Today's news brought a problem that has nothing to do with prices and everything to do with whether a thing actually works. Sui, one of the seven cryptocurrencies held in the Autopilot portfolio, suffered its second major network outage in five months. An outage means the blockchain stopped processing transactions for a stretch of time, and the people who run it, called validators, had to coordinate an emergency fix to bring it back online. Nobody could trade, move funds, or use any application built on it until it was restored. This raises a question every investor eventually faces: when does a product defect stop being bad luck and start being a reason to sell?
First, the basics. A blockchain is a shared public ledger that records who owns what and who sent what to whom. A Layer 1, or L1, is a blockchain that runs on its own, settling its own transactions rather than relying on another network. Bitcoin and Ethereum are the two largest. Sui is a newer challenger L1, and its entire pitch to investors is performance: it claims to be faster and able to handle more transactions at once than its older rivals. That sales pitch is exactly why repeated downtime is so damaging. A chain that sells itself on speed and reliability cannot afford to keep stopping.
The framework for thinking about this is what professional investors call a deselection rule. Most investing frameworks ask "what should I buy?" A deselection rule asks the opposite and arguably more important question: "what would make me drop something I already own?" Writing that rule down in advance matters because in the moment, every problem feels explainable. Founders issue reassuring statements, the price often bounces back, and it is psychologically easy to forgive one bad event. A deselection rule protects you from your own tendency to rationalize.
The specific rule worth borrowing here comes from a simple piece of investigative wisdom often summarized as: once is an accident, twice is a coincidence, three times is a pattern. Applied to operational failures, one outage is a data point. It happens to almost every young network, even good ones; Solana, now a top cryptocurrency, suffered several outages early in its life and recovered its reputation. Two outages in a short window, like Sui's two in five months, is a warning that demands attention but not yet a verdict. A third would cross the line from misfortune into a thesis of deselection: evidence that the reliability problem is structural rather than a fluke, and that the original reason for owning the asset, its performance promise, is broken.
Today's data is why Autopilot is flagging Sui rather than selling it. Sui makes up about 10% of that portfolio, which is a meaningful but not dominant slice, and this is its second outage, not its third. The portfolio's rules also only allow allocation changes on its scheduled Monday check, not in reaction to a single weekend headline. So the disciplined action is to hold, raise the flag publicly, and set a clear line: a third major outage would trigger a full review of whether Sui still belongs in the book. That is the difference between reacting to news and responding to a pattern.
There is a broader lesson here for anyone investing in technology of any kind, not just crypto. Reliability is not a boring engineering footnote; it is a core part of the investment case. For an institution deciding whether to build on or buy into a network, uptime is as important as raw speed or how much money the network handles. A blockchain that is fast 99% of the time but unusable during the 1% of moments that matter most is, for serious users, no better than a slow one. The same logic applies to a cloud provider that goes down, a payments app that freezes on payday, or a factory that produces brilliant products except when it stops.
What would invalidate this caution? If Sui's team ships a credible, transparent fix that addresses the root cause, and the network then runs cleanly for a long stretch under heavy load, the two outages become history rather than a trend, much as Solana's early stumbles faded once it stabilized. Reliability problems can be solved, and a young network earning back trust can be a genuine opportunity. The point of a deselection rule is not to punish a single failure. It is to make sure that if the failures keep coming, you act on the evidence instead of on hope.