Deep Tech Monthly in Review - November 2025
Field notes from last month in Deep Tech startups and private markets — a strategic recap for Builders and Backers.
Dear Builders and Backers,
November felt less like a “news month” and more like someone quietly redrawing the blueprint beneath Deep Tech.
The physics barely moved; the assumptions did.
Power stopped being a backdrop. Magnets, mines, polymers, and molten salts stopped being niche. Capex, grids, exits, and rate bases stepped out of the appendix and into the first slide.
On the energy side, advanced fission, geothermal, sodium-ion storage, and metal-fuel systems all converged on the same question:
Where do the firm megawatts actually come from, and who underwrites them for decades, not quarters?
Microreactors left the whitepaper and walked into mines, processing plants, and data campuses. At sea, ammonia and methanol started to look less like “future fuels” and more like live design choices for shipping.
The common thread: power and heat as strategic inputs, not ESG talking points.
In materials and mid-stream processes, the month pushed hard on control. Quantum-grade diamond foundries, sovereign magnet capacity, bio-based high-heat polymers, molten-salt rare-earth recovery, biomining, and photochemical routes all landed in one place: giving operators levers over feedstocks, process risk, and geopolitics without rebuilding the entire plant.
Materials stopped being “inputs” in a spreadsheet and started behaving like policy tools and industrial moats.
Capital behaved accordingly.
The biggest checks didn’t chase generic “AI” or “climate” slogans; they went into grids, inference infrastructure, magnet and polymer lines, transmission, programmable biology, molten-salt and biomining platforms, industrial robots, and very specific ways to turn FOAK plants into copy-exact assets.
Hardware was not dead; it was simply being priced as infrastructure again—on the condition that it could be replicated, serviced, and absorbed by real balance sheets.
This November package is our attempt to stitch all of that into a single industrial playbook.
You’ll find briefings that treat fission, grids, magnets, maritime fuels, biomining, sodium-ion, and humanoids as one connected risk surface; capital movements that show how investors are quietly re-rating power, materials, and process tech; educational episodes that translate utility, grid, and automotive logic into startup moves; and slow-burn analysis on capex, photochemistry, and how exits in magnets and rare-earths actually got engineered.
Our goal is simple: help you stop treating power, materials, and capex as assumptions, and start treating them as design inputs.
If you think in megawatts, tons, lines, and rate bases—not just models and slides—November offers plenty of ways to turn working physics and biology into businesses the real economy cannot ignore.
Enjoy the read,
— Giulia & Nicola
Before We Dive In: Big News!
📙 Announcing Our Book!
A Practical Handbook for Building and Backing Companies That Actually Get Funded.
Hitting 100 episodes of Deep Tech Catalyst has revealed patterns that are impossible to ignore.
The startup books we all know were born in a software-dominated world: cheap experiments, fast iterations, low capital intensity, clean feedback loops.
Deep Tech lives by a different physics. It doesn’t care about “move fast” slogans: it’s fragmented, capital-intensive, and the risk stack runs from the lab bench all the way to policy and infrastructure.
Through Deep Tech Catalyst—by sitting down with some of the sharpest investors and industry operators in the world—we stopped looking for universal answers in Deep Tech and focused on the questions they kept repeating.
From quantum to materials to biotech, these worlds look unrelated—but zoom in on any cap table, IC memo, or founder call, and the same fundamental questions keep surfacing.
We’ve turned those patterns into a practical handbook for people who build, fund, and partner with Deep Tech ventures.
It will be released soon…
Want it before everyone else?
November at The Scenarionist: fission, magnets, biomining, molten salts, grids, and exits — all connected into a single industrial review.
DEEP TECH BRIEFING
What Actually Moved in the Deep Tech World
4 weekly briefings for the moments when someone around the table asks, “What really changed this month?” and you need a single coherent narrative.
Advanced Fission, Geothermal, Sodium-ion
If you sign off on megawatts, read this first.
This Briefing will force you to stop treating baseload and process heat as “climate topics” and start seeing them as multi-decade infrastructure on the same footing as fabs and ports.Quantum Materials and Sovereign Magnets
For when substrates and rare earths show up in your IC memo.
This edition gives you language to explain why a quantum-diamond foundry and onshored magnet processing are not esoteric science projects, but policy tools and control points for entire downstream industries.AI Freight, Maritime Fuels, Microreactors for Mines, and Data
Mandatory if you touch logistics, energy or data campuses.
Here, freight platforms, ammonia/methanol at sea, synbio, and microreactors are treated as one continuous risk surface—showing how contracts for fuel, power, and capacity are starting to blur.Magnets, Metals, Reshoring, “volcano homes”
Read this before you say “reindustrialization” again.
This Briefing quietly redraws who can actually execute on reshoring once you factor in rare earths, agricultural robotics, geothermal baseload and the physical limits of local grids.
DEEP TECH CAPITAL MOVEMENTS
What Capital Believes Before the Narrative Catches Up
A weekly X-ray of where serious money is really going in Deep Tech, and what that implies about how the next cycle will be priced.
Inference Infra & Materials Substitution
The issue to read before your next AI infra debate.
This edition shows why compatibility and end-of-life are starting to earn a premium: inference infrastructure, drop-in materials, and recyclability funded together as one thesis, not three buzzwords.Nine-figure Space & Power-line Rounds
Proof that “hardware is dead” is intellectually lazy.
This piece tracks nine-figure checks into space manufacturing and high-eight-figure rounds into transmission tech, and makes clear under which industrial logics hard assets are very much bankable again.Electrons for GPUs, Biomining, Metal Fuels
Your cheat sheet on how AI now runs through mines and substations.
From power ICs to metal-fuel systems and microbial biomining, this is the week that forces you to stop thinking of “AI infrastructure” as something that lives only in the cloud.Programmable Bio, Humanoid Brains, Molten-salt Rare Earths
For when you want to see what a real “platform bet” looks like.
Growth capital into programmable biology, robotics brains, molten-salt rare-earth recovery, and high-heat polymers—useful shorthand for which deep platforms investors quietly expect to compound over the next decade.
DEEP TECH CATALYST
How to Actually Build and Back Deep Tech Companies
Operator-grade conversations and playbooks for the moments when physics is not the problem anymore, but contracts, capex, and integration are.
Venturing into Grid Technology: VC Insights for Deep Tech Startups
The episode that fixes your pitch to utilities.
If you’re still pitching grid technology like SaaS, this will recalibrate you: how to be treated as capex-grade infrastructure, how rate bases work, and why behind-the-meter power is often the real entry point.Optical Interconnect: From Copper to Light
Required listening before you commit to photonics or AI networking.
Sets a realistic bar on cost, yield, packaging, and hyperscaler purchasing behaviour—so you don’t fall in love with throughput slides that will never clear procurement.Scaling Deep Tech in Automotive: What CVCs Look for
For anyone who still thinks “auto is too slow.”
A candid map of how OEMs and their CVCs actually take risk on hardware—from first pilot to appearing as an option on a spec sheet—and what “good” looks like along that path.10 Questions to Stress-test Any Deep Tech Startup
The filter to keep open on your second screen.
A distilled checklist from 100+ conversations; useful both for founders and investors who want to avoid beautifully engineered companies that will never survive the regulator, grid operator, or OEM reality.
RUMORS & ANALYSIS
Capex, Patterns, Exits: The Long Game
Slow thinking on the parts of Deep Tech that don’t fit into a weekly news cycle: industrial capex, process technology, and how exits are really engineered.
Capex That Compounds: Turning Industrial Spending into a Growth Engine
Non-optional if you are anywhere near a FOAK plant.
A framework for turning first-of-a-kind industrial assets into repeatable, lender-friendly units—so capex becomes a flywheel instead of a very expensive science experiment.Light-Driven Chemicals and the Quiet Revolution of Photon-Chemical Manufacturing
Read this before you dismiss photochemistry as “still early.”
Explores what happens when photons become a first-class process input in the largest industrial energy consumer on the planet, and what that does to energy use, feedstocks, and margins.Inside Deep Tech Exits, Engineered
The piece to read before you design your next cap table or JV.
A candid look at how advanced materials and rare-earth recycling exists actually happened—who bought, to close which vulnerabilities, and what Builders and Backers could have structured differently from day zero.
November: The Big Reality Check
1. Power and Heat: Megawatts as the Gating Factor
The clearest shift this month is that power and heat stopped being “context” and became the object of competition.
An advanced fission player listing on Nasdaq [1] put a public marker on decades of high-temperature, long-duration output. Behind the ticker, the real story is boardrooms beginning to price time and reliability in the same sentence as megawatts: baseload and process heat as strategic inputs for data centers, refineries, and heavy industry.
In the same window, modular fission, geothermal, and sodium-ion storage show up not as “alternatives” but as different ways to achieve the same thing: firm, local, politically durable energy that can sit under industrial assets instead of next to them. [1]
Later in the month, microreactors move from theory toward specific use cases: mines, processing plants, and data campuses in regions where the grid is thin, and diesel is a recurring tax on every ton moved or inference run [3]. At sea, ammonia and methanol trials demonstrate that shipping is treating fuel choice as a live design parameter, not a policy talking point [3].
By the time we get to late November, the constraint is spelled out in operational terms: data-center pipelines and industrial expansions are running into interconnection queues, constrained substations, and transformer bottlenecks, not just chip lead times [4]. Humanoid robots quietly complete months of work on an auto line, but the more important bottleneck sits outside the factory fence: where to source firm megawatts in the right postal codes—and how OEMs and their CVCs actually sequence risk from first pilot to showing up as an option on a spec sheet [11].
On the capital side, the same story runs through the deal tape. One week of Capital Movements is dominated by “electrons for GPUs”: high-density 800V power ICs for AI data centers, metal-fuel power systems built around recycled metals, and multi-hundred-million-dollar commitments to AI-aligned, renewable-powered data campuses [7]. Another week sends high eight-figure checks to transmission technologies that simply build and string lines faster [6].
From the operator side, the GridTech conversation on Deep Tech Catalyst makes the incentives explicit: utilities want solutions that can be capitalized and put into the rate base, while hyperscalers will move faster on behind-the-meter generation that shrinks their visible load [9]. Two hardware plays stand out there: tools that keep existing lines online (fault and ignition detection) and local generation that sits inside the customer fence.
For Builders and Backers: November’s power narrative is simple and uncomfortable. If your thesis relies on high-uptime compute or process heat, “assume power is there” is no longer a valid line item. You either help deliver it, finance it, or work explicitly around its constraints.
2. Materials and Semiconductors: From Inputs to Levers
The second line running through November is control over materials and mid-stream processes.
A quantum-grade diamond platform moving into a foundry model is a good example [2]. The physics isn’t new; the manufacturing posture is.
The focus shifts to:
reproducible recipes,
within-lot uniformity,
and process control robust enough for certification.
That’s what turns quantum materials from lab curiosities into substrates for real sensors, metrology tools, and compact accelerators that industrial buyers can specify.
In parallel, magnets and metals are pulled into the same strategic frame [4]. The reshoring analysis looks at rare-earth chains and mid-stream copper steps not as abstractions but as:
motors and generators,
drives and transformers,
EVs, turbines, and defense programs that sit downstream.
Magnets, in that reading, are not cheap components; they’re long-life infrastructure. The “moat” in that title is literal.
On the venture side, several deals push in the same direction. One week of capital flows features a French company designing drop-in alternative materials for constrained supply chains: substitutes that run on existing equipment so you can re-engineer risk without re-engineering the plant. The same issue tracks seaweed-based, marine-biodegradable plastics that behave like incumbents on current extrusion and molding lines [5].
Later, a Dutch team expands funding for bio-based high-performance polymers aimed at electronics, data centers, and aerospace. High-heat polyamides, polyimides, and industrial 3D-printing resins show up as another place where “materials” are really locational and geopolitical risk in disguise [8].
Move one layer down, and you hit the wiring of the system [10]. The optical-interconnect deep dive focuses on what happens when bits move as light instead of electrons:
Physics gives you more bandwidth and lower energy per bit,
Economics still favors copper until optics clear cost, yield, and packaging hurdles.
The conversation is very specific: it’s not enough to hit headline speeds; you have to clear energy-per-bit and bandwidth-density thresholds at a cost structure hyperscalers will actually deploy. The battle moves into packaging, thermals, and manufacturability.
Finally, there’s the question of how we run chemistry itself. The photon-chemical manufacturing piece starts with a sober view of chemicals as they exist today: the largest industrial energy consumer, a major direct emitter, and an asset base built on cheap fossil heat and feedstocks [14].
Into that world come:
LED-driven and laser-driven photocatalysis platforms,
continuous-flow photoreactors in GMP and specialty plants,
hybrid photo-thermal routes for fuels,
lignin-to-aromatics schemes,
and modular photo-reactors that generate hazardous intermediates on-site and on-demand.
For builders, the common denominator is that materials and processes are no longer “inputs”. They’re levers. If you can give an operator a way to switch feedstocks or process knobs without rebuilding everything, you’re not selling a product; you’re selling degrees of freedom.
3. November’s Capital Signals: Grids, Materials, Mining, Chemical Processes, and Industrial Robots.
Tracking November’s capital flows straight through, the pattern is pretty clear: money is chasing concrete ways to move electrons and atoms more reliably. [5,6,7,8].
Four signals stand out:
On the grid side, the checks go to things that either shorten the distance between generation and load or make that link less flaky. Faster transmission build, modular substations, and high-density power electronics all show up as tools to turn interconnection queues into firm, schedulable megawatts for data centers and industrial sites. Capital isn’t trying to “pick a fuel”; it’s trying to buy control over when and where power is actually available.
In materials, investors lean toward drop-in substitutions rather than full process rewrites. Alternative alloys, bio-based polymers, and compatible plastics are pitched as ways to de-risk existing plants: you change the feedstock, but keep the tooling, the lines, and the certifications. The real exposure is to supply-chain shocks and regulatory pressure, not to whether an entirely new unit operation ever escapes the pilot stage.
Mining and chemical processes follow the same playbook. The money goes into sensing, automation, and modular units that bolt onto current pits, concentrators, and reactors: autonomy layers that push more tons with fewer incidents, compact flow, and photochemical reactors that handle hazardous steps on-site instead of in faraway specialty facilities. Instead of betting on “new mine” or “new complex” stories, investors are buying incremental volume, lower diesel burn, and lower specific emissions on assets that already exist.
The robotics piece makes the operator logic very explicit. Humanoids and other industrial robots are packaged around uptime and availability—service contracts, per-task pricing, integration into line-level metrics—so automation shows up in opex and productivity KPIs instead of one-off capex fights. For operators, that makes robots another knob on throughput and quality; for backers, it turns hardware into usage-linked revenue that sits directly on top of real factories and warehouses.
Takeaways for Builders and Backers
Stepping back from the individual stories, a few working conclusions for November:
Power is now a primary design constraint.
Fission on public markets, microreactors for mines and data, metal-fuel systems and high-voltage ICs, grid hardware and transmission build-out —all point to the same reality: megawatts in the right locations are the scarcest resources in many Deep Tech theses.Materials and mid-stream processes are becoming levers of strategy.
Quantum-grade diamond foundries, sovereign magnet supply, drop-in alternative materials, high-heat bio-polymers, and light-driven chemical routes are being treated as ways to control exposure to geopolitics, carbon policy, and margin, not just as technical curiosities.Mining and circularity are being industrialized.
Biomining, molten-salt rare-earth and lithium recovery, and engineered exits in magnets and separation all treat mining and recycling as infrastructure plays. The winners are likely those who design for integration into existing flow sheets and for eventual absorption by industrial balance sheets.Capex and customer behavior are as strategic as the technology.
The capex playbook, grid and automotive buyer insights, and the 10-question framework converge on one point: deep-tech companies that align their engineering, unit economics, and capital stack with how utilities, OEMs, and chemical majors actually buy will have a structural advantage.Capital is present, but more discriminating.
The November Capital Movements series shows a market willing to finance hardware-heavy plays in energy, materials, and infrastructure—provided the story is grounded: concrete customers, credible replication plans, and exits that make sense in industrial terms.
Final Thoughts
November was a quiet reality check: the physics was never really in doubt, but the power lines, magnets, mines, and capex plans finally stepped onto the main stage.
Across the month, one pattern kept repeating: nothing moves without firm megawatts, reliable materials, and a buyer who can carry real infrastructure on their balance sheet.
Fission listings, sodium-ion banks, molten-salt recovery, biomining, humanoids on auto lines—they all point to the same question behind the scenes: who is actually underwriting the power, the inputs, and the first plant?
For Builders and Backers, that makes the work both harder and simpler.
Harder, because “we’ll find the power later” or “someone will want this material” no longer passes a straight-face test.
Simpler, because the brief is clearer: design companies that live inside grids, mines, ports, factories, and chemical plants as they really are—not as we wish they were.
That’s the thread we’ll keep pulling on.
We’ll stay close to the places where Deep Tech becomes concrete—substations, magnet lines, bioreactors, reactors, yards full of transformers—and keep turning them into practical tools: questions for builders and backers [12], patterns for cap tables, and checklists for FOAK plants that are meant to be copied, not celebrated once and written off [13].
If you’re willing to think in megawatts, tons, rate bases, and exit paths that end on industrial balance sheets, the opportunity set in front of you is larger, not smaller.
The frontier is still there—it just runs through mines, grids, and warehouses as much as labs.
See you soon!
— Giulia Spano, PhD & Nicola Marchese, MD
“Anything that won’t sell, I don’t want to invent. Its sale is proof of utility, and utility is success.” — Thomas Edison