Your SAN Refresh Has a Real Number.
Here’s What It Is.

The framework path comes in $186K under your existing budget placeholder.

A one-page read for finance and executive sponsors. The detail behind these numbers is in sections 2 through 4.

The situation in three sentences

Meridian Manufacturing is scheduled for a 360TB SAN refresh in Q3 2026, against a $750K budget placeholder already in the FY24 capital plan. The Pure Storage quote in hand ($1.18M) exceeds that placeholder by $430K, and three-year exposure on the Pure + VMware path totals $2.46M through 2028. The framework path — VergeOS deployed on your existing Dell PowerEdge servers with refurbished enterprise SSDs — runs the same workload at $852K over three years, comes in under the existing budget placeholder in year one, and recovers enough delta to fund the VMware exit inside the same fiscal envelope.

The three numbers that matter

Why this works without new spending

VergeOS runs on the 14 Dell PowerEdge R750 servers Meridian already owns. The ioGuardian host is repurposed from existing inventory. The only year-1 hardware spend is for the refurbished enterprise SSDs that replace the Pure array’s drives — and those land at roughly half the cost of new flash. The combination of “use what you already own” and “buy capacity at qualified secondary-market pricing” is what brings year-1 cost in under the placeholder budget.

What your refresh costs at current vendor list, projected through 2028.

The number doesn’t shrink if you defer. The deferral path is shown alongside the planned refresh path.

Your current footprint

Meridian currently runs 240 TB usable of storage on a Pure FlashArray //X50 R3, deployed in Q2 2022. Support expires September 30, 2026. The next refresh is scheduled for Q3 2026, with a target capacity of 360 TB at the primary site and 120 TB at the DR site to accommodate 20% annual growth and the compliance retention obligation.

The vendor quote in hand

Pure Storage has provided a refresh quote dated April 18, 2026: $1.18M all-in for FlashArray //X70 (primary) and //X20 (DR), with a 30-day validity window — down from 90 days last cycle. The quote breaks down as $820K hardware and $360K Evergreen subscription over five years. The annual escalator written into the support component is 7%. Pure indicated that next quarter’s quote may be higher.

Compared to your 2022 deployment ($680K all-in for the same logical role), Pure’s 2026 quote represents a 74% increase — capacity grew 50% but cost grew 74%. The instinct that something has changed in the procurement math is correct, and the next table shows what the three-year exposure looks like if that path is followed.

Three-year cost trajectory

Cost component	2026	2027	2028	3-Yr Total
Hardware refresh (Pure quote)	$820,000	—	—	$820,000
Evergreen support / maintenance	$72,000	$77,000	$82,500	$231,500
DR array support (//X20)	$28,000	$30,000	$32,100	$90,100
Capacity expansion (20% growth)	—	$94,000	$103,000	$197,000
VMware vSphere Foundation	$316,000	$371,300	$436,300	$1,123,600
Storage subtotal	$920,000	$201,000	$217,600	$1,338,600
VMware subtotal	$316,000	$371,300	$436,300	$1,123,600
Status-quo 3-year exposure	$1,236,000	$572,300	$653,900	$2,462,200

Above table reflects vendor-quoted pricing plus the stated 7% storage escalator and a 17.5% midpoint VMware escalator (Broadcom guidance: 15-20%). Excludes the operational risk of running the existing Pure //X50 in unsupported mode if the refresh slips past Sept 30, 2026.

What deferring does

Deferring the refresh by one year does not reduce the number. Industry forecasts (TrendForce, IDC, Counterpoint) converge on NAND/DRAM normalization no earlier than 2028. Deferring to 2027 increases hardware cost by an estimated 30-40% on the same SKU, while exposing the existing array to 12+ months of unsupported operation. The deferral path is a higher number, not a lower one — and it carries operational risk the SLA cannot absorb.

Worth flagging directly: your CFO’s instinct that storage shouldn’t be 74% more expensive over four years is correct as a long-term observation but wrong about this specific cycle. The supply-side dynamics driving the increase (HBM reallocation, hyperscaler 2026 sellout, 30-day quote validity) are structural and forecast to persist through 2027.

The same workload, the same SLA, on the framework.

VergeOS deployed on your existing 14 Dell PowerEdge servers. Refurbished enterprise SSDs at qualified supply. ioGuardian repurposed from existing inventory. No new server purchases.

The architectural shift in one paragraph

The framework path consolidates compute and storage on your existing 14 Dell PowerEdge R750 servers running VergeOS as the operating system. The dedicated Pure array is retired. The Brocade fabric becomes the cluster backplane. Storage capacity is delivered by qualified refurbished enterprise SSDs pooled across the cluster as a single logical tier. RF3 (three synchronous replicas, N+2 tolerance) plus ioGuardian (active failure absorber) replaces RAID-based protection with a model that mathematically tolerates two simultaneous host losses without service impact — and a documented production event of four-of-six host loss with zero service impact (see Appendix).

The drives we’re sourcing for Meridian are not salvage. They come from hyperscaler lease cycles with R2v3 chain-of-custody documentation and 80-95% rated write life remaining. This is qualified secondary market inventory, not used hardware. The procurement framework filters every drive through six certification gates before deployment, and the architecture absorbs the residual failure rate operationally.

Sizing for Meridian

Based on your 14-host / 672-core / 340-VM footprint and the workload mix (30% general-purpose, 20% databases, 25% app servers, 15% backup, 5% VDI, 5% other), the recommended VergeOS cluster configuration uses your existing servers in three roles:

• HCI nodes: 8 of your existing R750s, balanced compute and storage
• Storage-only nodes: 4 of your existing R750s, packed with refurbished flash for database and file workloads
• Compute / GPU nodes: 2 of your existing R750s reserved for the predictive maintenance pilot (late 2026)
• ioGuardian host: 1 server repurposed from your existing inventory — no new server purchase required
• Total flash capacity: 360 TB usable at primary, 120 TB at DR, sourced from refurbished enterprise SSDs across multiple suppliers (Samsung, Micron, Solidigm)

Your existing Dell servers stay. They are 3 years into a 5-year refresh cycle, healthy, and have empty drive bays. The architectural shift uses what you already own.

VergeOS subscription pricing

VergeOS is licensed per server, at $5,000 per server per year. For Meridian’s 14-server cluster (including the repurposed ioGuardian host), that lands at $70,000 annually, or $210,000 over the three-year analysis window. Storage management, replication, ioGuardian, and continuous telemetry are included — there is no separate storage support contract because there is no separate storage system.

Hardware delta — line by line

Line item	Status quo (Pure)	Framework path (VergeOS)	Delta
Storage hardware (year 1)	$820,000	$410,000	($410,000)
Server hardware	—	$0 (existing + repurposed)	$0
VergeOS subscription (3-yr · 14 servers × $5K/yr)	—	$210,000	+$210,000
Storage support / maintenance (3-yr)	$321,600	included in subscription	($321,600)
Capacity expansion (years 2-3, refurbished)	$197,000	$74,000	($123,000)
Migration / professional services	—	$42,000	+$42,000
3-year storage total	$1,338,600	$736,000	($602,600)

Year-1 budget reconciliation

Year-1 framework spend lands as: $410K refurbished SSDs + $70K VergeOS year-one subscription + $42K migration services + $42K initial capacity headroom = $564K total. Against your $750K placeholder, that is $186K under budget — money that returns to the FY26 capital plan or stages forward to fund the VMware migration in Phase 2.

Performance and availability — equivalent or better

Your current 99.99% SLA requirement is met by RF3 alone. ioGuardian extends tolerance beyond the mathematical baseline — the 4-of-6 host failure event documented in the Appendix occurred at a similarly-sized customer with zero service impact and zero data loss. Your peak performance target of 85K IOPS and 4 GB/s falls within the validated envelope of a properly-sized 14-node VergeOS cluster on enterprise refurbished media. Tier-1 sub-1ms latency targets for Oracle and SQL are met or exceeded; we are happy to put this in writing as part of the test-drive evaluation.

Note on your specific concerns

On the “refurbished” optics question for the CFO: the language we use in this report — and that you should use with Janet — is “qualified secondary market inventory” and “hyperscaler lease-cycle drives with R2v3 chain-of-custody.” These are the words that match what’s actually being procured. The word “refurbished” appears in industry literature; the substance is a different category of supply.

On SAP HCL: VergeOS is on the supported hypervisor list as of February 2026. Certification documents are available on request — we can send these to your SAP basis team directly.

One cycle. Two wins. The math, and the path.

The refresh is required. The VMware exit is increasingly required. The framework executes both inside one budget envelope — and stays under it.

The VMware math

Meridian currently spends $316K annually on VMware vSphere Foundation across 672 cores at $470/core, with a renewal date of April 30, 2027 — your first full annual cycle under Broadcom. Your account team has indicated a 15-20% increase at next renewal (we use 17.5% as the working midpoint). Three-year exposure on the current path:

The $116K VMware figure assumes VMware decommission completes by April 30, 2027 (renewal date), with a four-month pro-rated VMware co-existence window during migration phases 1 and 2. Combined with the $602,600 storage savings from section 3, the 3-year total framework cost is $852,000 against $2,462,200 status-quo.

The one-cycle, two-win migration plan

The hardware delta from section 3 — $602,600 over three years, with $186K under the year-1 budget placeholder alone — is the funding mechanism for the VMware migration. The migration runs in waves, with VergeOS deployed alongside the existing VMware environment during the transition. No capital request beyond the storage refresh budget already in planning. No fight with the VMware renewal cycle.

Customer-specific dependencies — addressed

• SAP HCL: VergeOS confirmed on supported hypervisor list as of Feb 2026. Documents available for your basis team’s review pre-Phase 2.
• Veeam Backup & Replication v12: Hypervisor-agnostic. No change to backup architecture during migration. Existing job schedules and retention policies carry forward.
• Pure Cloud Block Store integration to Azure: This is the trickiest dependency in your environment. We recommend transitioning to native Azure Block Storage with VergeOS replication targets — design session in Phase 1 covers the cutover plan.
• VDI Horizon brokers: Migration to Omnissa Horizon (multi-hypervisor) or Citrix DaaS planned for Phase 3, after tier-1 VM migration is validated. Citrix conversion path is well-documented for VMware exits.
• Oracle DB licensing: Per-core licensing follows the host regardless of hypervisor. Math changes slightly but doesn’t block. Your Oracle account manager should be engaged in Phase 2 to update license positioning.
• Tom’s VMware skills: VergeOS administration is taught in a 2-day course; Tom’s existing virtualization fundamentals carry over. Recommend training Tom in Phase 1 so he leads Phase 2 cutover.