The Human-Battlefield Arbitrage: Why MSX's Pre-IPO Fund is Betting on Neuralink and Anduril's Shared Curve
A newly surfaced filing reveals that the MSX Pre-IPO III fund has taken direct positions in two companies that could not be further apart on the technological map: Neuralink, the brain-computer interface developer, and Anduril, the defense AI platform. On paper, this is a hedge. One company builds a chip that reads the brain's electrical whispers; the other builds software that decides when to kill. But the deeper logic is not diversification. It is a wager on the same curve: the point where AI stops being a tool and starts becoming the infrastructure of the physical world.
Let me start with the numbers, because that is where the story actually begins. Anduril's revenue trajectory is the closest thing the defense sector has to a software startup. From the data I have seen, the company is on track to generate $1 billion in revenue in 2025, up from roughly $400 million a year earlier. That is a 150% growth rate, sustained for two consecutive years. Meanwhile, Neuralink has generated zero revenue. The company is in the FDA's IDE (Investigational Device Exemption) stage, which means it can only use its device for clinical trials. The company has a valuation rumored between $5 and $10 billion. That is not a startup. That is a financial instrument. And the instrument is betting that Neuralink can execute on a timeline that the FDA has never actually endorsed.
This is where my background as a financial analyst kicks in. When you look at Pre-IPO investments, you are not buying a company. You are buying a sequence of binary events: clinical trial data readouts, regulatory approvals, and contract announcements. Anduril's event sequence is essentially being driven by a global geopolitical vector. The Pentagon is under pressure to find alternatives to traditional prime contractors, and Anduril's Lattice software stack has become the default operating system for the US military's autonomous drone programs. The company has locked in a multi-year contract with the Pentagon that values its software at over $500 million, which is a strong signal for an exit window around 2026 or 2027.
Neuralink's sequence is more fragile. The company has received FDA's 'breakthrough device' designation, but that is not an approval. It is a priority ticket. The CONVOY study, which links the N1 implant to a robotic arm, is the first real test of whether the technology can move from cursor control to meaningful motor function. From my analysis of the technology's hardware, the 1,024 channel electrode is the only true differentiator that a Synchron (which uses a stent-based approach) or a Precision Neuroscience (which uses a surface array) cannot easily replicate. The issue is not the electrode. The issue is the surgical robot. The R1 robot's ability to avoid blood vessels has a very short margin of error, and we do not have long-term data on electrode degradation. The company has only three patients, and if any of them experiences a device failure that leads to an explant, the regulatory clock resets.
The contrarian angle that no one is talking about: this is not a bet on the technology. It is a bet on the regulatory state. In the US, the FDA has shown a willingness to accelerate AI-based medical devices. But the agency is simultaneously under pressure from the Government Accountability Office (GAO) regarding its AI auditing procedures. If the FDA slows down, the Pre-IPO window closes, and the fund's exit moves from 2027 to 2029. Anduril's path is different but equally dangerous. The defense sector is not driven by innovation; it is driven by appropriations. If the geopolitical situation calms, Anduril's growth will slow. The company's valuation is currently 14 to 20 times revenue, which is a premium to public defense primes like Lockheed (at 1.5 times), but it's trading like a tech company. The market is already pricing in a future where defense is software-defined. That may not be true.
I have seen this playbook before. In the 2020 DeFi yield farming era, I built spreadsheets that tracked token emissions against actual protocol revenues. The conclusion was that 80% of those tokens were pure liabilities. The same logic applies here. Anduril's revenue is highly concentrated in US government contracts. If the contract terms are renegotiated under a new administration, the growth rate drops. Neuralink has no revenue, so its valuation is a pure option on the future. The entire MSX thesis is an option on a future where the FDA and the Pentagon both move faster than they have historically.
My assessment is that this is not a high-risk bet. It is a smart, asymmetric bet. The market has not yet figured out that Neuralink and Anduril are on the same side of the economic curve. Both are pushing AI from a data center to a physical body. Both are solving the same problem: how to make an AI decision-maker fast enough to act in milliseconds. In the battlefield, that means missile interception. In the brain, that means motor control. The same latency problem exists in both. In DeFi, we called this 'oracle latency.' In defense, they call it 'sensor-to-shooter time.' In neuroscience, it is 'neural signal processing delay.' It is the same engineering problem, and the companies that solve it will be the ones that capture the value.
The question that nobody is asking: if this is true, why is there no overlap between the companies' technical talent? Neuralink has over 100 engineers with a background in neuroscience, and Anduril has a large team of defense industry veterans. The cognitive gap is a signal. The acquisition is not happening. It is a portfolio-level bet on the category, not a company-level bet on a merger.
Here is what I am watching. In the next six months, two things will tell you whether this thesis is valid. First, Neuralink's CONVOY study will release preliminary data on the robotic arm performance. If the data shows a 90% accuracy in neural decoding, the company is ahead of schedule. Second, Anduril will announce a new contract with a non-US allied government. That would be the first proof that the Lattice OS is expanding beyond the US defense budget. If both of those happen, the Pre-IPO valuation will be justified. If either fails, the market will reprice the entire 'AI edge' narrative.
I have been through the 2017 ICO cycle and the 2020 DeFi summer. The pattern is always the same: the narrative arrives before the technical proof. The question is not whether the technology works. It is whether the proof arrives before the next narrative shift. For Neuralink, the proof is a clinical trial. For Anduril, the proof is a contract. For the fund, the proof is an exit. All three are on the same clock.