‍

The Art of Partnerships—Combos that Compound
Alone a spark is a flicker; paired with the right tinder it becomes a torch. The first quarter of 2025 reminded us—daily, dazzlingly, dizzingly—that progress remains a profoundly social pursuit. Every advance this winter inside the Lux network carried an invisible subscript: co‑written. From the co-founders building on the frontiers of the possible to our co-investors funding the future alongside us, plus our colleagues co-leading the deals that will determine destiny, we must constantly return to that most universal of truths: that to build the future requires cooperation in the present demanding trust in the past. A slick deal, a promise betrayed, an ill-gotten gain received today rather than tomorrow: yesterday’s actions determine today’s reputations and tomorrow’s results.

Cooperation is the antidote to today’s chaos. As the world ricochets away from collaboration toward zero-sum negotiations, we never forget that the future is forged in faith: faith of the possible, faith in others and faith in the power of human ingenuity and invention. It’s not through brutal individualized combat that humanity has progressed, but rather through a negotiated series of partnerships that tether us together in an intractable trajectory toward truth.

In this quarterly letter, we’ll discuss the mindmeld between man and machine, why partnerships are the bedrock of innovation, how we are incorporating Riskgaming to reduce surprise and accentuate our acumen, as well as highlight the leverage of small and cloistered partnerships to change the world. Progress begins with a pairing, and only through the right combos can they compound into the inventions everyone desires.

The Mindmeld between Man and Machine
The defining partnership of our era is the ancient yet accelerating covenant between Man and Machine. 

Consider the humble bicycle—among the most energy‑efficient devices our species has ever conjured. A morsel of cereal and some ounces of coffee get converted into miles of momentum through crank, chain and wheel. A person riding a bicycle travels farther per calorie than a salmon swimming upstream or a jet slicing the stratosphere. That simple frame neutralizes gravity with gyroscopic grace, turning balance into speed and inertia into freedom, all while letting its rider feel the wind as co‑pilot rather than foe.

Successive inventions turned our partnership with horses into partnerships with horsepower. Locomotives stitched continents, cars democratized distance, boats carved trade lanes, airplanes yanked geography taut, and elevators plus escalators granted us the vertical axis—human ambition rising floor by effortless floor. In every case, metal or silicon extended sinew, gear meshed with desire, and partnership—not replacement—reigned: the engineer still navigates the rail, the pilot still trims the flaps, the helmsman still tacks against the endless waves.

Motion is more than kinetic; it is emotional. Cameras became engines for empathy, transforming photons into feelings, letting a storyteller choreograph laughter, tears and even existential dread at 24 frames per second. A well‑timed cut can jolt an audience from comfort to horror; a particular color‑grade can bathe them in nostalgia or trigger uneasiness. Machines in partnership with our senses move man across our deep interior landscapes as surely as trains cut across the plains of Kansas.

Each of these inventions started as manual labor before rudimentary interfaces became enduring man-machine partnerships. The toiling labor of the draught animal was supplanted by rickety gas-powered tractors and then ultimately by today’s intelligent equipment that uses sensor fusion and geospatial analysis to autonomously adapt to the changing conditions of a farm.

Computers started as humans—women mostly—sitting in rows of desks to first tabulate artillery trajectories, then to tabulate ledgers, and eventually to tabulate words themselves. The invention of ENIAC opened the door to a new relationship, one where computers could partner with humans and carry out their wishes. Yet, this was not an easy partnership: punch cards and assembly languages frustrated all but the most experienced early coders. Today’s interfaces remain just as unyielding, demanding a memorized incantation of clicks and strokes.

Artificial intelligence is upending the implacable with the improvisational. Mice and macros are being replaced by chat and voice that can flexibly adapt to human intentions. Engineers now copilot the production of software alongside coding agents in a constant dialogue between mind and mechanical muse. Cinema directors are replacing the cold interfaces of editing software for the creative extensionalism of RunwayML, conjuring worlds from thoughts in an unceasing quest for the perfect shot.

Norms will change. The word processor begot spellcheck, and then grammar‑check followed. Today, LLM “style‑check” asks whether a paragraph was birthed by the human mind or statistical echo. We detect plagiarism, yes, but also something subtler—call it prompterism, prose coaxed from silicon’s latent space rather than summoned from lived experience. It is neither theft nor originality, a liminal authorship that demands new nuances in our culture and communications.

The same evolution of partnership can be seen with robots, our remote muscles. Bomb squads steer unmanned grippers, surgeons pilot Auris arms, astronauts tele‑operate rovers on moons dusted with solitude. At first, the human hand held the joystick in dominance; increasingly, the feedback loop runs both ways. The robot suggests trajectories, the avatar pre‑positions itself, the autonomous system flashes, “Are you sure?” before we even articulate intent. Anduril’s AI hunts drones but still salutes the human in the loop; meanwhile, Physical Intelligence is building generalized models of the world that will embed robots in all aspects of our daily lives. Agency is becoming a shared protocol, not a unilateral birthright. The glass, steel and code we forged are beginning to forge queries of their own.

What emerges is neither dystopia nor utopia but a new species of symbiosis. Some inventions remain just tools—from the microscopes and telescopes that let us see near and far to the microphones and speakers that silently suck in our voices and spout them out louder and farther. Yet, other inventions have also become our tutors, prompters and partners—co‑authors of movement, memory and meaning. Inside Lux we embrace that transition. Partnership persists, upgraded: man supplies context, conscience and narrative arc––for we will always be storytelling primates—while machines contribute acceleration, aptitude and tireless recall. Together they compose futures that neither entity—carbon nor code—could script alone.

The Power of Pairing in the Pursuit of the Possible
The story of any breakthrough begins with the power of pairing. History offers abundant examples. Crick’s ambitious theorizing was modulated by Watson’s careful empiricism in a balanced braid not unlike the helical strands of DNA they would discover. Lennon needed McCartney’s melodic optimism; McCartney needed Lennon’s serrated wit. Scorsese frames the shot while De Niro steals it. Warhol lent Basquiat a velvet‑rope audience, who repaid him with raw street voltage. These duets remind us that intellectual tension, properly harnessed, is not conflict but propulsion.

Our Lux founders ignite that creative thrust in their own founding partnerships. In a world always grasping toward a new Entropic Apex, it’s their attuned equilibrium that defines resilience. One's strength can be another's weakness. One's emotional highs can balance another's lows. The optimist invents the airplane; the pessimist, the parachute. The stronger the relationship, the more readily it can absorb the surprises of everyday life.

That inception is a combo that begins to compound in a nested network in pursuit of the possible: partnerships with employees, partnerships with customers and partnerships with investors. At the heart of every one of these partnerships is conviction, the belief that there is opportunity ready for the taking for those intrepid enough to seize it.

Lux’s conviction is the first gift we hand to a founding team. Long before Wall Street analysts christen an acronym or Gartner a quadrant, we wire the seed and signal the shoulder‑tap guarantee: we believe before others understand. When Cameron McCord founded Nominal to power mission-critical engineering, we didn’t manufacture excuses—we wrote a check. When Alex Wiltschko of Osmo.ai was ready to spin out of Google Brain with a few research papers on digital olfaction in tow, we didn’t give him a sniff test—we offered a check. When Chris Power of Hadrian decided to rebuild the factories of American aerospace and defense suppliers, we didn’t offer lectures on nuts and bolts—we endorsed a check. And when Josh DeFonzo of Mendaera embarked on inventing robots to transform healthcare, we didn’t strong arm and needle him—we sent a check. Capital paid the bills; conviction paid the courage tax.

We arrive at conviction not through relentless strongheadedness but rather through our own partnership and its orchestrated composition of competing ideas and experiences. Consider the table at our Monday partner meetings, where a mélange of priors sits—each very different, for if any two people were the same then one would be unnecessary. Electrical engineers and biologists intellectually spar with mechanical engineers and software coders as well as former journalists, diplomats and Delta Force commandos. Debate is ritual not rancor, because we've internalized that collective error diminishes as the boundaries of thought expands. 

Many of our friends have offered their counsel to advance our decision-making, which like in-phase and out-of-phase waves can constructively amplify or destructively cancel ideas. Our late friend Danny Kahneman’s siren of confirmation bias shrieks when consensus comes too quickly. Annie Duke challenges our convictions and whispers, “would you still bet on it?” Meanwhile, Michael Mauboussin reminds us that a flawless process can still yield a coin‑flip outcome—but a sloppy one rarely wins a second toss. By the time a decision emerges from our crucible, it carries the tensile strength of shared conviction, or at least one table-pounding partner who gets their one silver bullet. The competition over ideas never ceases.

The signing of a term sheet is merely the start of our alliance with a founding team, for then we must turn to constructing the coruscating and combinatorial cascade of strategics that will unleash that nucleus of brilliance at the heart of every Lux startup into a force that can change the world. Consider literally: Varda’s W‑series capsule is a concerto in collaborative logistics: boosted to orbit on a SpaceX rideshare and steered by a Rocket Lab Photon bus, it hosts an Air Force Research Laboratory spectrometer and NASA heat‑shield tiles, and then slips through the Australian dawn to deliver American micro‑gravity experiments and hypersonic data all in the same payload.

In parallel, Anduril, working shoulder‑to‑shoulder with Australia’s Royal Navy under the AUKUS banner, is co‑designing Ghost Shark—a stealthy autonomous submarine that shows how allied co‑development can sprint at startup speed. Saildrone fastened real‑time hurricane flux data to Palantir’s Foundry, turning Category Five chaos into climate clarity. Applied Intuition stitched a synthetic twin of Porsche’s next‑gen electric vehicle so Stuttgart engineers can crash‑test software, not sheet metal. Nvidia, the silicon sovereign, teamed with Together AI and Sakana AI, embracing a ‘coopetition’ loop in which every researcher accelerates the demand curve for H100s. Together, drawing on Nvidia‑accelerated infrastructure, now serves LLM deployments to Fortune 100 companies that in turn feed training insights back to Nvidia—an upward spiral born of enlightened self‑interest. Japan’s Ministry of Economy, Trade and Industry underwrote Sakana’s evolutionary diffusion models; America’s DIU, in turn, matched the grant, turning an AI lab into a strategic champion of sci‑tech diplomacy.

This is not the symmetrical grid of the planner, but rather the emergent pathways of the observant opportunist constantly seeking just one more connection that increases the speed of the flywheel. Some asset allocators believe in a linear approach to investing, with LPs investing capital into funds who invest into startups who invest in their employees who invest into products. We believe in consilient networks and not linearity. Our LPs are allied stakeholders, not silent partners. They call for counsel on direct deals in cutting-edge tech and we answer forthrightly, trust earned in restraint. In return, they swing open sovereign doors, anchoring financing rounds and helping accelerate deals like Variant Bio's Novo Nordisk partnership on metabolic disease. Reciprocity can compound faster than interest. So too for our partnerships with our peers with whom we both syndicate and compete with. Game theory and good moral compasses ensure we stay long-term greedy and never short-term transactional. It's algebra for adults: be positive-sum to win in today's zero-sum zeitgeist.

These compounding combos are never fixed, but extend through time. Most venture stories fade out at exit or shutdown, yet durable alliances telescope across decades. When a university endowment invests with us successfully, the lab they build in 2040 is invisibly penciled into today’s board slides. When Variant Bio shares upside with a coastal African community that provides genetic samples, a grandmother becomes, in effect, a downstream shareholder in a therapy her granddaughter may need in 2075. Even Anduril’s kill-chain code carries an implicit IOU: algorithms compiled this spring must still be legible to lieutenants not yet born. Partnerships are temporal bridges, letting present capital borrow insight from future users and future users borrow trust from present stewards. Compounding isn’t magic; it’s the dividend of promises kept—stretching from this quarter’s cash burn to a decades-scale mission log.

The global pursuit of Sci-Tech Diplomacy
The geopolitical map now resembles a circuit board more than a chessboard: nodes of talent and capital linked by diodes that permit forward current but block reverse betrayal. Alliances are no longer forged in marble chancelleries; they are coded in APIs, soldered on PCBs and inked in SAFE notes. Brett McGurk—who once fused 86 nations into the counter‑ISIS coalition—briefs our founders on Gulf diplomacy one day and Pacific deterrence the next. As we push "sci-tech diplomacy," we are increasingly engaged and invested with American allies from the United Kingdom, United Arab Emirates and Saudi Arabia to Japan and even Lithuania, where Atrandi seeks to upend the instrumentation of single-cell biotech.

We continue to fund cutting-edge aerospace and defense companies at home and in U.S. allied partners, and we also see big opportunities in tech, energy, supply chains and infrastructure in an emerging geopolitical strategic alliance that counters China's Belt and Road Initiative and will span from India onwards to the UAE, Saudi, Israel and ultimately to the European Union. The Silk Road is rapidly turning into the Silicon Road.

It’s not a simple path: export controls tighten, tariffs ricochet and electioneering threatens to politicize pure research. We respond by layering in redundancy—multiple launch providers, parallel fabs, mirrored data lakes—and by practicing what we'll call the "Diode Doctrine": current, capital and candor flow freely forward; betrayal meets an immediate open‑circuit. Defect once and you may win a round; cooperate consistently and you win the season. Long‑term greedy beats short‑term scarce in every simulated game and audited return.

In the current rollout of tariffs, we are surprised by the surprise of others. The one predictable pattern in the first Trump administration was unpredictability. Either the current rollout was the worst policy mistake in history, or more charitably, it may force countries and their leading companies to secure what is suddenly scant: stability and predictability. Might the likes of India, Japan, South Korea and Germany pay a premium to lock-in certainty in order to get capital spending and consumption back on track while possibly taking U.S. market share from competitive China-based suppliers? The odds may be higher than many think, and the long-term consequences much more inscrutable than people reading today’s A1 headlines appreciate.

Riskgaming and the lessons of adversarial collaboration
Uniting the race between the U.S. and China in electric vehicles, the Pentagon’s procurement of software and AI, the defense against deepfakes, and the climate adaptation of our nation’s navy is the alchemy of adversarial collaboration. Once more we turn to our late friend Danny Kahneman who championed this model—a nobler enterprise in which intellectual rivals join forces to fashion sturdier explanations. Think of Bohr and Einstein volleying letters about quantum uncertainty or DARPA’s Red Team / Blue Team gauntlets, where friendly fire tempers brittle code into battle-ready software. Competition sharpens arguments, cooperation aligns aims, congeniality oils the gears, conflict flushes out weak logic and consensus—earned, not imposed—lets the room adjourn with legitimacy intact.

Lux, for its part, designs our partnership to echo this model on purpose: many voices with no passengers, all incentivized to stress‑test a thesis until only the antifragile core remains. We aim to learn from and mirror the best because small rooms often write the biggest chapters. Belief without governance is bravado, so we institutionalize candor. Every Lux partner carries two envelopes: one labeled courage—introductions, recruiting, late‑night pep talks—and one labeled candor—the brutal truths murmured behind closed doors. Our mantra that failure comes from a failure to imagine failure means we rehearse downturns before they arrive, run pre‑mortems on rosy roadmaps, and simulate low-probability, high-magnitude black swans through Riskgaming, our simulated scenarios of complex science, technology and policy challenges.

Admirals and generals, congressmen and senators, asset allocators and startup prodigies, leading scientists and brilliant philanthropists have all stood around our Lux Riskgaming tables to role‑play the very scenarios highlighted at the start. The exercise forces stakeholders who may never share a cap table into a shared concert of consequence, each discovering that the other’s weakness is a future self-insurance policy. The insights are immediate: shipyards crack when trust thins; electric vehicles short out as the voltage between players spikes; artificial intelligence security breaks down where adversarial intelligences misalign.

Riskgaming itself is an engine for partnerships with brilliant minds who want to see tomorrow, today. Born at Lux, we have co‑hosted scenarios with Mike Bloomberg’s Bloomberg LP; with leading Canadian pensions whose trillions finance tomorrow’s factories; with prominent U.S. Senators whose dogged work protects America every day; and with prominent think tanks, peer venture funds and powerful diplomatic delegations who convene cohorts of the curious. The result is optionality: founders walk away with backup suppliers and back‑channel phone numbers; LPs harvest foresight before headlines price it in; and Lux gets up-close takes on talent and a trove of new investable ideas not yet appreciated by others.

The large impact of small partnerships
Partnership flourishes in cloisters—literal and metaphorical—where a handful of humans huddle to hammer out destinies for the multitudes. All kinds of institutions exist to channel ambitions and rivalries into productive outcomes for all. The institution of science itself—with its collaborative co‑authors and competitive peer‑review—serves as civilization’s open‑source operating system, an error‑correcting network where doubt is not a defect but rather the impetus that iterates us toward that asymptote of truth.

We see similar leveraged networks across every arena where authority concentrates. Inside the Pentagon, compact cross‑functional planning cells—often fewer than a dozen senior officers, analysts and legal advisers—assemble pre‑vetted target packages for rapid‑strike options. Their recommendations still travel the full National Command chain, yet the groundwork they lay can compress what once took days into hours, steering multi‑billion‑dollar assets with sobering precision. In Hollywood, a half‑dozen development execs—a green-light gang—debate whether an untested director gets $100 million to reboot a superhero franchise. Their verdict will tilt culture for a generation. The NBA’s Competition Committee of roughly a dozen owners, executives, coaches and player representatives debates proposals to change the fundamental rules of basketball—such as testing the Elam Ending or widening the lane—and then forwards its recommendations to the Board of Governors. A single tweak will echo from playgrounds to primetime, reshaping the moves kids emulate and the professionals undertake.

Some councils operate in darkness: three traders inside a macro hedge fund decide to yank liquidity from a sovereign bond market, and a currency collapses by cocktail hour. Others sit before footlights: the Writers Guild negotiating caucus—roughly 30 pens wielding millions of metaphors—went toe‑to‑toe with studio lawyers, striking not just for pay but for the very definition of authorship in this AI age. And then there’s the International Commission on Stratigraphy, a wonkish quorum of geologists whose multi‑year ballots and globe‑trotting conferences determine whether the Anthropocene should be enshrined as a formal epoch—a semantic stroke that recalibrates textbooks, museum exhibits and climate‑policy budgets worldwide.

Or consider the Papal conclave, that Renaissance‑calibrated choreography now swirling in Rome and which flowed through American cinema last year with Ralph Fiennes’ iconic performance in Conclave. Doors close, windows are sealed and a micro‑parliament of scarlet shepherds locks itself away with nothing but ballots, Latin liturgy and centuries of precedent. Shakespeare would savor the scene: daggers in smiles and blessings in whispers, each cardinal weighing the legacy of grace against the limits of geopolitics while white smoke waits like an anxious exit poll above the Sistine rafters. A billion believers will wake up one day to a new name, a new rhetoric and a new set of priorities—all because a small, self‑contained partnership navigated a gauntlet of competition and cooperation, congeniality and conniving, conflict and consensus.

These conclaves, committees and cabals—call them what you will—are essentially microchips: few inputs, vast outputs. They remind us that scale is not merely a function of mass; it is the product of leverage. Twelve jurors decide guilt or innocence for crimes that alter futures and reverberate through families; five Supreme Court justices can rewrite the social software of a continent. Even in Silicon Valley, a three-member board can choose whether to open‑source a model that will empower cutting-edge coders or disinformation despots instead. The settings are intimate; the stakes just as often enormous.

Progress is the Product of Partnerships
So where does that leave us in mid-2025? With a portfolio whose value is measured not merely in dollars and patents but in the resilience of its relationships. With LPs who phone us for foresight, founders who text us in triumph and terror, governments that increasingly engage us to close capability gaps, and competitively-motivated corporates that want to partner with Lux companies to leap-frog other incumbents.

If the last era lionized the individual genius—a myth of the narcissists, no doubt—the next may just exalt the deliberate duet, the orchestrated ensemble tuned to compound, not just compete. All of the above examples are not disparate anecdotes but harmonic overtones of a single theme: progress is the product of partnerships. We deeply value yours. Fiat Lux.