The term sheet landed with two names that don't typically share a line item. Neuralink—the brain-computer interface company with three human implants and a valuation narrative built on a future that hasn't arrived yet. Anduril—the defense AI firm with $1 billion in revenue and contracts from the Pentagon and beyond. MSX's Pre-IPO Phase III fund has put both companies in the same portfolio, and the signal is louder than either name alone.
This isn't a diversified bet on "AI" as a sector. It's a thesis about where AI terminates: into the human nervous system and into the battlefield kill chain. From brain to battlefield, MSX is constructing a portfolio that spans the outermost boundaries of AI application. The logic is seductive. The risks are structural.
Let me walk through what this investment actually means, starting with the technology stack.
The Technological Extremes: N1 Implants and Lattice OS
Neuralink's N1 implant represents a specific engineering bet: high-density flexible electrodes threaded into brain tissue via the R1 surgical robot, which can avoid blood vessels during insertion. The system captures neural signals across 1,024 channels—an order of magnitude beyond what competitors like Synchron's Stentrode can achieve with their 16-channel non-invasive approach. Three patients have received implants since January 2024. The FDA granted breakthrough device designation in 2025, and the "CONVOY" feasibility study approved in November 2025 will test whether paralyzed patients can control a robotic arm with their thoughts.
This is proof-of-concept territory. The technology works at a demonstration level. What it doesn't have is longitudinal safety data, manufacturing scale, or a regulatory pathway that guarantees commercialization. The gap between three patients and a viable medical product is measured in years and billions of dollars.
Anduril operates on a completely different maturity curve. Its Lattice platform functions as an AI-driven operating system for defense—sensor fusion, autonomous decision-making, and coordinated response across heterogeneous hardware. The product stack spans counter-drone systems, autonomous underwater vehicles, and the Roadrunner missile interceptor. In June 2025, Anduril announced a strategic partnership with OpenAI to develop counter-UAS systems for military bases. In December, it moved to acquire the UK-based 3D printing drone company Callum for approximately $1.3-1.4 billion. The company projects $1 billion in 2025 revenue, a 150% year-over-year increase.
The Lattice approach is "software-defined defense"—wrapping AI software around hardware components to create a closed loop of perception, decision, and action. This isn't theoretical. These systems are deployed, operational, and contracted.
The maturity gap between these two companies is enormous, yet MSX treats them as comparable investment opportunities. That alone tells you something about the fund's thesis: they're not betting on current fundamentals. They're betting on narrative endpoints.
Two Commercialization Models, One Exit Window
The commercialization paths couldn't be more different. Neuralink follows a "medical first, consumer later" trajectory. Right now, the company operates under FDA IDE (Investigational Device Exemption) for research purposes only—there is no commercial revenue. The earliest realistic pathway to market approval runs through 2026-2027, with device pricing likely in the $10,000-$50,000 range based on competitor benchmarks. Consumer applications like memory enhancement or direct human-AI interaction are post-2030 scenarios, contingent on solving safety, ethical, and regulatory hurdles that don't even have frameworks yet.
Anduril is a B2G machine. Its revenue comes from defense contracts—the U.S. Department of Defense, the UK Ministry of Defence, and allied governments. In December 2025, the Defense Innovation Unit awarded Anduril a $250 million contract for "large-scale expeditionary operations." The company expects to reach break-even by 2026 and profitability by 2027.
From an investment perspective, these are two completely different risk profiles masquerading as one "AI frontier" thesis. Neuralink is a binary bet on regulatory approval and clinical success. Anduril is a compounding bet on defense budget cycles and geopolitical tension.
The valuation math reflects this divergence. Neuralink was valued at approximately $5 billion in its 2023 funding round, with current Pre-IPO estimates ranging from $8-10 billion—against zero revenue. Anduril raised at $14 billion in 2024, with current expectations around $25-30 billion, against roughly $1 billion in revenue. That's a 14-28x price-to-sales ratio for Anduril, versus an infinite P/S for Neuralink.
The unspoken question is whether these valuations hold if the AI narrative cools. Pre-IPO funds typically target 3-5x returns over a 2-3 year holding period. That math requires both companies to deliver on aggressive timelines: Neuralink hitting FDA approval milestones and Anduril continuing its contract momentum.
The Competitive Matrix: Who's Actually Winning?
In the brain-computer interface space, Neuralink holds the technical lead but faces a crowded field of alternative approaches. Synchron's non-invasive Stentrode has been implanted in 10 patients—more than Neuralink's three. Precision Neuroscience has performed 30 procedures with its minimally invasive Layer 7 electrode array. Paradromics and Blackrock Neurotech are also advancing high-channel-count systems.
The competitive tension here is fundamental: Neuralink's invasive approach offers higher bandwidth but carries greater risk. Non-invasive approaches sacrifice bandwidth for safety and regulatory simplicity. If Synchron or Precision Neuroscience can demonstrate adequate performance with lower risk profiles, Neuralink's technical superiority may not translate into market leadership.
In defense AI, Anduril faces a different kind of competition. Palantir—with its $300 billion market cap and deep government relationships—competes on the AI decision layer, though it lacks Anduril's hardware integration. Shield AI focuses specifically on autonomous flight with its V-BAT systems. Traditional defense primes like Lockheed Martin and RTX are pursuing their own AI transformations, albeit at the pace of organizations with legacy infrastructure and procurement obligations.
Anduril's "software-defined" approach gives it a first-mover advantage in a market that's still forming. But the defense industry rewards incumbency, relationship depth, and security clearances—advantages that don't automatically accrue to technical innovation. The company's real challenge is whether it can scale hardware manufacturing to match its software ambitions.
Industry Disruption: The Double-Edged Sword
The broader industry implications of this dual investment are worth unpacking. Neuralink's technology could reshape treatment for neurological conditions affecting hundreds of millions of people globally—epilepsy affects approximately 50 million patients, Parkinson's disease 15 million, and spinal cord injuries 10 million. If brain-computer interfaces move from experimental to therapeutic, they create an entirely new category of "neuro-digital therapeutics."
Anduril's model is disrupting the defense industrial base itself. The shift from hardware-centric to software-centric defense systems represents a structural change in how military capabilities are developed and deployed. This isn't incremental improvement; it's a paradigm shift in procurement, development, and operational doctrine.
But both trajectories carry unresolved ethical and security questions. Neuralink's invasive technology raises issues of autonomy, identity, and privacy that no regulatory framework has adequately addressed. The data from neural interfaces is the most sensitive personal information conceivable—a breach could constitute something close to thought exposure. Anduril's autonomous systems operate in a domain where failure means casualties, where AI decision-making under battlefield conditions introduces unpredictable risks, and where the international community has not reached consensus on lethal autonomous weapons systems (LAWS) governance.
Composability is a double-edged sword in this context. The same AI capabilities that enhance human potential can be repurposed for cognitive surveillance. The same autonomous systems that protect military bases can be deployed in ways that trigger arms races. MSX's investment isn't just a financial decision—it's a position on which side of these dual-use equations will dominate.
The Investment Thesis: Beyond the Hype Cycle
Let me be direct about what this Pre-IPO allocation really represents. MSX is making a portfolio-level bet that AI's next phase of value creation will emerge at the physical and biological boundaries of computation. The "brain-to-battlefield" narrative provides a coherent story for LPs—it's the kind of framing that resonates in fundraising decks and private placement memoranda.
But I've been through enough cycles to recognize the pattern. The 2017 ICO bubble taught me that narratives about transformative technology are cheap; execution against technical milestones is expensive. When I modeled the liquidity flows of 50+ Ethereum ICOs, I found a critical correlation between whitepaper buzzwords and short-term price pumps—and almost no correlation between those buzzwords and actual product delivery. The market rewarded storytelling until it didn't.
The question for MSX is whether Neuralink and Anduril represent genuine execution or compelling narratives. Anduril's revenue growth suggests real execution. Neuralink's clinical progress—while early—shows technical competence. But the valuation gap between narrative and fundamentals is where Pre-IPO investments either create outsized returns or destroy capital.
Algorithms don't fail; models do. The model here assumes that AI's frontier applications will command premium valuations in public markets, that regulatory pathways will clear on schedule, and that competitive pressures won't erode first-mover advantages. Any one of these assumptions failing changes the risk calculus significantly.
The Contrarian Angle: Decoupling from the AI Trade
Here's where I diverge from the consensus read. The market is treating Neuralink and Anduril as part of the broader "AI trade"—the same thematic bucket as Nvidia, OpenAI, and every startup with "GPT" in its pitch deck. This conflation is a mistake.
Neuralink and Anduril are not software companies with AI features. They are hardware companies with AI cores. Their value chains involve manufacturing, regulatory approval, government procurement, and physical deployment. They don't scale like software; they scale like industrial enterprises. This means their risk profiles, margin structures, and valuation dynamics are fundamentally different from the AI software layer.
This distinction matters for investors. The AI software trade has been characterized by extraordinary multiples and rapid adoption curves. The AI hardware trade is slower, more capital-intensive, and more exposed to execution risk. If the AI trade decouples into "software AI" and "physical AI" segments, Neuralink and Anduril belong to the latter—with very different performance characteristics.
There's also a geopolitical dimension that deserves attention. Anduril's defense contracts tie its fortunes to U.S. defense budgets and allied military modernization programs. A shift in political priorities, a budget sequestration, or a de-escalation in global tensions would directly impact its revenue trajectory. Neuralink's medical focus is less geopolitically exposed but more regulatory-dependent. The two companies face entirely different macro sensitivities.
What the Market Misses
The market's attention is focused on Neuralink's next clinical milestone and Anduril's next contract award. But the more significant development is what this investment structure signals about the Pre-IPO market for "frontier AI" companies.
We're seeing the emergence of a two-tier system in private markets. Tier one consists of companies with genuine revenue and clear paths to profitability—they command high multiples but with some fundamental support. Tier two consists of narrative-driven companies with transformative potential but no current earnings—they command high multiples based entirely on option value.
Neuralink sits firmly in tier two. Anduril is transitioning from tier two to tier one. MSX's decision to bundle them in the same fund suggests either a sophisticated understanding of their complementary risk profiles or a narrative-driven approach that prioritizes story coherence over fundamental analysis.
Based on my experience auditing the 2020 DeFi composability crisis—where I calculated systemic risk when over-collateralized loans became highly correlated—I see a similar pattern here. The correlation isn't in the underlying technologies; it's in the market's willingness to pay for "AI frontier" exposure. When that willingness contracts, both positions will draw down together, regardless of their fundamental differences.
The Positioning Play
For investors watching this space, the takeaway isn't about Neuralink or Anduril specifically. It's about the structure of frontier AI investing in the current cycle. The Pre-IPO market is pricing in successful outcomes that are still years away and contingent on variables that aren't fully within these companies' control.
The smart positioning isn't to chase these narratives at current valuations. It's to identify where the value chain will be disrupted next—the component suppliers, the enabling technologies, the adjacent markets that emerge as brain-computer interfaces and autonomous defense systems move from experimental to operational.
Cross-border payments are evolving in similar ways. The infrastructure layer—settlement, clearing, compliance—is where value will consolidate as frontier technologies mature. The same logic applies to AI: the picks-and-shovels of neural interfaces and defense AI are where sustainable value creation will occur, not in the headline-grabbing applications themselves.
The bubble may not burst; the lessons will remain regardless. What matters is whether you're positioned for the maturation phase, not the hype cycle. MSX is making a bet on the maturation phase. The question is whether the timeline matches the valuation.