Listen to the AI Narrated commentary overview of the post:
One year ago, we published our first Rfx as a signal and an invitation as a way to convert “interesting futures” into actionable filters.
That 2025 Rfx didn’t just serve as a write-up, we were able to convert it into workflow. It directly shaped four new investments into our portfolio for 2025 and it helped catalyze the incubation of two new initiatives in our growing emrgnce ecosystem.
If 2025 taught us anything, it’s this: we’re not living through a “shift.” We’re living through a compression of timelines: climate volatility, institutional brittleness and AI-driven capability jumps all arriving faster than our legacy decision systems can metabolize.
The future isn’t coming. It’s already here… just unevenly distributed, violently compounding and poorly governed.
Re-introducing Rfx: Request for X
As a reminder, Rfx is not our “predictions for the year” list. It’s a standing request: for startups, systems and strategies that deserve deeper exploration, especially the ones that look strange until they become obvious.
Think of it as a beacon, not to summon hype, but to attract builders, operators and aligned capital partners toward the same handful of leverage points.
Agency Under Compression
A Note: There is an underlying theme that connects 2026’s Rfx I’m about to share, a simple lens: when the old pathways to stability stop feeling real, people (and systems) reach for short-horizon agency, and the volatility that follows becomes the upstream context for everything else.
I still remember when fintech felt like adult infrastructure. Mint.com nudged you when you overspent. Wealthfront made it easy to route a slice of every paycheck into the slow, boring magic of compounding. The promise wasn’t adrenaline, it was stability & growth.
But the macro beneath it has shifted and the product layer evolved with the incentives.
When housing becomes a moving target for many, wages start chasing inflation and “the career ladder” feels like a museum exhibit, people don’t stop wanting agency… they start renting it in smaller, faster doses.
So the interfaces optimized for hope: tiny decisions, instant feedback, lottery-shaped payoffs. Robinhood didn’t invent that impulse, but it perfected the aesthetic: the Robinhood-ification of finance, where complexity is packaged as empowerment and “swinging for the fences” becomes a daily ritual.
Zoom out and you see the same pattern everywhere. Same-day (0DTE) options compress time horizons into hours. Sports betting turns live games into a stream of micro-wagers. Prediction markets move from niche to mainstream posture — Polymarket positioning for a regulated U.S. return, Kalshi expanding “event contracts,” and even mainstream brokers experimenting at the edges of this category.
I won’t opine on the morality on this casino-ification of fintech. For me this transcends the moral debate, but appears as a systems signal. When a society can’t offer a believable path to compounding stability, it manufactures volatility-as-a-service. The casino isn’t the disease; it’s a pressure valve. And pressure valves don’t reduce pressure, they redirect it.
Eventually, the demand for agency leaves the screen and shows up in the street. Iran feels like a live boundary condition in plain sight: a collapsing currency, persistent inflation and “modern life” getting harder to finance, intersecting with energy constraints and water stress, producing strikes and protests that have forced the question of legitimacy back onto the table these past three days.
This is the reality we’re exploring in Rfx 2026: human-made brittleness (affordability, debt, institutional fragility) increasingly overlapping with physics (heat, water, energy, food shocks). The second- and third-order effects are the real story, where “economic stress” becomes “political rupture,” and where policy whiplash becomes its own accelerant.
With that as the operating backdrop, here are the seven Rfx themes we are looking for in 2026.
1) Causal Actors: Gravity vs Butterfly
Not all causal forces are equal and not all are required forever.
Some causes sustain a system (gravity). Others initiate a chain reaction (butterfly). Some are horizontal and continual; others are vertical and hierarchical. Too many strategies fail because they treat all causes as interchangeable.
If you misidentify the kind of cause you’re fighting, you’ll win battles and lose the system.
We’re interested in frameworks and interventions that can name:
initiating vs sustaining causes
structural constraints vs sensitive triggers
continual vs one-time causal actors
Question: What changes when we stop treating causality as a story… and start treating it as a design space?
Concept Exploration:
This is the distinction between triggers and load-bearing structure. The butterfly is real, small perturbations can reorder trajectories in complex systems but “sensitivity” is not the same thing as “importance.” Gravity is the deeper truth: enduring constraints, incentive gradients, and structural causes that keep the system behaving like itself even after the headlines change. The practical reference frame here is intervention thinking: don’t just ask “what correlates,” ask “what can be changed,” and then separate what starts a chain reaction from what sustains it. Strategy becomes less about reacting to sparks and more about redesigning what keeps catching fire.
The world doesn’t break evenly, it breaks along its load-bearing causes.
2) Causal Intelligence: Decision Traces as Infrastructure
We’re taking 21st-century problems and running them through decision systems that were never built for uncertainty.
LLMs, what Jensen Huang frames as “universal function approximators” can map inputs to outputs, but they don’t preserve the why; they risk designing the future by remixing the past. Whilst ‘cutting-edge VC’ attention is expanding LLMs to “world models” (see: the LAM Rfx we flagged in our Rfx 2025); what we see next is causal intelligence: decision systems that can explain, stress-test and update their reasoning.
Inside organizations and especially inside government, there’s a missing layer: the why behind decisions. The context. The precedent. The exceptions. The invisible architecture of judgment.
If you can’t replay the reasoning, you can’t improve the system.
A few months back, we began our first causal explorations as design partners to Anthos. As we dive deeper, we are looking for partners (operators, agencies, implementers) who can turn better causal thinking into operational reality: workflow changes, procurement designs, feedback loops and “learning systems” that don’t collapse into theater the moment uncertainty arrives.
Question: What becomes possible when institutions can remember causality, not just outcomes?
Concept Exploration:
A lot of institutional failure is epistemic, not moral: we demand linear plans in nonlinear worlds and we reward legibility over learning. The classic warning is the “legibility trap”, simplifying complex systems into neat categories until you erase the ecology that made them stable in the first place (the forestry parable is the clean metaphor everyone remembers). The modern version shows up in delivery: policies that look correct on paper but collapse in implementation because feedback loops are weak, incentives are miswired and uncertainty is treated like a defect rather than a condition. The most compelling examples of causal intelligence aren’t dashboards, they’re decision trails: capturing context, exceptions and precedent so institutions can learn without pretending certainty.
A society that can’t remember why becomes a society that can’t learn how.
3) Tipping Realities: Mapping the Cascades, Not the Headlines
We’re interested in systems that model non-linearity as a first-class citizen: feedback loops, compounding risks, correlation breakdowns, and brittle equilibria.
Not everything needs to be “predicted.” But it does need to be postured for.
The goal isn’t certainty. The goal is robustness when certainty is impossible.
This includes climate-driven supply chain cascades, food system fragility, grid instability, and social tipping dynamics that turn “economic stress” into “political rupture.”
Question: What does a tipping-aware operating system look like for a company, a city, or a portfolio?
Concept Exploration:
Tipping work gets serious when it stops being about “the one big cliff” and becomes about linked thresholds and cascades; how a shock in one domain turns into stress in another: heat → grid → water → food → prices → unrest, or flood → logistics → insurance retreat → municipal credit stress. The most useful mental models here look less like forecasts and more like failure choreography: what breaks first, what breaks second, and what containment looks like when multiple systems slip at once. In other words: the real value isn’t predicting the domino, it’s seeing the table it sits on.
The cliff rarely announces itself. It just becomes the new baseline.
4) Emerging Markets as Early Signals (Not Afterthoughts)
Emerging markets aren’t “later.” They’re first — first to feel extremes, first to confront broken infrastructure, first to invent under constraint. They are where solutions get stress-tested against reality, not theory.
The edge isn’t a dead end. It’s the proving ground.
We’re looking for businesses that can win on unit economics inside volatility: intermittent power, high heat, supply constraints, informal housing, climate migration, and policy uncertainty.
Question: What products become obvious when you design for instability as the default?
Concept Exploration:
Emerging markets are where the future shows up first because constraints arrive earlier: extreme heat that reshapes productivity, grid volatility that reshapes product design, supply fragility that reshapes unit economics, and policy uncertainty that reshapes distribution. The best reference points here aren’t “impact stories”, they’re constraint stories: humid heat thresholds (where cooling becomes survival, not comfort), internal displacement dynamics (where migration is not a surprise but a downstream physics-and-economics outcome), and informal infrastructure realities (where solutions win only if they are rugged, cheap to maintain, and culturally adoptable). If a solution thrives there, it often becomes globally exportable because it was forged in real conditions, not ideal ones.
The frontier is where the future stops being theoretical.
5) The Doomsday Belt: Bangladesh, Jakarta, New York (and the Next 20 Cities)
Some geographies are emerging as the world’s stress concentrators: where sea level, subsidence, heat, migration pressure, and institutional capacity collide. We’re especially interested in solutions that work in places where failure has nowhere to go.
When resilience fails in a megacity, it doesn’t fail locally, it exports instability.
This theme includes coastal defense and retreat logistics, salinity and water security, heat survivability, insurance and liability redesign, and any infrastructure that turns “inevitable stress” into “manageable strain.”
Question: What gets built when the customer is not a consumer but a city that can’t afford collapse?
Concept Exploration:
This “belt” becomes legible when you stack three kinds of maps:
(1) where water rises,
(2) where land sinks, and
(3) where institutions thin.
Bangladesh is the canonical story of slow violence: salinity intrusion turning fresh water, soils, and livelihoods into a long negotiation with the sea.
Jakarta is the stark reminder that sometimes the ground is the crisis: subsidence turning flooding into a structural condition rather than an emergency event.
New York is the mirror: wealth doesn’t delete exposure, it just delays acceptance, and once coastal flooding becomes “routine,” the real drama is insurance, liability, and who gets priced out of safety.
Some cities are where the planet collects its debts.
6) Doomsday Interventions: Buying Time Where Time Collapses
There are domains where “adaptation” is too polite a word. Some risks are not gradual, they’re phase changes. We’re interested in interventions that reduce the probability, speed, or blast radius of catastrophic outcomes, especially around the cryosphere and sea-level-linked cascades.
Doomsday isn’t a date on a calendar (unless referring to December 18, 2026). It’s a curve that quietly steepens until it becomes a cliff.
What we’re looking for: pragmatic, testable approaches that acknowledge real constraints (energy flows, boundary conditions, governance, maintenance) while still bending the trajectory.
Question: What does it look like to build a credible “time-buying” intervention — one that’s modular, measurable, and politically survivable?
Concept Exploration:
The “Doomsday Glacier” framing (Thwaites) is useful not because it’s cinematic, but because it forces a mechanical mindset: grounding lines, warm-water intrusion, buttressing loss, fracture dynamics: the kinds of failure pathways that don’t negotiate. From there, sea-level rise becomes less a number and more a systems problem: timelines, uncertainty bands and the gap between “planning-grade expectations” and “politically speakable truths.” The most illuminating work in this space tends to pair physics with legitimacy: what it takes for an intervention to be measurable, governable and deployable without becoming geopolitical theater.
When the timeline collapses, time becomes the scarcest asset on Earth.
7) C2B: Consumer-to-Business — Turning “Free Labor” into Public Good
The modern economy already runs on “free labor”: our clicks, our preferences, our data exhaust, our micro-decisions: all translated into training signal and monetized upstream.
C2B (consumer-to-business) asks: can we build systems where that surplus is intentionally harvested and routed toward communal resilience?
If value can be extracted from participation, then participation can fund resilience.
We’re exploring models where the value extracted (training data, behavioral signal, local intelligence) exceeds the cost of providing value — and the spread becomes a new kind of civic financing layer.
This is where the IKEA effect matters: people value what they help build. The participation doesn’t just produce signal, it produces ownership, affinity, and continuity (the psychological glue most “community” projects fake with branding).
Question: Can we decouple “cost of engagement” from “value created,” and redirect the surplus toward shared outcomes?
Concept Exploration:
Modern platforms already run a quiet arbitrage: millions of micro-actions (attention, preferences, behavior) become training signal and market power upstream. C2B is simply is asking whether we can route some of that surplus into resilience without turning participation into extraction. The investable leap is designing consent-first systems where participation produces business value and a community dividend… a flywheel, not a guilt trip.
If the economy already taxes attention, we should at least decide where the revenue goes.
A Call to Action
If you’re building something that aligns with any of the seven themes or you’re sitting on a weird, high-conviction insight that deserves to be pressure-tested, we want to hear from you.
Rfx is how we keep our foresight honest: make it legible, make it actionable, make it accountable to outcomes.
De-sci but make it civic infrastructure.