In short: blockchains upgrade without stopping by shipping new rules in node software that stay inactive until enough of the network has upgraded, then switching them on at an agreed point. Hard forks change the rules in a way older software rejects, soft forks stay backward-compatible, and systems such as Solana‘s feature gates activate changes one at a time, testnet first, once enough stake runs compatible software. The key points:
- No single party can force an upgrade; operators must choose to run new software.
- A contentious hard fork can split a chain into two coins, as with Bitcoin and Bitcoin Cash.
- Staged rollouts and testnets catch bugs before real money is at risk.
- In Australia, coins received from a chain split have tax consequences when you sell them.
Solana’s validator documentation from Anza covers how feature activation works, Ethereum‘s roadmap shows how a major network sequences upgrades, and the ATO’s guidance on crypto asset investments covers how forked coins are taxed.
Updating a normal app is simple, push a new version, users download it, done. A blockchain can’t work that way. Thousands of independently-operated validators and nodes need to agree on exactly the same rules at exactly the same moment, or the network risks splitting into two incompatible versions of itself. Understanding how blockchains actually manage this problem, coordinating an upgrade across a decentralised network that can never simply be switched off, reveals some of the more elegant engineering happening quietly underneath every major chain.
The Core Problem: Nobody’s in Charge
On a centralised system, an upgrade is straightforward because one party controls every server. A blockchain deliberately has no such single point of control, which is exactly the property that makes it valuable, but it also means an upgrade has to be voluntarily adopted by thousands of independent operators before it can safely take effect. If even a meaningful minority of nodes refuse to upgrade while the majority moves ahead, the network can split into two separate, incompatible chains, each with its own version of history from that point forward.
Hard Forks vs Soft Forks
Understanding the difference between a hard fork and a soft fork is the essential starting point here. A soft fork tightens the rules in a backward-compatible way, nodes still running older software continue recognising new transactions as valid even though they don’t understand the new feature themselves, letting the upgrade roll out gradually without forcing every single node to update immediately. A hard fork is a more fundamental break, introducing new rules that old nodes can’t validate, meaning any node that doesn’t upgrade gets left behind on a separate, incompatible chain once the fork activates.
When a Hard Fork Creates a New Coin
Bitcoin Cash is the clearest real-world example of what happens when a hard fork doesn’t achieve full community consensus, a meaningful portion of the network disagreed strongly enough about a proposed change that the chain split into two, each continuing forward as its own separate, independently valued asset. This is the most extreme possible outcome of a failed coordination attempt, and it’s exactly why most major chains have moved toward more careful, phased activation mechanisms designed specifically to avoid this kind of contentious split wherever possible.
Solana’s Feature Gate System
Solana’s specific approach to this problem, called feature gates, is a well-engineered example worth understanding in detail. Rather than forcing every validator to upgrade at one fixed moment, a new feature is built into validator software but stays dormant, inactive, until a feature authority formally requests its activation by submitting a transaction that creates a dedicated feature account onchain. The runtime checks for pending activation requests at the start of each new epoch, and only activates a feature once a sufficient majority of the network’s stake is already running software capable of supporting it.
Why Testnet Comes First
Solana’s feature activation process is deliberately cluster-agnostic, meaning a new feature typically activates first on Testnet, then Devnet, and only finally on Mainnet, once confidence has been built through real-world testing on lower-stakes networks. This staged rollout matters enormously, catching a bug or an unintended consequence on Testnet costs essentially nothing, while the same issue surfacing on Mainnet after real economic value is already flowing through the new feature could be costly. Generally, only one feature gate is activated at a time per network specifically to keep this process manageable and to make isolating the cause of any unexpected issue considerably simpler.
The Version Floor Concept
As each feature gate successfully activates, Solana raises what’s called the version floor, the minimum software version a validator needs to be running to keep participating meaningfully in the network. This gives the ecosystem a clean, verifiable way to track exactly which features are considered safely and fully adopted across the network at any given time, rather than leaving that as an ambiguous, informally-tracked state that different participants might interpret differently.
Kaspa’s Own Phased Approach
This same underlying philosophy, careful, staged rollout rather than one disruptive flag-day switch, shows up across other major chains too. Kaspa’s roadmap, which raised the network to 10 blocks per second with its Crescendo upgrade in 2025 and plans programmability through a later upgrade called Toccata, reflects the same deliberate sequencing, shipping and stabilising one major capability fully before building the next one on top of it, rather than attempting to launch a fully-featured, maximally complex network all at once from day one.
User-Activated Soft Forks
Not every network upgrade is driven primarily by developers or a formal feature authority. A user-activated soft fork is a coordination mechanism where node operators and users themselves signal support for a proposed change directly, and once enough of the network’s participants demonstrate genuine support, the change activates even without unanimous agreement from the core developer team or miners specifically. This mechanism exists precisely because blockchain governance isn’t always top-down, sometimes the community itself needs a credible way to drive a change forward when other coordination channels have stalled.
Governance’s Role in Deciding What Gets Activated
Before any feature gate or fork gets proposed for activation at all, someone needs to decide it’s worth building in the first place, and this is where onchain governance tokens often play a role, letting holders vote on which protocol changes a project should prioritise. Understanding this earlier stage of the process, the actual decision of what to build and in what order, matters just as much as understanding the technical activation mechanism itself, a well-engineered feature gate system is only as valuable as the governance process deciding what gets fed into it.
What Actually Happens if Coordination Fails
Understanding how a 51% attack works provides a useful, if extreme, contrast to voluntary upgrade coordination, both scenarios involve a meaningful share of network participants acting in a way that diverges from consensus, but a failed upgrade is a coordination breakdown among honest, well-intentioned participants disagreeing on direction, while a 51% attack is a malicious actor deliberately exploiting majority control. Keeping this distinction clear matters, a contentious hard fork like Bitcoin Cash’s split wasn’t an attack on Bitcoin, it was a legitimate, if disruptive, disagreement about the network’s future direction playing out through the same underlying fork mechanism a malicious actor could theoretically misuse for very different purposes.
Running your own full node is one way to verify exactly which rules a network is operating under at any given moment, rather than simply trusting a third party’s summary of what’s currently active. For anyone deeply invested in a specific chain’s long-term integrity, understanding how to check this independently is an useful skill, particularly around a period of active, ongoing protocol upgrades.
How This Connects to Trading Around Upgrade Events
Major, well-publicised network upgrades often coincide with periods of heightened price volatility, as the market prices in both the anticipated benefits of the upgrade and the genuine uncertainty around execution risk. Understanding how to protect yourself from volatility is worth doing specifically around a major, scheduled upgrade event, and applying sound position sizing rather than dramatically increasing exposure purely in anticipation of an upgrade’s success is a sensible discipline, upgrades that look straightforward on paper can still encounter unexpected delays or technical issues once real-world conditions are involved.
Being aware of your own confirmation bias around a project you’re already invested in matters here too, it’s easy to interpret every incremental testnet milestone as confirmation an upgrade will definitely succeed on schedule, when genuine, unbiased research would weigh both the progress made and the real technical risk still remaining before the upgrade fully activates on the main network.
Comparing Upgrade Philosophies Across Major Chains
Different projects have made different tradeoffs in how cautiously versus how quickly they push new features live. Comparing Bitcoin’s and Ethereum’s historically differing approaches to protocol change is a useful reference point, Bitcoin has generally favoured a slower, more conservative upgrade cadence prioritising stability above all else, while other major chains have been willing to move faster and accept somewhat more execution risk in exchange for shipping new capability sooner. Neither approach is objectively correct, understanding which philosophy a specific project has actually adopted, and whether that matches your own risk tolerance as a holder, is the more useful question to ask.
Why This Matters for Anyone Holding the Asset
For a holder rather than a validator or developer, phased activation systems matter because they directly reduce the risk of holding an asset through a major protocol change. A well-executed, gradual rollout with extensive testnet validation meaningfully lowers the odds of an unexpected chain split, a stalled network, or a critical bug reaching Mainnet with real capital already flowing through the affected feature. Applying genuine DYOR to any project you’re holding through a major upgrade, specifically checking whether it uses a careful, staged activation process like this rather than a single, high-risk flag-day switch, is an useful signal of engineering maturity worth factoring into your own risk assessment.
Tax Implications of a Hard Fork
If a hard fork results in you holding a new asset, the way Bitcoin Cash holders found themselves holding both BTC and BCH after that split, it’s worth understanding how hard forks are taxed in Australia, since receiving a new asset through a fork event carries its own tax consequences. Under ATO guidance, if you hold crypto as an investment, coins received from a chain split are not treated as income when you receive them, but they have a nil cost base, so the full sale price is a capital gain when you dispose of them.
How This Relates to Layer-2 and Rollup Upgrades
The same underlying coordination challenges apply, in a somewhat different form, to layer-2 networks upgrading independently of their base layer. A rollup or sidechain generally has more centralised control over its own upgrade process than a fully decentralised layer-1, which can make coordination faster, but it also means holders are trusting a smaller, more concentrated set of operators to manage that process responsibly rather than relying on the broader, more distributed validation Solana’s feature gate system or Bitcoin’s node network provides. Understanding this tradeoff is worth doing before assuming every “upgrade” across the crypto ecosystem carries an identical risk and coordination profile.
Smart Contracts and Upgrade Risk
For projects with a live smart contract layer, a base-layer protocol upgrade can sometimes introduce unexpected interactions with existing deployed contracts, particularly if the upgrade changes how transactions are processed or validated at a fundamental level. This is exactly why a proper smart contract audit process, and genuine bug bounty programs, matter so much around any major network upgrade specifically, catching an unexpected interaction before real funds are exposed to it is considerably cheaper than discovering the problem after the fact.
Researching a Project’s Upgrade History
Before holding a position through any major announced upgrade, reviewing a project’s actual track record of past upgrades, did previous ones land on schedule, did they require emergency patches shortly after activation, is an useful part of broader DYOR and altcoin research practice. Reading a project’s own whitepaper and technical documentation directly, rather than relying purely on marketing material describing an upcoming upgrade in only the most favourable terms, gives a considerably more grounded picture of genuine execution risk.
Understanding a project’s market cycle position and applying sensible portfolio diversification, rather than concentrating heavily around a single anticipated upgrade event, remains sound practice regardless of how confident the surrounding narrative feels in the lead-up to activation. Thinking through an appropriate investment time horizon also matters, a major protocol upgrade’s true, long-term impact on a network’s usefulness and adoption often only becomes clear well after the initial activation headlines fade.
Key Takeaways
Whether the network is Bitcoin, Ethereum or Solana, upgrading a blockchain safely is a harder coordination problem than updating ordinary software, thousands of independent, voluntarily-participating operators need to move together without a single party able to force the change through directly. Mechanisms like Solana’s feature gate system, with its epoch-based activation, staged testnet-to-mainnet rollout, and formal version floor tracking, represent real, considered engineering solutions to this problem, built specifically to avoid the kind of contentious, damaging chain split Bitcoin Cash represents.
Understanding how a project actually manages its own upgrade process, rather than assuming all blockchain upgrades carry equal risk, gives an useful lens for evaluating engineering maturity across DePIN networks, exchange platforms and every other project you choose to hold, particularly through any period of significant protocol change.