Cloud as a Path to Better EBITDA

By Lou Perugini | CloudTechCFO.com

The Myth: Cloud Hurts EBITDA

Spend five minutes with a finance leader debating cloud strategy and you’ll hear it:

“Cloud spend drags down EBITDA because it’s OPEX, not CAPEX.”

On the surface, it makes sense. In the on-premises world, servers, networking gear, and data

centers were purchased as capital expenditures. Those investments didn’t hit EBITDA directly

because they were capitalized and depreciated over time.

Cloud flips the model. Infrastructure becomes operating expense, recognized immediately. For

CFOs steeped in CAPEX discipline, that looks like an automatic EBITDA headwind.

But here’s the truth: cloud doesn’t destroy EBITDA — unmanaged cloud does. When

governed properly, cloud spend is often more accretive to EBITDA than legacy models ever

were.

I’ve modeled this for dozens of customers across industries. Using reasonable assumptions —

utilization rates, autoscaling, avoided refresh cycles — I have never seen a case where staying

on-premises produced a better EBITDA outcome than cloud.

Why the CAPEX vs. OPEX Lens Is Too Narrow

The problem with the “cloud hurts EBITDA” mindset is that it fixates on accounting treatment

instead of economic reality.

•  CAPEX isn’t free. Buying servers doesn’t make costs disappear. It just spreads them

across depreciation schedules. EBITDA may look cleaner in the short term, but cash is

still consumed and assets still sit on the balance sheet.

•  On-premises requires overbuilding. Enterprises must size for peak season, maximum

throughput, and a safety factor on top. Servers then sit idle much of the year. Studies

show on-prem servers often run at only 30–40% utilization, sometimes in the single

digits.

•  Cloud is elastic. With autoscaling, you pay for what you use. Spiky workloads scale up

and down seamlessly, eliminating the structural waste of idle capacity.

•  Efficiency compounds. With proper governance — reserved instances, savings plans,

and enterprise discount agreements — unit costs bend down over time. On-premises

assets, by contrast, tend to get less efficient as they age.

•  Transparency is unmatched. In the cloud, with proper account structures and tagging

strategies, 90% or more of spend can be directly attributed to individual products or

teams. Only 10% (or less) needs to be spread “like peanut butter” across the portfolio. By

contrast, even the best Technology Business Management (TBM) system on-premises

rarely achieves more than ~50% attribution. The rest stays trapped in shared services and

allocations.

Seen through this lens, cloud isn’t an EBITDA drag. It’s an opportunity to align costs tightly

with revenue, expand gross margin, and create true operating leverage.

Case Story: Oil and Gas — Busting the VMware Myth

I once worked with a large oil and gas customer who was convinced their on-premises VMware

environment was cheaper than the cloud. The CIO pointed to capital depreciation schedules and

thought the math was clear.

But when we built a full analysis, the picture looked very different. We included:

•  The Bill of Materials for servers, storage, and networking.

•  The headcount required to operate the environment — measured in servers managed per

FTE and TB of storage per FTE. On-premises was far more labor-intensive.

•  Ongoing software licensing and support contracts.

•  Hidden costs outside the CIO’s budget, such as real estate and facilities costs for the data

center.

When all costs were accounted for, the on-premises environment was 10% more expensive than

a comparable cloud footprint — and that was before layering in benefits from autoscaling and

operational agility.

The conclusion was clear: cloud improved costs and EBITDA when viewed holistically.

Case Story: Analytics at Cloud Speed

Another client, a global information and analytics company, wanted to test whether moving its

big data environment to the cloud would make sense. Running that experiment on-premises

would have meant acquiring racks of servers, installing them, and carrying the risk of stranded

assets if the initiative didn’t pan out.

Instead, we spun up 100,000 CPU cores in a single weekend using cloud spot instances. The

entire exercise cost less than $30,000.

The team got its answers quickly, with no sunk CAPEX and no leftover assets to depreciate.

More importantly, they were able to validate strategy in days rather than months.

That’s the EBITDA story in action: experimentation at low cost, with no long-term drag on

the balance sheet.

Case Story: Rethinking the EBITDA Mindset

At another enterprise, the CFO resisted optimization efforts. His concern: if customers were

paying for cloud usage with a markup, then reducing cloud spend would reduce revenue.

The reality was the opposite. Optimizing infrastructure didn’t shrink revenue; it reduced unit cost

while keeping revenue constant. Gross margins improved, and the EBITDA profile strengthened.

Once the model shifted from totals to unit economics, the CFO saw that efficiency was not a

threat to revenue, but a tailwind for profitability.

How Cloud Improves EBITDA

So, if CAPEX vs. OPEX is the wrong debate, what’s the right one? It’s about governance and

alignment.

1. Transparency Beats Black Boxes

Cloud costs can be directly attributed down to products, teams, or revenue streams at levels on-

premises could never match. That makes accountability real and cost control precise.

2. Unit Economics Drive Margin Expansion

When you measure cloud spend per transaction, per customer, or per workload, optimization

becomes actionable. If unit costs fall while revenue scales, EBITDA margin expands.

3. Centralized Commitments Lower Rates

Savings plans, reserved instances, and enterprise discount agreements reduce effective rates. A

central FinOps function should manage these, freeing engineering teams to focus on efficiency.

4. Forecasting Strengthens Guidance

Governed cloud spend is more predictable. Finance can run scenarios, sensitivity tests, and build

reliable forecasts. That makes EBITDA guidance to investors stronger and more credible.

5. Efficiency Frees Capital for Growth

Waste reduction isn’t about shrinking innovation. It’s about redirecting spend toward higher-

value growth bets. This drives both sides of the EBITDA equation: lower OPEX and higher

revenue.

Talking to Investors: Change the Narrative

When boards or investors push back with, “Cloud hurts EBITDA because it’s OPEX,” here’s the

counter-narrative:

•  On-premises CAPEX ties up cash, creates stranded assets, and requires overbuilding for

peaks.

•  Cloud elasticity avoids idle capacity and aligns cost to demand.

•  Unit rate optimization and autoscaling reduce total cost below the on-premises baseline.

•  With tagging and account structures, 90%+ of spend is directly traceable to products,

compared with less than 50% attribution in most on-premises environments.

•  Every point of efficiency drops straight to EBITDA without cutting growth.

  
  

In short: cloud doesn’t reduce EBITDA — it improves it, when governed.

Closing: EBITDA as a Cloud Advantage

EBITDA is one of the most scrutinized metrics for later-stage SaaS companies. The old CAPEX

mindset creates a myth that cloud adoption is an automatic headwind. The reality is the opposite:

unmanaged cloud erodes EBITDA, while governed cloud strengthens it.

CFOs who adopt this mindset can:

•  Improve EBITDA margins without starving innovation.

•  Tell a stronger efficiency story to investors.

•  Redeploy avoided capex into growth opportunities.

•  Use cloud governance as a lever to trade at higher valuation multiples.

• 
  

I’ve built pro forma models of on-prem versus cloud for many customers. Using reasonable

assumptions — utilization, scaling, refresh cycles — I’ve never seen an on-premises case

outperform cloud on EBITDA.

Cloud, when managed correctly, is not an EBITDA drag. It’s an EBITDA advantage.

Want help reframing your cloud spend as a path to stronger margins?

LET'S TALK