While the market sleeps, the ledger does not lie. And this week, the ledger reveals a fund manager placing bets on the two most extreme frontiers of artificial intelligence. MSX's Pre-IPO Fund, in its third round, has secured allocations in Neuralink and Anduril. This isn't a diversified portfolio. It's a thesis on the endgame of AI—one targeting the biology of the human mind and the machinery of modern warfare. The message is clear: the next phase of AI investment is not in software that writes code, but in systems that read neurons and control the battlefield.
The Context: An Unlikely Pair
On paper, these companies could not be further apart. Neuralink, under the controversial leadership of Elon Musk, is threading hair-thin electrodes into brain tissue to decode neural signals. Their goal is to bridge the gap between the organic mind and digital computation. Anduril, on the other hand, has built a fortress around the Lattice software platform—an AI-powered operating system for defense that connects drones, submarines, and missile warning systems into a cohesive, autonomous network. One company is fighting to interface with the human mind; the other is building the autonomous machinery of war.
The market sees these as two separate "AI" investments. That is the surface-level reading. The deeper reality, visible to anyone who has spent time tracking capital flows in this sector, is that MSX is not buying two companies. They are buying a single, overarching narrative: AI is penetrating the biological and physical worlds. The in-between layers—the digital world of text, code, and data—are already saturated. The new alpha is at the edge.
The Core: Two Technologies, One Direction
The technical specifics are where the strategy reveals its substance. Neuralink is currently in its POC-to-production transition. They have successfully implanted their N1 chip in three human patients. The device is a high-density flexible electrode array with 1,024 channels, inserted by the R1 surgical robot, which is designed to avoid vascular damage. In November of last year, they received FDA approval to begin the "CONVOY" feasibility study, which will test whether a patient with a spinal cord injury can control a robotic arm with their thoughts. The technical goal is simple to state but difficult to execute: achieve high-bandwidth signal acquisition and minimal invasive impact.
Anduril is further ahead in the production lifecycle. They are already generating significant revenue, an estimated $1 billion in 2025, with roughly 150% year-over-year growth. The architecture is a combination of the Lattice AI platform and heterogeneous hardware. Their recent acquisition of the UK-based 3D printing drone company Callum for $1.3-1.4 billion shows they are building their own supply chain to create cheap, expendable drones. Their December 2025 contract with the US Department of Defense's innovation unit, valued at $250 million for "large-scale expeditionary operations," is a signal that the Pentagon views software-defined defense as the future.
The commercial model is the inverse of traditional tech. Neuralink is a "medical-first, consumer-later" play. They are walking through the FDA's Investigational Device Exemption (IDE) phase, aiming for commercial approval in the 2026-2027 timeframe. Anduril is a pure B2G play. They are not selling to consumers; they are selling to the US and its allies. The revenue streams are locked in by government contracts. This is the Pre-IPO logic: get in before the public markets wake up.
The Contrarian Angle: The Blind Spot in the Narrative
The consensus interpretation is that MSX is buying into the two most exciting frontier tech stories. The contrarian view—the one I find more compelling—is that they are buying into a highly probable liquidity squeeze. The problem with these frontier bets is not the technology; it's the financial engineering around it. Neuralink is valued between $8 and $10 billion on its next round. That's a massive number for a company with zero revenue and a handful of implants. Anduril is looking at a $25-30 billion valuation. Their P/S ratio is around 14-28x, which seems reasonable against Palantir's 30-50x, but this ignores that Palantir has a diversified software revenue stream. Anduril is a hardware company with a software coat of paint.
There is a fundamental mismatch in the "Pre-IPO" liquidity narrative. The exit path for these funds is typically 2-3 years. Neuralink is targeting an IPO in 2027-2028. Anduril is targeting 2026-2027. If the FDA delays Neuralink's approval—and the FDA is conservative—that timeline slips. If the US defense budget is cut, Anduril's contract flow slows. The market narrative is priced for a frictionless, on-schedule exit. The chain is built on the assumption of human and governmental efficiency. That is a fragile foundation.
Furthermore, the "bio-tech" and "defense-tech" segments have a fundamental flaw that the market ignores: the ethics of the underlying product is a regulatory risk. Neuralink is facing the issue of electrode longevity and potential signal degradation, which could lead to forced re-operation. Anduril is facing the shadow of lethal autonomous weapons (LAWS). The international community has not yet reached a consensus on this issue, but the pressure is building. If any single major event, such as a serious adverse effect on a Neuralink patient, occurs, the stock price is not the only thing that will be affected; the entire Pre-IPO round's viability will be questioned.
The Takeaway: The Next Watch
This is not a story about two companies. It is a story about the "AI next frontier" narrative being priced in. The strategy is clear, but the execution is complex and potentially delayed. The next watch is not the quarterly P&L of the companies, but the FDA's response to the CONVOY trial and the speed of the DoD's next contract. If the FDA approval comes in line with the timeline, and if the defense contracts continue to flow, these bets will pay off. If the machine tries to scale before the human body and the legal system are ready, the ledger will show the true cost of a frontier that is not yet ready to be conquered.
