Building resilient Web3 infrastructure in an unstable world: Interview with Pauline Shangett, Chief Strategy Officer at ChangeNOW

Infrastructure rarely makes headlines in ￰0￱ without it, no DApp, wallet, or exchange could survive a single ￰1￱ sat down with Pauline Shangett, Chief Strategy Officer at ChangeNOW and Strategic Advisor at NOWNodes, to discuss why resilience is more important than the cloud-versus-hardware debate and how the next wave of blockchain adoption will depend on getting the foundations ￰2￱ and risk Pauline, you’ve said that in Web3, the scariest moment isn’t necessarily a hack, but the silent collapse of ￰3￱ do you see that as such a critical distinction? Because when you’re hacked, at least you know what happened – there’s an attacker, an exploit, and a narrative you can point ￰4￱ when your infrastructure fails unexpectedly, it feels like the ground is disappearing beneath ￰5￱ missed patch, one fire, and one overloaded endpoint, and suddenly your service isn’t hacked or censored; it’s just ￰6￱ a Web3 team, that’s ￰7￱ don’t care whether it was malicious or accidental; they just see ￰8￱ can’t move, transactions fail, and trust erodes instantly.

That’s why I emphasize this point: infrastructure may be invisible when it works, but it becomes the only thing that matters when it doesn’t. The industry tends to romanticize hardware as the “decentralized” alternative to the ￰9￱ in your talk, you cautioned against seeing hardware as a silver bullet. Why? Because decentralization doesn’t magically protect you from real-world ￰10￱ threats to hardware are often physical, not ￰11￱ at South Korea in 2022: a single fire in a KakaoTalk data center brought down payments, logins, and even access to one of the country’s largest exchanges, ￰12￱ wasn’t a ￰13￱ was smoke in a server ￰14￱ case is important because it shows how fragile “single points of failure” can ￰15￱ these aren’t rare freak accidents; they’re part of a broader landscape of physical risks: blackouts, floods, cable cuts, and political ￰16￱ things are happening more often, not ￰17￱ if your Web3 infrastructure is built on the assumption that the physical world is stable, you’re setting yourself up for ￰18￱ resilience So how do you actually prepare for instability?

What does resilience mean in practice for NOWNodes? For us, resilience starts with ￰19￱ operate across multiple strategic regions – Europe, North America, and Asia, with physical presence in countries like Germany, Finland, the Netherlands, the U. S., and Singapore. That’s ￰20￱ don’t want all your eggs in one basket, whether that basket is political, geographic, or technical.

Second, it’s about ￰21￱ infrastructure follows a 2N+1 model: for every critical component, whether it’s compute, power, or network, we maintain two backups and an additional ￰22￱ if one fails, the system keeps ￰23￱ the backup fails, the spare takes ￰24￱ goal is continuity without the user ever ￰25￱ third, we constantly test ￰26￱ shut down systems in mirrored environments, we simulate attacks, and we cut regions offline on purpose, because the worst time to discover your failover doesn’t work is when you’re in the middle of an actual ￰27￱ has always been a deciding factor between cloud and ￰28￱ do you see the economics shifting today? The cloud used to be the obvious financial winner; you avoided upfront investment, you only paid for what you used, and scaling felt ￰29￱ now, the reality is very ￰30￱ big cloud providers like AWS, Google, and Microsoft control the market, and with limited competition, prices only move upward.

We’ve seen compute costs rise by more than 20% year-over-year, and companies reporting surprise bills jumping 25% or more in a single cycle. Meanwhile, hardware is becoming more predictable. Yes, you need to invest upfront, but when you stretch that investment over seven to ten years, the math actually favors ownership. A server that costs $1,100 today ends up being about $110 a month over a ￰31￱ equivalent in the cloud?

Two to seven thousand dollars per month. That’s a massive ￰32￱ beyond dollars, there’s ￰33￱ bare metal, you control your patches, your deployment, your ￰34￱ the cloud, you’re limited to what your provider decides you’re allowed to ￰35￱ makes sense technically, but what about the human factor? What happens if a provider doesn’t support you when it matters? That’s where most teams underestimate ￰36￱ the best hardware nor the most optimized cloud setup matters if your provider disappears when you need ￰37￱ our clients consistently tell us that their loyalty isn’t about the hardware we use; it’s about the way we ￰38￱ value that our engineers answer in minutes, not ￰39￱ trust that we can scale with them without surprise ￰40￱ appreciate that we support more than 115 blockchains, including the more complex ones that others ￰41￱ perhaps most importantly, they know that if something breaks at two in the morning, we’re online, fixing it live.

That’s not marketing. That’s the day-to-day reality of why they stay with ￰42￱ and pricing models are often pain points in Web3 ￰43￱ does your approach differ? In support, we’ve taken a hard stance: no chatbots, no endless ticket ￰44￱ you reach out to us, you’re talking directly to an engineer who can solve your problem in real time, whether it’s through Slack, Telegram, or another ￰45￱ average response time is under three ￰46￱ complex bugs, resolution happens in hours, not ￰47￱ pricing, we reject the “gotcha” model that’s so common in the cloud world, where invisible limits throttle you, or a sudden traffic spike turns into a tripled ￰48￱ subscriptions are ￰49￱ know exactly what you’re paying for, and scaling up is fast and ￰50￱ stability allows teams to plan for growth without fearing that their infrastructure costs will suddenly sink ￰51￱ future of a sustainable Web3 To wrap up, is hardware the future of sustainable Web3 infrastructure?

Or is the cloud still king? Honestly, ￰52￱ real future isn’t about choosing one or the other. It’s about ￰53￱ build resilience by combining smart backups, wide distribution, transparent economics, and human-centric ￰54￱ or hardware alone won’t save ￰55￱ ￰56￱ resilience isn’t rented. It’s built, tested, and earned over time.

That’s the mindset we bring to NOWNodes , and it’s the mindset I believe the entire industry needs if it wants to grow ￰57￱ thoughts for founders and teams building in Web3 right now? My message is simple: infrastructure is boring until it isn’t. But when it fails, nothing else about your product ￰58￱ your users can’t transact, if your RPC endpoints are down, if your service disappears without explanation – you’ve lost their ￰59￱ don’t treat infrastructure as background ￰60￱ it as the backbone of your ￰61￱ with the assumption that something will ￰62￱ it ￰63￱ teams that survive the next cycle of Web3 won’t be the ones with the flashiest UI or the best meme marketing.

They’ll be the ones whose infrastructure kept them online when the world threw them a curveball.