Titanic Lessons


The Greek myths of the Titans have been taught over millennia, primeval paragons of progress and perilous pride whose stories speak to the verities and vanities of mankind. The four brothers descended from Iapetus and Clymene are archetypes, ones both divine and ever-so-human. There’s Prometheus, the rebel Titan who defies Zeus to offer fire to the newly-formed yet ill-equipped humans, his revolt affording knowledge and civilization. There’s Atlas, the enduring holder of Earth who would come to symbolize strength through adversity and the indomitable spirit of exploration.

‍

Then there’s Epimetheus and Menoetius. Unsurprisingly, the two brothers who represent human weakness have held far less glamorous positions in mythology than their brethren. Epimetheus impulsively accepts a gift from Pandora — Zeus’s trap to unleash horrors on humanity — ignorant of the consequences of his decision. Menoetius, meanwhile, is struck down by Zeus, a lightning bolt against his reckless hubris.

Lux (Latin for “light”) has always used past lessons to inform our work on the future edges of human progress. If for Orwell, “Who controls the past controls the future. Who controls the present controls the past,” a corollary is that those who fail to understand the past can’t comprehend the present, and thus have no handle on the future. The evolution of science and technology varies, but human nature is invariant. The Titanic Lessons of the canon are the kernels of comprehension.

In preview then, a letter of four parables. The Prometheus Promise, our thesis-driven framework of “extensionalism” — investing in ventures that technologically transcend and extend human limits, from our senses of vision, hearing and olfaction to the embodied intelligence of robots and ultimately to the xenotransplantation of critical human organs. The Menoetius Moment, the dangerous pride of too many AI investors amid a generational frenzy. The Epimetheus Escape, how Lux is using Riskgaming as a tool for deliberative foresight across portfolio construction and risk analysis. Finally, the Atlas Answer, how four strategic initiatives at Lux strengthen our enduring exploration of the frontiers that can deliver humanity the progress we’ve all been waiting for.

The Prometheus Promise: Extending Human Limits


Prometheus was said to have molded humanity out of clay, creating a “blank slate” species that was at first powerless in an unforgiving world. Rebelling against the wishes of Zeus, he would offer fire — an elemental technology — to humans, affording them warmth and salvation against hunger, while also ushering in the seeds of violence through flame. The Prometheus Promise emphasizes the rebellious nature of progress, that one must thoughtfully defy strictures to reach a new and better state of existence.

‍

Millennia later in 1798 that blank slate would reappear, with English philosopher John Locke describing the human mind as a “white paper, void of all characters, without any ideas” that only acquired knowledge through sensory experience. Our senses — sight, sound, touch, taste, and smell — were the gatekeepers of understanding, the bridge between our consciousness and reality. Today, we find ourselves at a remarkable inflection point: technology is not just augmenting our natural senses, but fundamentally extending and transforming them in ways that would seem like science fiction to Locke and the Greeks. This is the essence of extensionalism — the 'overedge' where human potential transcends its boundaries into uncharted territory.

Today, Lux companies are at the forefront of this sensory revolution. Consider vision: from van Leeuwenhoek's microscope unveiling the cellular world to the James Webb Space Telescope peering into the depths of the cosmos, progressive inventions have expanded our visual frontiers. Matterport and Eikon are carrying this torch forward, transcending the limitations of human sight. Matterport's technology creates digital twins of physical spaces, allowing us to see and interact with environments in ways our eyes never could — a spiritual successor to the camera obscura that first let us capture reality. Building on Nobel Prize-winning research, Eikon’s technology lets us watch individual molecules dance inside living cells — a feat that would have seemed impossible even to the pioneers of electron microscopy.

Sound has followed a similar wave. From the earliest ear trumpets to Alexander Graham Bell's telephone, we've continuously sought to extend our auditory reach. Now, Lux companies Chromatic and Phonic are revolutionizing how we process and enhance auditory information. Chromatic’s AI-powered hearing aids don't just amplify sound like the crude acoustic horns of the past. Instead, they use sophisticated yet power-efficient algorithms to identify and enhance speech while suppressing background noise, effectively giving us superhuman hearing in even the noisiest environments. Phonic, meanwhile, generates new audio through advanced AI models, shaping the future of aural experiences.

For touch and motion, humanity's progress has been equally dramatic. The lever and pulley of antiquity gave way to the steam shovel and excavator, the forklift and crane — each extension multiplying human strength and reach. Today, we've seen remarkable advances beyond mere mechanical advantage, from CTRL-Labs (acquired by Meta) translating neuromuscular signals to Physical Intelligence developing modular AI brains for diverse robotic hardware.

The final frontier may be our most ambitious yet: digitizing smell. While humans have long captured and recreated visual and auditory experiences, scent has proved to be far more elusive. Through Osmo, we're finally cracking the code of olfaction, using AI and combinatorial chemistry to not just identify but reproduce scents. By unlocking an entirely new dataset for model training, the applications are as boundless as the air around us. Just as the invention of photography let us preserve and analyze visual information in new ways, digitizing smell could revolutionize everything from disease detection and environmental monitoring to designing consumer experiences.

Even the modern state's quest for legibility — what James Scott in Seeing Like a State saw as the drive to make complex systems measurable and manageable — has found its technological apotheosis in Lux-family companies extending human sensing capabilities across sea and space. Saildrone's autonomous vessels now traverse vast oceanic expanses even through the epicenters of megastorms like Hurricane Milton, rendering the once-inscrutable seas into streams of actionable data. Planet advanced omniscient observation through satellites, transforming terra incognita into a real-time atlas. As we extend human senses to scales and spaces previously reserved for science fiction, we transform the un-sensed, opaque and illegible into the seeable, knowable and quantifiable.

From Sci-Fi to Sci-Fact: The eGenesis Breakthrough


The story of Prometheus isn’t complete without his divine punishment for offering fire to humanity: every day thereafter, an eagle sent by Zeus comes to eat his liver as he is bound to a rock, only for his liver to regrow and the punishment to be repeated. It’s a story about suffering, but unfortunately for too many people today, their suffering is from a lack of organs — and not their regeneration.

‍

Roughly 100,000 Americans are waiting for a kidney transplant alone, tethered to dialysis machines while hoping for a donor. The current organ procurement system operates in perpetual crisis, racing against biology's strict timeline to deliver precious organs from unexpected donors to desperate recipients. The cost is staggering — a typical kidney transplant runs $442,500 — while the human toll is immeasurable.

Perhaps no investment better exemplifies our ability to transcend the limits of human physiology than eGenesis, in which we recently led a $191 million financing — the largest initial check in Lux's history. eGenesis uses CRISPR gene editing to make xenotransplantation (cross-species organ transplants) not just possible but practical. Earlier this year, the company achieved a landmark milestone: successfully transplanting a gene-edited porcine kidney into a 62-year-old patient at Massachusetts General Hospital. This was more than a medical breakthrough — it was proof that those with sufficient vision and persistence can bend science fiction into science fact.

Xenotransplantation could transform a crisis-driven organ system to a planned, preventative one. Rather than waiting until patients are at death's door, we could intervene earlier, dramatically improving outcomes while rapidly reducing costs. That’s the Prometheus Promise: using cutting-edge science to expand well past human limits, saving lives while building massive new markets. It’s precisely the kind of ambitious bet that venture capital should be making.

The Menoetius Moment: The AI Infrastructure Race


Menoetius is the forgotten Titan child, left behind because of humanity’s penchant to ignore the hard-fought lessons of the past in pursuit of a boundless and beckoning future. His excessive and arrogant conduct in the conflict with the Olympians ultimately leads Zeus to strike him down with a lightning bolt, symbolizing the power of light (lux) to enervate the hubristic behavior of a Titan out of control.

We feel that artificial intelligence is hitting its Menoetius Moment, when the prideful ignorance of too many reckless prophets of the future will finally be revealed. Amid this soaring frenzy, the only salve in this Titanic Lesson is the ever-deeper unearthing of ground truth.

Take the collision of AI and traditional infrastructure, which is creating important second-order effects. The conventional narrative centers on an arms race around scaling laws: larger models, more data, bigger clusters, greater power. Today's largest AI clusters use around 100,000 Nvidia H100 chips and consume roughly 100 megawatts of power. Next year's clusters will scale to 300,000-500,000 chips requiring nearly a gigawatt of power — about the consumption of a small city. When Elon Musk's xAI couldn't secure sufficient grid power in Memphis, it brought in mobile natural gas generators to add 35 megawatts. Meta's new $1 billion Arizona facility alone will require 12,000 tons of steel — the weight of 30 fully-loaded Boeing 747s.

These aren't rounding errors. While the headlines focus on ChatGPT and semiconductor chips, the real story is one of physical constraints. The tech giants who built empires on weightless bits and bytes are now grappling with atoms: steel, copper, water rights, and critically, natural gas. In the past five years, just three companies — Microsoft, Meta and Alphabet — have spent over $370 billion on new physical infrastructure, with their latest quarterly run-rate exceeding $140 billion. We still remember the graveyard of past internet infrastructure bets, including the optical fiber networking giants 25 years ago and the cloud-shifting colocation providers 15 years ago, where unwary venture capitalists went bust while distressed investors made a fortune. Once again, the bottlenecks in AI are shifting rapidly, presaging massive dislocations in infrastructure. In 2023, the limiting factors were chip-on-wafer-on-substrate technology and high-bandwidth memory. In 2024, it has been Nvidia’s H100 chips and networking equipment that can only effectively coordinate clusters of about 100,000 units. Looking ahead to 2025 and beyond, power infrastructure may become the primary constraint: America only has a handful of locations that can provide the gigawatt-scale power these massive clusters will require. While we’re bullish on the seeming resurgence of elemental energy — our rebrand of nuclear power — abundant natural gas from the Texas Permian seems a wiser bet.

The initial phase of any major innovation often emphasizes raw power and scale, but the lasting advantages ultimately accrue to those who figure out how to do more with less. Just as the personal computer eventually overtook mainframes by being more efficient and accessible, we believe the future of AI will be shaped by companies that can democratize access through clever engineering rather than scaling brute-force computation.

Thus, we see two different futures emerging simultaneously. Distributed computing approaches can eliminate the need for massive monolithic clusters, as exemplified by our investment in Together AI. Together has become one of the fastest companies to reach $100 million in annualized revenue in our history by enabling organizations to build and run generative AI workloads more efficiently.

The second future may come from an unexpected quarter: Apple. While the industry fixates on bigger data centers and more powerful chips, Apple has quietly published research showing how to run large language models directly on devices with limited memory. By leveraging novel storage techniques and the inherent sparsity in AI models, the company can reduce memory requirements by 50% while maintaining performance. Advances like these could enable future iPhones and Macs to run sophisticated AI models locally and privately — a development that could radically reshape the entire AI infrastructure landscape.

Lessons from our AI Summit in New York City


The Menoetius Moment demands perspicacity, which we continually refine through our private conversations with technically adept leaders who are pioneering the pathbreaking innovations on the AI frontier. At our recent second-annual AI Summit in New York City, we brought together over 150 CEOs, founders, researchers and policymakers to explore AI's horizons. The Summit was electrifying, reflecting NYC's emergence as a major AI hub with 20 new unicorns in 2023 and 10 more already in 2024. Beneath the surface excitement though, we identified critical insights that are grounding our investment approach from the irrationality of the markets.

National Security and AI


Senate Intelligence Committee chair Mark Warner emphasized how national security must be redefined beyond traditional military assets to encompass technological competition. The divide isn't left versus right, but past versus future. As West Point's Dean, Brigadier General Shane Reeves, starkly put it: "if our officers are not AI-enabled, we will lose." This recognition of AI's critical role in modern warfare underscores why Lux continues to invest in dual-use technologies that can serve both commercial and defense applications.

Infrastructure + Efficiency


Leaders from Together AI and Hugging Face highlighted the growing tension between AI's computational demands and economic reality. As Vipul Ved Prakash of Together noted, "you start to realize what your yearly bill is going to be, and it just kind of makes sense to invest in [custom models] and make something more efficient." This creates opportunities for companies building more efficient AI infrastructure and specialized models.

Democratization + Applications


From Runway's stunning video-generation capabilities to tldraw's collaboration innovations, we saw how AI is becoming more accessible and useful across industries including cinema, design, protein folding and sensing. Fears of bread lines and mass unemployment have been replaced with optimism around the opportunities for productivity growth and further empowerment.

The Epimetheus Escape: Portfolio Construction + Riskgaming


We turn now to our third parable. Epimetheus’s contribution to Greek myth is his impulsiveness, notably his relationship to Pandora. Humanity suffers the tragedies of Pandora’s box (actually a jug) due to his naivety and lack of forethought. More well-intentioned than his brother Menoetius, Epimetheus represents the danger of divorcing motivation from consequence. The critical Titanic Lesson is that the benefits of the future start with assessing the tradeoffs of the present constrained by the decisions of the past.

In markets: equities, gold, and Bitcoin are at record highs, bank earnings are ample and credit markets haven’t cracked. Inflation is still above the Fed’s target, while Powell risks repeating the mistake of the 1970s by easing interest rates before inflation has been fully slayed — especially in light of the new administration’s policy goals. Higher for longer is our posture on the natural cost of capital, regardless of what the Fed does.

While many assets are awash in capital, the venture ecosystem is traversing a Great Liquidity Desert: distributions to LPs have hit a bone-dry 5.2% of NAV — the lowest since 2008’s Great Financial Crisis and a brutal eight-quarter stretch of single-digit returns in an industry accustomed to nearly 17% quarterly distributions over the past decade. While the techno-optimists focus on frothy greenshoots, they’re missing the industry’s further desertification: over 30% of 50,000 private companies are trapped in late-stage limbo while over 730 unicorns are trapped in exit-less purgatory, with nearly half of them over a decade old and dragging down returns like sea-sinking anchors.

Remember our aggressive prediction in prior letters that 30-50% of VCs would exit the industry. The inferno is spreading faster than even we anticipated: 25% of active investors from 2023 have already vanished (from over 15,000 to under 11,500), while surveys show 40% of GPs have either abandoned plans for their next fund or indefinitely delayed fundraising. We repeat that what others call "dry powder" is actually sopping wet — the data shows just 23.1% of supposed dry powder was deployed in deals this past year, the lowest ratio since 2010. With 2024 fundraising projected to be less than half of 2021's $352 billion peak, this isn't just a cycle; it's a structural cleansing, set to reshape venture capital we think for the better.

For those with truly available capital and discipline, the next 18 months will separate the quick from the dead. What’s needed is the Epimetheus Escape, the careful use of foresight to avoid the deadwood in pursuit of new growth.

That’s where Riskgaming comes in. Our Lux Riskgaming Initiative has evolved into a powerful tool for anticipating and preparing for future challenges. Recent scenarios have explored climate-driven defense infrastructure failures and deepfake foreign election interference to the future of AI-driven national security and the auto industry. These exercises help us and our portfolio companies think through complex possibilities and prepare for contingent futures. One particularly relevant scenario explored the future of electric vehicle manufacturing and China's growing dominance in this sector. The insights we gained have informed our investment thesis around supply-chain resilience and domestic manufacturing capabilities. We can escape Epimetheus, with competitive deliberation and the constant rebalancing of risk.

The Atlas Answer: Lux Labs, Corporate Spin-offs, Tactical Global Opportunities and Fixware


We turn now to our fourth and final Titanic Lesson, that of Atlas. Atlas may be the most visually recognizable of the brothers, burdening the weight of the world as a punishment from Zeus. But over time, Atlas became a symbol of navigation and enduring exploration, eventually becoming the definition of maps bound together thanks to Gerardus Mercator (of the Mercator projection). Similar to Locke’s tabula rasa, Atlas harkens us to the terra incognita, those ‘overedges’ of the map that remain undiscovered, save for an intrepid explorer whose ambition exceeds their risk of demise and infamy.

Riskgaming has highlighted a structural shift toward a rising cost of capital. Regardless of future Fed rate cuts, we believe the natural cost of capital is increasing. This shift demands strategic adaptation, leading us to develop four key strategies for deploying capital in this environment: Lux Labs, corporate spin-offs, targeted tactical global opportunities and finally, maintenance (“fixware”).

America remains unrivaled in economic and scientific dynamism, yet we systematically undermine our own success through sclerotic institutions and myopic policy choices. We're rich on paper but poor in practice — our infrastructure crumbles while financial engineering hollows out industrial titans like Boeing and Intel. Our scientific institutions are calcified and risk-averse, our immigration system actively repels the world's brightest minds, and our researchers spend nearly half their time on grant applications rather than discovery.

The silver lining? This dysfunction creates opportunities for ambitious builders willing to route around the institutional roadblocks. That's why we launched Lux Labs — to give brilliant scientists and technologists an entrepreneurial escape velocity from the inefficient scientific-industrial complex.

Lux Labs formalizes our nearly twenty-year track record of transitioning breakthrough science from the laboratory to the marketplace. We've observed repeatedly: excellent ideas ready for transition to the real world remain mired in academic and corporate research labs, constrained not by technical limitations but by institutional friction. Sometimes the friction comes from university intellectual property procedures. Sometimes it's the logistics of spinning out an under-funded corporate team. Often, it's the challenge of aggregating talent, of creating the critical mass needed to bring interested but employed teammates together.

We've already proven this model works through more than 20 companies, including Aeva in next-generation autonomous vehicle sensing; Aera in transformative genetic medicines; Cajal Neuroscience in neurodegenerative disease; Kurion in novel technologies for nuclear waste remediation; and Variant Bio in genetic-outlier discovery. Like Atlas, we’re constantly seeking to take on the burden of freeing world-leading talent to pursue their entrepreneurial ambitions.

Turn now to the largest tech companies. When capital costs were low, it acted like a tractor beam for the future, pulling 20-year, far-out ideas into 20-month frenzied projects. Now, as the cost of capital rises, time horizons are contracting, and experiments are being canceled amid tighter fiscal discipline. This creates opportunity.

Alongside Lux Labs, we have successfully spun out or have been first investors behind de-risked technologies and teams from big tech. Examples include Osmo and Physical Intelligence from Google, Aeva (and another not yet disclosed) from Apple, and Evolutionary Scale from Meta. These unlock value by giving breakthrough teams and technologies the funds, freedom and focused entrepreneurial attention to thrive.

Our pursuit of breakthrough technologies and pathbreaking teams knows no borders. In Japan, we've made strategic investments like Sakana AI, which recruits world-class talent in Tokyo to pioneer new approaches to large language models. The company's name, meaning "fish" in Japanese, evokes individual fish forming a school with emergent behaviors — a perfect metaphor for their work in nature-inspired intelligence and complex evolutionary systems.

Sakana represents another departure from the "bigger is better" paradigm in AI infrastructure discussed earlier. While one large model provider needed 75,000 GPUs to train their latest model, Sakana's approach required just eight GPUs to achieve impressive results. This isn't just about cost savings, but about fundamentally rethinking how AI systems can be architected.

Israel too has captured our attention, particularly in aerospace and defense. Current geopolitical challenges have only intensified the country’s focus on real technologies from highly-technical operators. We've made our first two investments there, co-investing alongside our partners at Sequoia and Founders Fund, respectively, in companies that will emerge from stealth mode in due course. We’re backing founders who, like Atlas, were forged in fire and we expect will embody Israel’s greatest generation as they carry the burdens of freedom forward.

Finally, let’s turn to the burden of maintenance. For over a decade of low rates, investors and companies prioritized growth at all costs. Now, with rising costs of capital, attention is turning to the other half of capex: maintaining existing assets. We might call this emerging category "fixware" — new technologies that upgrade, modify and maintain existing systems. This opportunity spans software, sensors, automation and robotics, touching everything from buildings and transit systems to military installations and satellites.

We’ve been investing in companies that are developing innovative solutions for maintaining and upgrading infrastructure, such as Nominal and Epsilon3 in logistics and aerospace, Impulse Space in space missions, Hadrian in aerospace supply chains, and Kinetic Auto in electric vehicles.

The defining symbol of Atlas is his enduring burden of maintaining the Earth’s position on his back, of securing it against chaos and entropy while searching the stars for what’s next. Through all four of these strategic initiatives, we track down the best entrepreneurs in the world and burden them with a grand mission: to carry forward human progress for another generation.

From Past to Future


The Titanic Lessons can be summarized thusly: compete ferociously to secure human freedom and progress (Prometheus), be wary of the hubris of groupthink (Menoetius), be proactive in anticipating the consequences that present decisions have on the future (Epimetheus) and finally, cooperate with others to take on the great burdens of society (Atlas). From our thesis-driven framework of extensionalism to our focus on transforming science fiction into fact, from our orthogonal views  on AI infrastructure to our Riskgaming-derived portfolio strategies — each represents a different facet of how we compete and win the future through the lessons of the past.

Yet winning, for us, has always meant more than financial returns. It means expanding the boundaries of human capability. It means turning long-shot scientific breakthroughs into world-changing companies. It means finding and funding those rare founders who see possibilities where others see only problems.

In 1798, Locke's "blank slate" represented the ultimate human potential, waiting to be written upon by sensory experience. Today, we face a different kind of blank slate — not of the individual mind, but of our collective future. Through breakthroughs in sensory enhancement, AI, and biotechnology, we're not just writing on this slate — we're expanding its very dimensions.

We often say at Lux that we believe before others understand. This requires not just vision, but conviction — the ability to see potential when others see only risk, to recognize patterns in seeming chaos, and to keep the faith when conventional wisdom suggests otherwise. The chain of belief of course starts with you, our partners. You believed in us first, which allowed us to believe in our founders, and we believe many of our founders will significantly reward us all.

As we look to the remainder of 2024 and beyond, several trends bear watching: the acceleration of AI capabilities beyond language models into biology, chemistry and physical systems; the growing imperative for domestic manufacturing and supply-chain resilience; the continued convergence of commercial and defense technology; the evolution of experimental science through automation; and finally, the transformation of healthcare through sci-fi breakthroughs like xenotransplantation.

In the end, we return to where we started: compassionately competitive, motivated and mission-driven, helping to write the future, one breakthrough at a time.

We're betting on the impatient optimists who refuse to wait for permission to build the future.

Fiat Lux.

‍