster revenue.
For builders
• Learn enzyme kinetics, catalytic modeling, bioprocess optimization, and how to integrate biology with industrial columns.
• Develop intuition for hybrid climate-bio systems, this is where the next wave of innovation will happen.
Future Prediction
By 2028, enzyme-enhanced DAC will likely outperform traditional sorbent systems on both efficiency and cost. When you can get more output from the same hardware, operators don’t need much convincing. The companies owning enzyme IP and hybrid sorbent technologies will shape the economics of the industry.
If I had to bet…
If I had to bet, by 2028 enzyme-enhanced carbon capture will outperform traditional systems on both cost and speed. When operators can pull in more CO₂ using the same equipment and the same energy budget, the shift happens fast. That’s usually how industrial adoption works.
And I think the real value will flow to the companies building the inputs: enzymes, catalytic systems, hybrid sorbents. These become the levers that improve performance without requiring anyone to rebuild entire plants. Biology stepping into climate tech feels like a genuine turning point.
Signal #3 | Solar Is Becoming a Material, Not a Panel
What’s happening
The NUS/SERIS team built a stable 27.1% triple-junction perovskite–silicon solar cell. By stacking three specialized photovoltaic layers and solving a long-standing voltage mismatch problem, they achieved higher efficiency using far less surface area.
Why it matters
Solar just broke out of its efficiency plateau.
This unlocks high-density applications: aerospace, drones, EV skins, military systems, and dense urban environments.
We’re shifting from “solar mounted on surfaces” to “solar built into materials.”
For investors
• Look closely at perovskite deposition equipment, precursor chemistry IP, and multi-junction fabrication tools. These are the chokepoints for global scaling.
• Early winners will be the companies integrating high-density PV into drones, satellites, EVs, and autonomous systems where weight and surface area matter.
For founders
• Build for sectors where every watt is strategic: aerospace, robotics, EVs, remote sensors, defense systems.
• Think beyond panels. The future is solar-integrated materials and structures.
For builders
• Learn perovskite chemistry, interface engineering, thin-film fabrication, encapsulation design.
• Focus on bridging the gap between lab-scale efficiency and manufacturable products, extremely valuable expertise.
Future Prediction
And by 2029, I expect multi-junction solar to become the standard in sectors where weight and surface area matter. Solar won’t be something mounted on top of structures anymore. It becomes something built into them, which is a very different way of thinking about energy.
If I had to bet…
If I had to bet, by 2029 multi-junction solar will become the go-to choice anywhere power density matters. Aerospace, drones, EV surfaces, defense hardware, compact city systems (all of these benefit from getting more energy out of less material).
The real breakthroughs will come from teams who solve perovskite deposition, interface stability, and lightweight integration. Solar is starting to shift from something we attach onto structures into something that becomes part of the material itself, and that change opens a completely different design space.
Prediction Drop
By 2027, I expect robots with human-level touch to start taking over the kinds of tasks that have always been too fragile for machines. Think kitchens, clinics, warehouses, and eldercare environments. Once robots can handle the physical world with the same nuance we do, a lot of the friction around adoption disappears.
By 2028, I think enzyme-enabled DAC will clearly outperform traditional sorbent systems on both speed and cost. When you can capture more CO₂ with the same equipment and the same energy use, the industry tends to shift quickly. My guess is this becomes the default approach faster than people expect.
By 2029, I see triple-junction solar becoming the standard in sectors where energy density matters most. Aerospace, drones, EV surfaces, and compact urban systems all benefit from getting more power out of less material. Solar stops being something added onto structures and becomes part of the structure itself.
❤️ Why I Care?
I share these signals because I believe the right people: people who build, invest, and pay attention early, can shape the future if they see the patterns before everyone else.
Most of the real breakthroughs don’t show up in headlines.
They hide in the weird, boring and quiet places: the PDFs nobody reads, the grants nobody follows, the prototypes nobody talks about yet.
That’s why I built DeepRadar™.
To make sure those signals don’t stay hidden.
You don’t need more noise.
You need clarity.
And you need it early.
These are the signals that matter this week. Where they go next is where the leverage really starts to show.
Stay early. Stay sharp. I’ll keep bringing you the signals that matter.
—
Eden Djanashvili
Author, Invest Deeptech
See you next Tuesday!
DeepRadar™ is already scanning for what’s coming next.
🛑 For information only. Not financial advice.