Beyond the Cloud: Building Resilient Web3 Infrastructure, interview with Pauline Shangett, Chief Strategy Officer at Changenow

In this interview we sit down with Pauline Shangett. She is the Chief Strategy Officer at Changenow and Strategic Advisor at NowNodes, to discuss her recent speech on WebX entitled “Head in the Clouds: Is Hardware the key to sustainable web3 infrastructure?”.

Pauline Share her perspective on the changing landscape of Web3 infrastructure, the myths around cloud versus hardware, and why real resilience is more about people, processes and strategy than just technology.

In the following, the CSO on Changenow Share her insights about some of the most urgent matters for web3 and then.

Pauline, in your webx -keynote, you started with a striking rule: “What is the scary for a CTO? The hack? No, it is when everything fails without warning.” Can you explain what you meant by that?

Absolute. When we talk about “scenarios” in web3, the heads of most people immediately go to hacks, exploits or malicious actors. And yes, they are terrifying. But in reality, the situations that shake teams to their core are often much simpler and everyday: one fire, one missed update, one overloaded end point and suddenly your entire product is offline.

This is not theoretical care. We have seen that large platforms have not been brought down due to advanced cyber attacks, but due to a power outage, a defective cable or a wrongly configured failover system. And when that happens, you don’t just lose the uptime. You lose user confidence, transaction volume and in some cases your reputation.

That is why I framed my keynote as a ‘reality check’. I am not a CTO, I don’t write a code every day, but I constantly talk to teams, founders, infrastructure pipes and chain developers. And what I have learned is this: In web3 your greatest vulnerability is often the one you never expected, the one outside your attack area, but in your operational risk model. Infrastructure is not interesting until it fails. And then the only thing that matters is.

Let’s talk about Cloud. Everyone knows the benefits such as scalability, speed, ease of use. But in your speech you argued that people overlook the security angle. What did you mean?

The story about the cloud has always been: “It’s easy, it’s fast and it scales.” And that’s true. But when people leave the cloud, the most common justification is security. They will say: “I don’t want my nodes or infrastructure to be controlled by a centralized provider that can censor, cut or expose me to surveillance.”

There is truth in that care. Centralization in the infrastructure layer introduces risks. But ironically I see that when people dump the cloud for ‘safety’, they completely underestimate a different type of risk: physical risk.

Think about it: the greatest power of the cloud is not only elasticity, it is redundancy. If an AWS region goes down, there are several layers of fallback. When you host your hardware in one facility, you don’t have that safety net. And that is where people can be blinded.

You gave the example of the Kakaotalk Data Center Fire in South Korea, which paralyzed entire services, including Upbit. Why is that case so important for the web3 industry?

Because it demonstrates something fundamentally: failure does not have to be malignant to be catastrophic. When a fire broke out in only one data center in 2022, frozen services throughout the country.

It was not a hack. It was not ransomware. It was smoke. Yet the consequences were huge, users could not log in, transactions were blocked and the government had to intervene. That is a disruption at national level caused by a single failure point.

In Crypto we often talk about ‘crypto -winters’ in terms of market decline. But I think the more urgent crypto -winter is operational: wars, floods, fires, cut cables, rolling black -outs. Those are not “border cases”. They are part of the world in which we live. And if you don’t plan for them, you are essentially guessing with your infrastructure.

So how does Nownodes approach resilience differently? What does “planning of failure” look like in practice?

Bee NOWNODESOur philosophy is very simple: do not ask if something will go wrong – ask when. Because it will. That is the only certainty in infrastructure.

Our systems are deliberately spread over several regions: the EU, the US and Asia, with physical presence in countries such as Germany, Finland, the Netherlands, the United States and Singapore. That is not just a check box for compliance. It is a survival strategy: placing junctions where they can withstand political, geographical and technical risks.

We also work on a 2N+1 architecture. That means that for every critical component such as strength, calculation, network, we do not only have one backup. We have two, plus a reserve. So if one system fails, the traffic shifts immediately. If the backup also fails, the reserve will take over. It is a layered safety net.

And we don’t just trust the system blindly. We regularly perform failover simulations. We deliberately conclude systems in mirrored environments to see what breaks. We do stress tests, regional tests, even attack simulations. Because you don’t want you to test the resilience when a real crisis happens.

For years, Cloud was considered the cheaper option. But you have suggested that the comparison is changing. Can you lead us through it?

Five years ago, Cloud was the obvious choice. You avoid a massive preceding capex, you only paid for what you used and the scales felt effortless. But that story has shifted dramatically.

Nowadays the “Big Three” cloud providers, such as AWS, Google Cloud, Azure, the market. And as happens in every near-monopoly, price price trends up. AWS calculation costs rose, for example, by more than 20% in just one year. Almost 40% of the companies reported the cloud defenses that spiked by more than 25% in the past 12 months by more than 25%.

In the meantime, hardware has become more predictable. Yes, you pay more on the first day – servers, racks, strength. But when you spread that investment for 7 – 10 years, the economy flips. One engineer calculated famous that a $ 1,100 server costs around $ 110 a month for a decade. Compare that with $ 2,000 – $ 7,000 a month for equivalent sources in the cloud, and mathematics speaks for itself.

And after the costs, hardware gives your freedom. You are not limited to the functions and configurations that your cloud provider offers. You can patch, tweak and implement how you need it. That level of control can be the difference between flexible scale and growth with bottlenecks.

But even with great hardware or a strong cloud setup, what if your provider simply fails at a crucial moment?

That is exactly the point. Neither Cloud nor Hardware will save you if your provider haunts you at 3 o’clock in the morning. Infrastructure is only as reliable as the people behind it.

At Nownodes, when we ask our customers why they choose us, their answers rarely have to deal with the details of our servers or our architecture. Instead, they talk about the human side, the fact that we respond within minutes, that we are seamlessly scales without hidden billing relationships, that we support more than 115 block chains, including the less obvious, and that when their RPC crashed at two o’clock in the morning, our team was repairing it in real time.

That is the reality: infrastructure is just as good about trust and responsiveness as about technology.

Let’s comment on the details: back -ups, multichain range and support. What makes your model different?

Backups first. Junctions fail. Freeze chains. Updates break compatibility. The real question is: when a disaster strikes, what condition do you recover? Last week? Last month? We have geo-distributed backups so that your “worst day” becomes a small bump, not a total blackout.

On Multichain range, most providers fill with 50-70 block chains. Only a handful, maybe 3-5, support 100+. We support 115+ chains and we constantly add more. It is important that we do not only support the trendy chains. For example, we are the only provider that offers an infrastructure with shared nodes for some of the most complex and overlooked block chains: Monero, Ecash, Nano and more. Because your users cannot wait for you to ‘support their possession’. They leave.

Finally support. We do not believe in chatbots or endless ticket uses. So our customers get real engineers in their telegram or weak. Average response time: under three minutes. Resolution time: hours, not days, not even for deep technical bugs. And that is not a premium upsell. That is our basic line.

Price models in infrastructure can be notoriously opaque. How do you approach transparency?

Most RPC providers rely on complex layers, hidden gas valves and surprise costs. One day everything looks good, the next day you have crossed an invisible threshold because a botnet distributed your traffic, and suddenly triples your account. Or even worse, they just broke you off.

We have followed a very different approach. Our prices are clear, based on subscription and predictable. You always know exactly what you pay for. If you need more capacity, scales are fast, reasonably priced and transparent. There is no hostage negotiation.

That predictability is one of the main reasons why our customers stay with us. Because there is uncertainty everywhere in web3 – in the markets, in regulation, in adoption. The last thing you want is uncertainty in your infrastructure account.

So back to your original question: Is hardware the key to sustainable web3 infrastructure?

No. And neither is cloud. The real key is resilience.

Resilience comes from smart backups, distributed systems, people-oriented support, transparent prices and real Multichain range. It’s not something you rent. It’s something you build.

Infrastructure is boring until it is not. Until your end point falls, your TVL will disappear, your users will stop and say your logs nothing. That is when you realize that the infrastructure is more than servers. It’s trust. It is the silent contract between your product and the people who keep it alive.

Last thoughts? What should Web3 teams take away from your message?

I would say this: stop treating infrastructure as a side issue. It is the foundation of your product. You can have the best user interface, the smartest tokenomics and the most loyal community, but if your infrastructure fails, it doesn’t matter.

Don’t ask: “How do I save money this month?” Question: “How do I survive the crisis I haven’t seen yet?” Because it’s coming. The teams that will still be here in five years are not those who cut the corners. They are the ones who build resilience from the first day.