The “Scale Now” Dilemma 

By Greg Hudson, MD at Ingenuity Cloud Services

In theory, scaling compute in 2025 should be frictionless. The cloud promises near-infinite resources, ready to deploy on demand, but for modern CTOs, the reality is far more complex. The central dilemma for technical decision makers like myself is the need to navigate a political, governance, security and cost landscape vs. selecting the right technology to appease without stifling innovation and business growth.

As digital services grow – especially in latency-critical sectors like gaming, media, and VPNs – the pressure to deliver high-performance infrastructure across global markets has never been more intense. But scaling rapidly often introduces hidden trade-offs in efficiency, flexibility, and regulatory compliance. In this article, I am summarising the themes that keep coming up in conversations with our customers and that make us continuously look at how we better serve the markets we operate in.

1. Latency is the new uptime

For users, speed is everything. In cloud gaming, a few milliseconds of lag can ruin the player experience and in n blockchain/web3, every 10ms of latency adds huge increase cost of infrastructure. VPN users want to tunnel into remote regions without performance hits. The challenge is not just adding more compute but placing it strategically. Centralized clouds can be simpler but often introduce higher latency for faraway users.  Distributed deployments offer performance gains but are harder to manage.

Without a consolidated edge strategy and ultra-low-latency network backbones, scaling up is likely to worsen user experience – creating a false sense of readiness. Striking the right balance between centralization and geographic distribution is therefore a  core challenge, especially for companies in sectors such as game development, web3, VPNs.

2. Cloud Spend: The Hidden Crisis 

While provisioning cloud instances is easy, controlling costs has become a top concern. In fact, according to Flexera’s 2024 report, 82% of IT leaders in 2024 cited cloud cost management as their biggest headache. Elastic scaling leads to elastic bills, and on-demand flexibility comes at a steep premium.

For high-load applications – like 24/7 streaming services or online games – scaling on hyperscalers can quickly become unsustainable. This has led some companies to consider “cloud repatriation,” moving workloads back on-prem or to optimized hosting to regain control over costs.

3. One Size Doesn’t Fit All

Cloud providers offer fixed instance types, but modern workloads are far from generic and as we’ve noticed in the markets we operate. A machine learning app might need GPUs; a game server demands fast CPUs and low jitter; a streaming platform thrives on high-bandwidth encoders.

When infrastructure can’t match workload needs, performance suffers – or teams waste effort adapting software to fit the hardware. The future lies in bespoke infrastructure – purpose-built, workload-optimized environments that match exact technical requirements. But most providers don’t offer that level of customization.

4. Compliance as a Scaling Bottleneck

If your cloud vendor doesn’t offer certified regions or local availability zones, your roadmap hits a wall. For developers, this means managing multiple deployment environments and region-specific configurations; for CIOs, it means juggling vendor lock-in, legal risk, and fragmented architecture. Compliance isn’t just a legal checkbox – it’s a core design constraint that complicates multi-cloud strategies and undercuts the “scale anywhere” promise of the cloud.

Examples can range from U.S.-based media platforms expanding into Europe that must ensure GDPR compliance, or which requires processing and storing user data in EU-based, certified data centres. Similarly, any healthcare application targeting U.S. users must leverage HIPAA-compliant infrastructure – often narrowing the list of eligible cloud providers.

5. Expect the Unexpected Surges

Modern digital platforms live and die by their ability to handle surges. A new game release, a global event, or a geopolitical crisis can send traffic through the roof. And while autoscaling works in theory, scaling massive infrastructure – say, 1,000 GPU nodes – is rarely instant.

Systems that can’t scale now during spikes risk failure at critical moments. But overprovisioning “just in case” is financially wasteful. This paradox is particularly acute in gaming, streaming, and VPNs – industries where performance, not just availability, is mission-critical. 

The Need for a Smarter Path

The “scale now” dilemma boils down to this: today’s cloud providers offer quick fixes, not deep solutions.

Hyperscalers deliver global reach but can be expensive and inflexible. What organizations need is scaling without compromise. That means deploying infrastructure where it’s needed most, optimizing for latency and cost, meeting compliance requirements by design, and tailoring environments to the unique demands of each workload – from GPU-heavy gaming clouds to compliance-sensitive healthcare stacks. Organizations no longer want just more capacity; they want the right capacity, at the right time, in the right place. If we can deliver this level of precision, flexibility, and performance, we will keep defining the future of cloud – because in 2025 and beyond, scaling smart is the only way to scale sustainably. 

About Greg Hudson  

Greg is an experienced technology and commercial strategy leader, passionate about solving customer problems and creating value for the organisations he works with.

Ager more than 18 years’ experience in the technology and IT sector, Greg leads the Cloud Services division at THG Ingenuity and oversees the product and technology functions, ensuring strategic investments and growth roadmaps bring profitable growth.