The True Cost of Cloud Scalability

Hey legend, I hope you’re living the dream.

Your SaaS is growing. Revenue’s up. Infra is “scaling as promised.” Then the cloud bill hits, and it doesn’t make sense.

Economies of scale aren’t kicking in. The bigger you get, the cheaper it should become to run your business per customer, and profits should go up.

And while you were promised infinite elasticity and efficiency, somehow performance is spiky, backups are duplicating across regions, and you just racked up $4K in egress fees.

Sound familiar? You’re not alone. 73% of SaaS businesses spend at least 6% of their revenue on cloud costs, and 31% spend 11% or more—much of which is waste.¹

But to me, managing cloud costs isn’t just about saving money. It’s about regaining leverage to compete and win in a competitive market.

¹CloudZero

We’ll cover three things:

1. Breaking down cloud costs and how to identify hidden cloud charges.

2. How we’re balancing performance and redundancy efficiently.

3. An approach to smarter scaling decisions that is working for us.

Was this email forwarded to you? Subscribe here

Cloud scalability is often praised for its flexibility and performance, although 50% of tech founders cite unexpected cloud costs as a major barrier to effective scaling.² Even more concerning is 74% of SaaS businesses overspend due to premature scaling.³

Scalability isn’t just stacking servers. It’s managing what you don’t see: waste, inefficiency, and surprises. Growth without discipline is just expensive chaos.

Here’s how it’s playing out in the arena:

• Forecasts are wobbling: unpredictable usage spikes are turning quarterly budgets into moving targets. Cloud spend can swing 10–15%, forcing CEOs to adjust plans mid-cycle to protect margins and hit profitability targets.

• Margins are eroding: with infrastructure spend climbing ~20% year-over-year⁴ and gross margins slipping by 3–6%⁵, CEOs are diverting cash from new feature development and go-to-market to cover infrastructure overruns.

• Roadmap velocity is decelerating: engineering teams spend more cycles on cost-optimization (rightsizing, instance tuning), slowing feature rollouts and lengthening time-to-market.

We’ve prioritized three areas to get our leverage back:

1. Embracing FinOps as a growth strategy.

2. Optimize what we already pay for.

3. Architecting cloud stacks for cost, not just scale.

²Hashicorp ³Duet Partners ⁴Gartner ⁵Spring Street Capital

This starts with treating cloud costs like a strategic lever.

• Tag everything and allocate by team: implement tagging systems that assign every cloud resource to a product, team, or customer unit. Then build dashboards that break spend down by owner and how it’s being used.

  • Why it matters: It shifts accountability from “someone in DevOps” to actual product and revenue owners. Teams can’t improve what they can’t see—and this visibility lets you tie spend to business value.

• Forecast and track like CAC: integrate tools to monitor key signals—logins, feature usage, and support trends. Create dashboards highlighting early warnings and automate alerts for your customer success team.

  • Why it matters: When cloud is our second-largest line item, guessing is expensive. Treating cloud like CAC makes it measurable, adjustable, and reportable to the board—just like every other growth metric.

Key takeaway: FinOps isn’t just about saving money but building operational leverage. When embedding cost-awareness into our team’s culture and systems, we turn cloud from a liability into a growth advantage and a more attractive business. SaaS businesses can achieve ~30% reduction in cloud costs after implementing FinOps⁶ processes, which gives us three things:

1. Cash flow freedom: high margins mean more cash per sale, giving us flexibility to invest in growth, product, and our team.

2. Resilience and optionality: we’re less exposed to economic downturns or VC dependence. We can scale sustainably or stay bootstrapped by choice.

3. Attractive for investment or exit: high-margin, capital-efficient SaaS businesses are gold for the bank and investors.

⁶FinOps Foundation

Play around with the model we whipped up for this letter

We’re squeezing more value from our existing infrastructure before negotiating discounts or adding more resources.

• Right-sizing everything: regularly auditing cloud usage (VMs, databases, storage) and scale down what’s oversized or underused. We’re shifting workloads to cost-effective configurations without sacrificing performance.

  • Why it matters: Oversized resources quietly drain our profits. Right-sizing aligns cloud spend with real demand, turning waste into margin and guesswork into discipline.

• Automate tiering and use smarter instances: moving cold data to cheaper storage and using spot or reserved instances for predictable workloads.

  • Why it matters: These changes can cut costs by ~45%⁷ with zero impact on performance. It’s the SaaS version of compound interest—tiny infra tweaks that add up to a massive margin over time.

Key takeaway: Cloud doesn’t get cheaper with scale, it gets smarter with thoughtful allocation. Before we build more, we fix the bloat. The fastest way to protect margins isn’t to grow revenue—it’s to stop wasting what we’ve already paid for.

⁷Intel

Play around with the model we whipped up for this letter

We’re spending more time making thoughtful architecture decisions early, so scale doesn’t come at the expense of our margins.

• Choose the right services for the job: we’re not defaulting to expensive virtual machines. We use managed services, serverless, or containers where they fit—especially for workloads with variable traffic.

  • Why it matters: We only pay for what we use. Serverless and PaaS options eliminate idle capacity, reduce operational overhead, and give us clean alignment between usage and spend, while delivering ~35%⁸ cost reductions.

• Minimize data transfer and wasteful patterns: we’re architecting our systems to reduce unnecessary cross-region or cross-zone data movement. We keep chatty services close to each other. Cache smartly.

  • Why it matters: Egress and transfer fees can quietly eat ~15% of our budgets⁹. We can’t cut what we can’t see—and bad design shows up as bloat in our cloud bills.

Key takeaway: Good architecture isn’t just scalable—it’s sustainable. When cost-efficiency is baked into the system from day one, you don’t just grow—you compound. Below are the GB vs. Sortage vs. RAM comparisons that helped us choose our vendors. Sharing to get your juices flowing, please do your own research as always.

Takeaways from the comparison above:

• Budget-friendly options: Digital Ocean’s Basic Droplet (1GB RAM, 1 vCPU, 25GB storage, 1TB data transfer) offers an affordable entry point at ~$4/month. Similarly, Hetzner’s CX11 plan (~$4.09 USD/month) provides competitive pricing for small-scale applications.

• Performance-oriented plans: OVHcloud’s Compute Optimized instance (~$0.0489/hour) and AWS T3 Micro (~$0.0104/hour) cater to workloads requiring higher compute efficiency. These plans include scalable infrastructure suitable for dynamic use cases.

• Scalable storage: Providers like Azure (B1s) and GCP (f1-micro) leverage SSD-based storage to support applications needing faster read/write operations, though they come at higher hourly costs than traditional HDD storage.

• Cost-sensitive projects: Digital Ocean and Hetzner are ideal for cost-sensitive projects such as startups or personal projects that need basic computing and storage at minimal costs.

• Compute-intensive workloads: OVHcloud’s Compute Optimized instances and AWS T3 instances stand out for their balance of performance and price.

• Speed and reliability: Azure and GCP offer robust SSD options for scalable and reliable storage for applications that demand speed and reliability.

⁸Sphinx Solutions  ⁹Redress Compliance

True craftsmanship means sweating the details no one sees. Managing cloud costs is our way of refining every corner of the business so we can move faster, build better, and lead the market.

Every dollar we save on waste is a dollar we invest in what matters most: our mission, our customers, and our future.

As mentioned in my recent letter, it will take me some time to get this letter right. Please bear with me as I get my writing, rhythm, and cadence in order.

Cheers to building together,
—Sanket

Selling your Vertical SaaS?

• I’m buying with $10M–$30M to spend on this criteria.

• I’m offering the fastest and most flexible exit that exists today; Entrepreneur.com called it a “reinvention of software exits.”

• It’s all happening at Spring Street Capital—the fund I founded to buy and scale great software businesses.

You likely built something special, and I’ll make sure it keeps growing that way.

Know someone selling? I’ll pay you $50K–$150K if your intro leads to me buying a business. Reply to learn more.