Epic on Cloud Implementation Partners — 12 Firms Compared [2026]

AdventHealth completed its 53-hospital Rackspace cutover on November 14, 2024 — 38,000 concurrent users, sub-two-hour transition, nine states. Three months later, Healthcare IT Today reported that Geisinger’s AWS migration, led by Deloitte, was “probably the largest public cloud-based instance of Epic in the industry”: 7,500 servers, three data centers, 1,500+ applications reduced to 1,100, on-premises footprint cut 40%, cloud adoption from roughly 10% to over 90%. These are the reference cases for production-scale Epic in the cloud. The remaining buyer questions are which partner to use and which deployment model to start with.

KLAS’s November 2024 report, Epic in the Public Cloud 2024, documented approximately 30 health systems in North America running Epic production workloads on AWS or Azure. Roughly 75% used third-party consulting firms. That count excludes Epic-Hosted deployments and the much larger private-cloud installed base served by Rackspace, so total cloud-hosted Epic volume is materially higher.

This article names the KLAS-validated firms, distinguishes the five deployment models most listicles conflate, and states the cost-parity finding KLAS documented without softening it. Read it before issuing an RFP.

---

The Epic-on-Cloud Myths to Clear Up First

Most articles about Epic cloud migration get three facts wrong. These shape every decision downstream.

Myth 1: Epic has a formal cloud partner certification program.

It does not. What exists: Epic’s own Hosted Epic program, where Epic itself operates customer environments and charges a managed-hosting fee; informal “Epic-approved partner” standing that firms earn by working closely with Epic’s Technical Services and Hosting teams over actual migrations; and KLAS validation, which is the only published, independent verification of which consulting firms have documented, live Epic cloud deployments. There is no AWS-Premier-equivalent badge that Epic issues. Any vendor claiming “Epic Certified Cloud Partner” is inventing a credential.

Myth 2: Azure and AWS are interchangeable choices.

Epic’s product roadmap favors Azure. Cogito Cloud (formerly Nebula), Epic’s analytics and reporting platform, runs on Azure and Microsoft Fabric. Egress costs from an Azure-hosted Epic environment to Cogito are materially lower than from AWS or GCP. That is an ongoing operating cost, not a one-time migration cost.

KLAS 2024 also found that AWS customers report higher support satisfaction than Azure customers in live Epic deployments. AWS holds roughly 40% of healthcare cloud market share. Both hyperscalers are viable; the economics depend on whether Cogito/Fabric workloads are in scope.

Myth 3: Cloud Epic saves money versus on-premises.

KLAS 2024 is direct: most cloud-Epic deployments cost the same or slightly more than on-premises in the near term. The Forrester Total Economic Impact study of Epic on Azure (2025) found 162% ROI over three years and payback under six months — but that ROI comes primarily from $46.7M in avoided hardware-refresh capital expenditure and reduced DR complexity, not from a lower monthly cloud bill. The 2026 case for most CIOs is agility, ransomware recovery, and AI-readiness. Expect cost parity, not savings.

---

The 5 Epic Deployment Models

“Epic on cloud” is not one decision. It covers five architectures with different cost profiles, risk levels, and implementation timelines. Most listicles treat them as one.

Deployment model	Description	Typical workloads	Reference deployment	Primary cost lever
Self-hosted on-prem	Own data center; customer owns all hardware and operations	All Epic workloads	Baseline for comparison	Capex: server refresh every 5–7 years
Epic-Hosted	Epic’s own data centers; Epic operates the environment	Full production EHR	Epic’s managed-hosting customer base	Per-user managed hosting fee to Epic
Cloud Production (full)	Production EHR on AWS, Azure, or GCP	Entire Epic footprint	Geisinger (AWS), Mount Sinai (Azure), Hackensack Meridian (GCP)	Compute + storage opex; egress to Cogito
Alternate Production / Failover	Production stays on-prem or Epic-Hosted; cloud holds warm DR site	DR, HA	Many AdventHealth-style deployments started here	DR infrastructure opex; RTO/RPO economics
Build, Train & Non-Production	Dev, training, and testing in cloud; production elsewhere	Non-prod Epic environments	Common entry point for orgs new to cloud Epic	Reduced non-prod hardware footprint
Isolated Recovery Environment (IRE)	Clean-room ransomware recovery target on cloud	Cyber-resilience	Jupiter Medical Center (AWS, Cordea) — $4M capex avoided	One-time setup; fast RTO (15–20 min)

The IRE pattern matters in 2026. The Change Healthcare cyberattack in February 2024 reset how health systems think about ransomware recovery. Cordea Consulting and AWS launched a no-cost Epic IRE-on-AWS for the first 25 hospitals signing by April 30, 2025. Organizations that had not yet moved any Epic workload to cloud are entering through the IRE — a lower-risk, lower-cost first project that also builds the AWS Epic footprint needed for later migrations.

The dominant entry sequence in 2025–2026: IRE → Build/Train → Alternate Production → Full Cloud Production.

---

Hyperscaler Positions

Microsoft Azure: The Structural Volume Leader

Azure has more live Epic deployments by count than AWS, driven partly by the Cogito Cloud relationship and partly by Microsoft’s hospital-system enterprise agreements. So long as Cogito/Nebula runs on Azure plus Microsoft Fabric, health systems that care about EHR analytics will pay lower egress costs from Azure than from AWS.

Named Azure deployments with documented outcomes:

• Mount Sinai (Accenture-led, five-year transformation, March 2022): Azure migration across one of the country’s largest academic medical centers.
• Franciscan Health: $45M in projected five-year savings; Epic response time improved 50%; failover time reduced from 8–12 hours to 30 minutes.
• Sentara: Approximately $2.5M per year in savings; roughly $7–8M over three years.
• Legacy Health (Portland, OR): 65% cost reduction on hybrid + DR architecture.
• St. Luke’s University Health Network: 3M patient records, 20 Epic environments migrated.

The Forrester Total Economic Impact study of Epic on Azure (2025/2026) quantified the composite case: $68.2M in total benefits, $26.6M in total costs over three years, $41.7M NPV, 162% ROI, payback under six months. The largest benefit bucket: $46.7M in avoided hardware-refresh capital over three years. Epic infrastructure footprint reduction: 90% over the engagement period.

Amazon Web Services: Customer Satisfaction Leader and Largest Single Instance

AWS holds approximately 40% of healthcare cloud market share and now hosts what is credibly “probably the largest public cloud-based instance of Epic in the industry.” The Geisinger migration, completed with Deloitte and detailed at AWS re:Invent 2022, produced documented results:

• 1,500+ applications consolidated to 1,100
• 7,500 servers migrated across three data centers
• On-premises footprint reduced 40%
• Cloud adoption from roughly 10% to over 90%
• 30–60% savings versus legacy on-premises infrastructure
• 60% reduction in EHR disaster-recovery environment IT spend
• 90% reduction in deployment time

Other named AWS deployments:

• Care New England (via Kyndryl, August 2023): First Kyndryl-led Epic conversion; Kyndryl absorbed approximately 160 IT staff as part of the ITO transition.
• Tufts Medicine: 3M+ patient accounts cut over to Epic on AWS in 71 hours, with roughly 800 IT and clinical staff involved. The equivalent on-premises path was estimated at approximately one year.
• JMC (Jupiter Medical Center, Florida): Epic on AWS in a 13-month implementation; isolated recovery environment with 15–20-minute RTO; contact-center base costs cut in half; $4M capex avoided.
• NGHS (Northeast Georgia Health System) (via Optimum Healthcare IT): Epic 20% faster post-migration; full Epic production on AWS.

KLAS 2024 finding: AWS customers report higher support satisfaction than Azure customers in live Epic deployments. The gap correlates with AWS’s longer track record in healthcare infrastructure rather than any single cause.

Google Cloud: Real but Limited Epic Footprint

GCP is the least-deployed public cloud for Epic production workloads, but it is not theoretical. Hackensack Meridian Health announced at HLTH 2022 that it would migrate Epic to GCP — the first health system to do so publicly. Non-production EHR workloads were live on GCP as of October 2023, with the full production migration on a multi-year timeline.

Google’s healthcare data infrastructure is genuine: Cloud Healthcare API, Vertex AI, and MedLM are more developed than most enterprises realize. The limits for Epic on GCP are Cogito/Fabric egress economics that favor Azure, and a thin bench of consulting firms with documented Epic-GCP migrations.

Two common misattributions to avoid: HCA Healthcare’s Google partnership is for analytics and generative AI — HCA runs Meditech Expanse, not Epic. Mayo Clinic’s ten-year Google Cloud deal (September 2019) covers AI and research data; no public source confirms Mayo runs Epic production workloads on GCP.

Rackspace Healthcare Cloud: The “But Actually the Biggest” Caveat

Rackspace is the world’s largest third-party Epic hosting provider as of late 2024. The AdventHealth migration completed on November 14, 2024: 53 hospitals, 9 states, 38,000 concurrent users, sub-two-hour cutover, 99.999% Epic SLA. Rackspace reports 2,500+ healthcare clients and 1,125+ combined years of Epic experience. Reference client: Seattle Children’s.

This matters for buyer research because most top-10-Epic-cloud-partners listicles are built around public hyperscaler partner pages (AWS lists only four Epic partner firms; Azure’s healthcare partner listing is not Epic-specific). Rackspace’s private-cloud Epic hosted base is larger by deployed-user count than the entire public-cloud Epic installed base documented in KLAS 2024. Any serious evaluation of hosted Epic options needs to include Rackspace alongside the hyperscaler-native firms.

---

Top 12 Epic-on-Cloud Implementation Partners

Scoring basis: KLAS validation status, named reference clients at documented scale, cloud platform specialization, and distinguishing operational capability. Where KLAS 2024 validation is noted, it is based on published KLAS data; the absence of a notation does not confirm absence of validation.

---

1. Rackspace Technology Cloud focus: Private Healthcare Cloud (Rackspace-managed) KLAS status: Documented in KLAS healthcare-cloud reports; not public-cloud hyperscaler certified Scale indicator: World’s largest third-party Epic hosting provider; 2,500+ healthcare clients; 1,125+ combined years Epic experience Distinguishing strength: 99.999% Epic SLA; end-to-end managed operations; ITO capability Named reference clients: AdventHealth (53 hospitals, 38K concurrent users, Nov 2024 cutover); Seattle Children’s

For health systems that want Epic in the cloud without building or managing cloud infrastructure, Rackspace is the default comparison. The AdventHealth deployment is the largest reference case in the industry.

---

2. Deloitte Cloud focus: AWS-primary KLAS status: KLAS-validated for AWS Epic migrations Scale indicator: Big-4; thousands of Epic-certified consultants globally Distinguishing strength: Enterprise transformation methodology; “self-funded business case” framework documented at AWS re:Invent 2022 (session PRT284); deepest AWS partnership depth in healthcare Named reference clients: Geisinger (7,500 servers, 3 data centers, 1,500+ apps, Deloitte-led; cited as “probably the largest public cloud-based instance of Epic in the industry”)

The Geisinger migration is the benchmark AWS case. Deloitte’s approach — build the business case so projected savings fund the migration — is what most large IDNs end up copying when scoping a multi-year cloud migration program.

---

3. Accenture Cloud focus: Azure-primary KLAS status: KLAS-validated for Azure Epic migrations Scale indicator: Big-4; largest Microsoft alliance bench in healthcare IT Distinguishing strength: Microsoft partnership depth; five-year transformation methodology; Fabric/Cogito integration expertise Named reference clients: Mount Sinai (five-year Azure transformation, March 2022)

If Cogito Cloud and Microsoft Fabric analytics are central to the organization’s roadmap, Accenture’s Azure bench is difficult to match. The Mount Sinai engagement is the most-cited Azure Epic reference case. See the broader Azure partner landscape for context on Accenture’s healthcare positioning.

---

4. Kyndryl Consult Cloud focus: Azure and AWS (dual-cloud ITO) KLAS status: KLAS-validated for Epic cloud transitions Scale indicator: 90,000 employees globally; IBM spinoff; largest independent IT infrastructure services firm Distinguishing strength: Only top-tier global IT services firm with a productized “Azure Cloud Deployment Services for Epic EHR” offering; ITO transition capability (absorbs client IT staff) Named reference clients: Care New England (August 2023 — first Kyndryl-led Epic conversion; absorbed approximately 160 IT staff)

Kyndryl fits when the organization needs to migrate Epic to cloud and outsource IT operations at the same time. The ITO-plus-cloud-transition combination is their differentiator. No other firm on this list has documented experience absorbing hospital IT departments as part of an Epic cloud cutover.

---

5. Optimum Healthcare IT Cloud focus: AWS-primary (also Microsoft Advanced Tier) KLAS status: KLAS-validated for AWS Epic deployments Scale indicator: AWS Premier Tier Services Partner; healthcare-exclusive Distinguishing strength: Pure-play healthcare IT shop that combines deep Epic certification bench with cloud-native AWS architecture; covers Epic implementation, optimization, and cloud migration in one firm Named reference clients: NGHS (full Epic production on AWS, 20% performance improvement); Baptist Memorial (Epic DR on AWS); Inova (Beaker laboratory implementation)

Optimum combines Epic-certified staffing with AWS cloud architecture — a pairing that is uncommon among healthcare-only firms. Health systems that need Epic application expertise and cloud infrastructure expertise without managing two separate vendors should evaluate Optimum early. For AWS-focused healthcare migrations, Optimum is one of the most cited pure-play options.

---

6. Cordea Consulting Cloud focus: AWS KLAS status: KLAS-validated for AWS Epic transitions Scale indicator: Mid-size boutique; AWS Marketplace seller Distinguishing strength: Cyber-resilience-focused Epic on AWS; originator of the Epic IRE-on-AWS entry-point pattern; launched no-cost IRE-on-AWS offer for first 25 hospitals (deadline April 30, 2025) Named reference clients: Jupiter Medical Center (13-month implementation; $4M capex avoided; 15–20-minute IRE RTO; contact-center costs halved)

Cordea specializes in organizations whose first cloud-Epic workload is ransomware recovery rather than full production migration. The IRE pattern they developed — a clean-room Epic environment on AWS that activates in 15–20 minutes after a cyberattack — is now a common 2026 entry point for systems that hesitated to commit to full cloud production. Post-Change Healthcare, most security teams treat the IRE as a requirement.

---

7. Sapphire Health Cloud focus: AWS and Azure (cloud-agnostic) KLAS status: Only KLAS-validated firm cited for both AWS and Azure Epic transitions Scale indicator: Healthcare-focused boutique Distinguishing strength: Hyperscaler-agnostic posture; validated on both platforms; useful when the organization has not yet committed to a cloud provider Named reference clients: Not publicly named in available sources

Sapphire’s dual-cloud KLAS validation is its primary differentiator. For CIOs still in cloud provider selection, a partner validated on both platforms can run a genuine bake-off rather than steering the decision toward their primary alliance.

---

8. Atao Consulting Cloud focus: AWS KLAS status: KLAS-validated for AWS Epic transitions Scale indicator: Emerging boutique Distinguishing strength: Focused AWS + Epic practice; useful for smaller health systems and community hospitals entering their first cloud-Epic engagement Named reference clients: Not publicly named in available sources

---

9. AirGate Technologies Cloud focus: Azure KLAS status: KLAS-validated for Azure Epic transitions Scale indicator: Mid-size boutique Distinguishing strength: Azure-focused Epic cloud practice; relevant for health systems on Microsoft-heavy infrastructure Named reference clients: Not publicly named in available sources

---

10. CDW Healthcare Cloud focus: Azure (primary for Epic) KLAS status: KLAS-validated for Azure Epic transitions Scale indicator: Fortune 200 technology solutions provider; large healthcare practice Distinguishing strength: Breadth of IT portfolio (devices, networking, cloud, security) makes CDW practical for mid-market health systems that need an Epic cloud migration to coexist with a broader technology refresh; existing CDW relationships reduce vendor management overhead Named reference clients: Not publicly named for Epic-cloud engagements in available sources

---

11. Nordic Consulting Cloud focus: Cloud-agnostic (optimization and advisory) KLAS status: Consistently KLAS-top-rated for Epic implementation and optimization; cloud-migration-specific validation not confirmed in available sources Scale indicator: 1,000+ Epic-certified consultants; healthcare-exclusive Distinguishing strength: The deepest independent Epic optimization and adoption bench outside of Epic itself; Community Connect expertise; post-migration optimization and clinical change management Named reference clients: Broad Epic client base; specific cloud-Epic references not publicly named

Nordic is not a cloud-native infrastructure firm. Its value in an Epic cloud migration is the application layer: Epic configuration, workflow optimization, training, and the clinical change management work that decides whether a technically successful cutover also produces clinical adoption. Pair Nordic with a cloud-infrastructure-focused firm (Optimum, Cordea, Deloitte) for a complete team.

---

12. CereCore Cloud focus: Cloud-agnostic (operational expertise) KLAS status: Healthcare IT managed services; Epic implementation pedigree Scale indicator: 700+ clinical and technical staff; implemented Epic in 400+ facilities; originated inside HCA Healthcare (Parallon spinoff) Distinguishing strength: Hospital-operator DNA — CereCore staff have operated Epic from the inside of a 180+ hospital system; strong Hyperdrive and Epic version-transition expertise; community hospital and rural health focus Named reference clients: HCA Healthcare network heritage; 400+ facilities with Epic implementations

CereCore’s value is operational depth rather than cloud-infrastructure specialization. For systems where a hyperscaler partner is running the cloud migration but the Epic application layer needs a firm that has worked inside large hospital operations, CereCore is the right complement.

---

Honorable mentions (growing practices, less KLAS-cited in 2024):

Tegria (spun out of Providence; UVM Health Network reference; strong in Pacific Northwest health systems), EHC Consulting (founded by Tim Calahan, former CTO at Michigan Medicine, who led an academic medical center through a full Epic-to-public-cloud migration; firsthand operator pedigree is uncommon at this size), Avaap, Impact Advisors, and Healthcare Triangle (HCTI) all have growing cloud practices and Epic expertise. None appear prominently in KLAS 2024 public-cloud Epic findings, but each serves specific market segments — regional health systems, academic medical centers, community hospitals — where their relationships and delivery models are well-regarded.

---

Cost Benchmarks

KLAS 2024 baseline: most cloud-Epic deployments cost approximately the same as on-premises or slightly more in years one and two. Long-term savings are expected but not yet widely documented at the time of publication.

Engagement type	Typical cost range	Typical timeline
Cloud assessment + business case	$80,000 – $200,000	6–12 weeks
Build, Train & non-production env migration	$200,000 – $600,000	3–6 months
Alternate Production / IRE setup	$300,000 – $1,500,000	3–6 months
Full Epic production migration (mid-size health system)	$5M – $50M+ all-in	12–24 months
Ongoing managed cloud Epic (opex vs. legacy hosting)	1.0× – 1.4× legacy hosting opex	Ongoing
Pro services for full migration (Forrester TEI Azure composite)	~$4,500,000	18-month engagement

The Forrester Total Economic Impact study (Epic on Azure, 2025/2026) is the most detailed public cost model available. Composite organization: academic medical center, 1,000 Epic users, $4.5M in pro-services investment over 18 months. Results over three years: $68.2M in total quantified benefits, $26.6M in total costs, $41.7M net present value, 162% ROI, payback under six months. The dominant benefit category ($46.7M) is avoided hardware-refresh capital — not reduced monthly cloud spend. See the HIPAA-compliant cloud providers guide for compliance cost context.

For IRE-entry-point projects, Cordea’s Jupiter Medical Center case shows $4M capex avoided on a 13-month engagement — a more representative benchmark for community hospitals and single-facility systems.

---

Decision Framework: Picking the Right Partner

Seven criteria, in priority order for most health system buyers:

1. KLAS validation for your target hyperscaler. AWS and Azure KLAS validation are not interchangeable. If the organization has chosen Azure, confirm the candidate has documented Azure Epic deployments in KLAS — not just general Azure cloud experience.

2. Named reference client at comparable scale. A 50-hospital IDN and a critical-access hospital are not the same engagement. Ask for references at your bed count, IT staff size, and Epic module footprint. The difference between Deloitte’s Geisinger reference and Atao’s community-hospital references is real.

3. Healthcare-exclusive vs. cross-industry shop. Healthcare-exclusive firms (Optimum, Nordic, CereCore) carry deeper Epic institutional knowledge; cross-industry firms (Deloitte, Accenture, Kyndryl) carry deeper cloud infrastructure scale. Neither is universally better. The trade-off depends on where your risk is: application complexity or infrastructure complexity.

4. IRE / DR / non-production entry-workload offering. Most organizations now start with IRE, not full production migration. A partner that can start there and build toward production reduces commitment risk and builds operational cloud experience inside the team.

5. Cogito Cloud / Microsoft Fabric experience. If the organization plans to use Cogito’s cloud-native analytics — and Epic’s roadmap points there — confirm the partner has delivered Fabric-integrated Cogito environments, not just Epic runtime on Azure.

6. ITO transition capability. If absorbing the client’s IT staff is part of the deal (common in community hospital and IDN situations), only a handful of firms — Kyndryl leads, Rackspace second — have demonstrated this at scale. Most Epic consulting firms are not ITO firms.

7. Post-migration optimization track record. Most of the optimization work happens in the twelve months after cutover. Ask for references twelve months post-go-live, not at go-live. The firms that excel at Epic optimization (Nordic, CereCore, Optimum) stay engaged through that year; many cloud-infrastructure firms do not. Review the healthcare cloud migration checklist for a full post-migration review framework.

Also evaluate HITRUST CSF certification of candidate firms — cloud Epic environments are HIPAA-regulated infrastructure, and partner HITRUST certification is a minimum bar for most health system security teams.

---

Common Pitfalls

Treating cloud Epic as a cost-savings program. KLAS 2024 documented cost parity as the current state. CIOs who sell the board on “we’ll save 30% versus on-prem” will face a hard conversation in year two. Sell the actual case: agility, ransomware recovery, AI-readiness, and avoided capex on hardware refresh.

Picking the cheapest hyperscaler without modeling Cogito/Fabric egress. Egress fees from AWS or GCP to Cogito Cloud on Azure add up at Epic’s data volumes. Run the three-year total-cost model including analytics egress before signing an AWS contract for a Cogito-heavy environment.

Going straight to full production migration without an IRE or non-prod step. Full production Epic migration is a 12–24-month, $5M–$50M commitment. Organizations that start with IRE or build/train environments gain cloud operational experience and de-risk the production move. Post-Change Healthcare, most health system security teams are already asking for the IRE — use it as the entry point.

Underestimating third-party application vendor readiness. Geisinger’s migration was slowed by ISVs whose products were not yet cloud-compatible. Conduct a third-party application cloud-readiness audit before committing to a production migration timeline.

Conflating cloud technical cutover with clinical adoption. AdventHealth’s sub-two-hour cutover and Tufts Medicine’s 71-hour migration are technical milestones. The 800 IT and clinical staff Tufts mobilized, and the months of training before go-live, are the clinical adoption work. The cloud infrastructure partner and the clinical change management partner are often different firms — budget and plan for both.

---

Further Reading

• Healthcare cloud migration checklist — pre- and post-migration verification framework
• HITRUST vs HIPAA: cloud compliance for health systems — compliance certification requirements for Epic cloud environments
• Cloud providers for healthcare data platforms — AWS, Azure, and GCP compared on healthcare-specific capabilities
• HIPAA-compliant cloud providers — BAA coverage and compliance posture across major providers
• AWS healthcare partners — full AWS partner directory with healthcare specialization
• Azure healthcare partners — full Azure partner directory with healthcare specialization
• Cloud migration consulting services — vendor-neutral cloud migration partner selection
• Industries: Healthcare — full healthcare IT consulting partner directory