Nova Intelligence raised $31.5 million in Series A funding on May 5 to point agentic AI at SAP's $89 billion S/4HANA migration wave — a market that runs every Fortune 500 supply chain, payroll system, and finance ledger built on SAP's legacy ECC platform. The round was led by Chemistry, with participation from Accel, Conviction, and SAP's own venture arm SAP.io, bringing total funding above $40 million. The startup's pitch is straightforward: 77% of the world's transactions touch SAP systems, the legacy ECC sunset is a hard deadline, and the consulting industry that has historically owned this market is too expensive and too slow for what comes next.

That last part matters. Migration consulting fees account for 45–60% of total S/4HANA project budgets, day rates have risen 30–50% since 2022, and most enterprise programs run 12–24 months. Nova's reference customers report 75% reductions in manual effort and 50% cost savings (calculate your potential savings). For CIOs and CFOs looking at quotes from Accenture, Deloitte, IBM, and CapGemini for migrations that range from $1 million for the mid-market to north of $400 million for the largest Fortune 500 deployments, that is not an incremental improvement. That is a decision-tree-changing event.

What Changed: Nova's Funding and the Migration Math

Nova Intelligence is 18 months old, San Francisco-based, and built around an unusual founding team. CEO Emma Qian was previously a research engineer at Google DeepMind and Meta AI. Co-founder Professor Alexander Zeier is the co-inventor of SAP HANA itself and former CTO of Accenture's SAP Business Group — meaning the man who helped design the platform CIOs are now migrating to is helping build the AI that automates the migration. Sam Yang, a repeat operator, rounds out the founding team. Justin Kershaw, formerly CIO of Cargill and a public reference customer for the platform, joined as Chief Customer Officer.

The $31.5M Series A was announced on May 5, 2026 and reported exclusively by Fortune. Existing investor SAP.io continued participation — a meaningful signal given SAP's commercial interest in keeping migrations moving on its preferred timeline. Nova currently runs at 20 employees.

The product is an agentic AI platform that analyzes, modernizes, and generates the custom ABAP code that runs payroll, supply chain, and finance functions inside SAP customer landscapes. It is not a code-completion sidecar. It treats the migration itself as the workflow — the agent reads the existing custom-code estate, identifies what needs to change for S/4HANA compatibility, and generates the modernized equivalents.

Two reference outcomes anchor the pitch. Kyndryl, in an April 2025 case study, reported a 75% reduction in manual migration effort and 50% cost reduction. Festo, the German industrial automation company, reported "complex programs" completing in days versus months — a roughly 5x acceleration on the slowest part of the migration (Nova Intelligence customer page). These are not pilot numbers. They are production migration outcomes, which is what makes the funding round different from the 2024-vintage AI-for-code startups that struggled to convert demos into renewals.

The deadline pressure is the demand-side story. SAP ends mainstream support for ECC on December 31, 2027. Extended maintenance — at roughly 9% additional fees on top of existing maintenance — runs through 2030. RISE with SAP private cloud customers can extend through 2033 under specific commercial terms, but at materially higher subscription costs in the 2031–33 window than earlier migration would cost. After 2027, no security patches, no compliance updates, no legal-regulatory changes for the customers who stay. (SAP community discussion, TJC Group analysis of post-2027 options).

The total addressable opportunity, including implementation, upgrades, and ongoing support, is $89 billion. Some individual enterprise migrations are budgeted above $1 billion. Some Fortune 500 programs run to $400 million. This is the largest enterprise-software event of the decade, and it is mostly still ahead of the buyers.

Why This Matters: Technical and Business Implications

Technical Implications (CIOs and CTOs)

The migration challenge is not the SAP standard objects — those upgrade reasonably well through SAP's own tooling. The challenge is the custom code. Most large SAP customers have accumulated 10–25 years of ABAP modifications: industry-specific posting logic, custom workflows, integrations with non-SAP systems, regulatory adaptations, and a long tail of tactical fixes that nobody documented. Migrating that custom estate to S/4HANA — which deprecates large portions of the old ABAP API surface — is the work that makes 24-month timelines and $1M+ consulting bills.

Agentic AI changes the unit economics of that work. Instead of staffing a consulting bench at $150–$350/hour to read and rewrite ABAP, an agent reads the entire estate, classifies what's compatible, what's deprecated, and what's redundant, and generates the modernized replacements with human review in the loop. The 75% manual-effort reduction at Kyndryl is the headline number. The harder-to-quantify benefit is consistency: the agent applies the same standards across every module, which legacy SI delivery — by its nature — does not.

The architectural question CIOs should be asking is whether the agent's output meets your non-functional requirements: code review ability, security posture, observability, audit trail. Nova's customer base inside SAP suggests the answer is yes, but the burden of proof is on the buyer to verify before committing. Sample the generated code. Run it through your existing static analysis. Have your principal architects pressure-test the worst 10 custom programs in your estate, not the easiest ten. If the agent handles the worst ones at 70%+ acceptance, you have a viable migration path.

Business Implications (CFOs and Business Leaders)

The financial reframing is more dramatic than the technical one. S/4HANA migration is one of the largest discretionary IT line items most enterprises will see this decade. The average enterprise migration runs $1M–$8M for mid-market and well into eight figures for large deployments, with consulting costs accounting for 45–60% of the total. Day rates for experienced S/4HANA specialists have risen 30–50% since 2022 — a function of constrained supply (every Big 4 firm is competing for the same ABAP talent) meeting fixed demand (the deadline doesn't move just because the consultants are expensive).

If agentic AI delivers anywhere near the 50% cost reduction Nova's references claim, the savings on a $5M mid-market migration is $2.5M. On a $50M enterprise migration, it is $25M. On the $400M Fortune 500 cases, it is $200M (Tachyon mid-market S/4HANA cost analysis). These are not sensitivity-table numbers — these are line items that should already be in your 2026 capital plan.

The deadline math is also a CFO problem. Extended maintenance through 2030 is 9% extra annually on existing maintenance. For a customer with $20M in existing SAP maintenance, that is $1.8M/year of pure cost-of-delay starting January 2028. RISE-extended-to-2033 is materially more expensive than that. Migration is going to happen. The question is whether you migrate before the consulting market gets even more constrained, after, or with AI tooling that compresses the timeline regardless of consultant availability.

For competitive positioning, there is a separate signal: 85% of private-equity buyers now factor AI-enabled finance and operations capabilities into target valuations (Fortune coverage of the consulting disruption thesis). A portfolio company stuck on ECC in 2028 is not just a technical-debt problem. It is a valuation discount.

Market Context: Why the Consulting Model Is Cracking

Three things are happening simultaneously, and Nova sits at the intersection of all three.

First, the foundation-model providers are vertically integrating into services. Anthropic's $1.5B joint venture with Blackstone, Goldman Sachs, and Hellman & Friedman launched on May 4. OpenAI's "Deployment Company" finalized at a $10B valuation with $4B from TPG, Brookfield, Advent, and Bain Capital the same week (TechCrunch coverage). Both ventures use the Palantir-style forward-deployed engineer model — embed the engineering talent directly with the customer, build the workflows on the model provider's stack, and own the relationship. Nova is doing the same thing for the SAP-specific use case.

Second, the SI partners are responding by launching their own forward-deployed programs. ServiceNow and Accenture announced their FDE program on May 4 — 300+ pre-built AI agent skills, joint engineering pods inside customer environments (Accenture press release). Cognizant launched Secure AI Services on May 7 with similar architecture. The Big 4 are not standing still — they are repositioning. The question is whether the repositioning produces structural cost reduction, or whether it just adds AI surcharges to the same hourly rates.

Third, the analyst data is converging on the trend. Gartner forecasts that 40% of enterprise applications will feature task-specific AI agents by 2026, up from less than 5% in 2025 (Gartner press release). Forrester predicts 30% of enterprise app vendors will launch their own MCP servers, and half of ERP vendors will launch autonomous governance modules. The infrastructure for agentic ERP work is being built in real time.

For a CIO standing in front of an S/4HANA migration decision in May 2026, the practical question is not whether AI will reshape this work. It is how to bid the project so AI tooling produces savings instead of just enriching the SI. That is what the framework below addresses.

Framework #1: SAP S/4HANA Migration Decision Matrix

The migration market has historically been a four-quadrant choice: SI-led (Big 4), boutique SAP specialist, internal team, or RISE with SAP. Agentic AI tooling adds a fifth quadrant — and changes the economics of the original four. Use this matrix to score your situation.

Five Migration Path Options

Option A: Big 4 SI-Led (Accenture, Deloitte, IBM, CapGemini). Full-service migration with the SI owning delivery end-to-end. Pricing: $150–$350/hour blended, 12–24 month timelines, 45–60% of total budget consumed by consulting fees. Best for: largest Fortune 500 with multi-region, multi-entity complexity and an existing SI master agreement.

Option B: Boutique SAP Specialist. Mid-tier firms (Tier 2 SI) with deeper ABAP specialization at lower day rates. Pricing: typically 20–30% below Big 4. Timeline similar. Best for: mid-market enterprise with one geography, moderate custom-code estate, and a CIO willing to manage delivery directly.

Option C: Internal Team (with SAP support). Build an internal migration team using contractors and SAP's own conversion tooling. Pricing: lowest direct cost but highest opportunity cost (your best architects pulled off run-the-bank work for 18 months). Best for: enterprises with deep existing ABAP expertise and the executive willingness to fund a multi-year internal program.

Option D: AI-Native Migration (Nova-style platform + thin services layer). Use agentic tooling to handle the 70–80% of custom code that is mechanical, supplemented by a small architect team for the long-tail edge cases. Pricing: typically 40–60% below Big 4 SI on the same scope, per Nova's reference outcomes. Best for: any enterprise with a heavy custom-code estate where mechanical analysis dominates the work.

Option E: RISE with SAP / Stay-and-Wait. Move to RISE private cloud and extend support to 2033. Pricing: highest 5-year TCO of any option, but lowest near-term capex. Best for: enterprises with active M&A activity, regulatory uncertainty, or executive teams that want to defer the decision through CEO transitions.

Decision Scoring (assign 1–5, sum scores, highest wins)

For each option, score these five dimensions:

1. Custom-code estate size. Big estate (>500 custom programs) favors Option D. Small estate (<100) favors Option C or B.
2. Deadline distance. Less than 18 months from go-live target favors A or D. More than 30 months allows B, C, or E.
3. In-house ABAP talent depth. Strong bench favors C or D. Weak bench rules out C, favors A or D.
4. Budget elasticity. $10M+ headroom favors A. $1–10M favors B or D. <$1M forces C, D, or E.
5. Regulatory or auditability constraints. Heavy regulation favors A or D (auditable AI output) over B or C.

A simplified rule: if your custom-code estate is large, your deadline is tight, and your budget is constrained — which describes the median Fortune 1000 SAP customer in 2026 — Option D is the path that did not exist in your last migration playbook. Pilot it on a contained module before committing to the full estate.

Framework #2: Three-Scenario S/4HANA Migration ROI Calculator

The case for AI-native migration is most compelling when you put numbers next to it. These are illustrative scenarios using public consulting rate data and Nova's referenced 50% cost-reduction outcome. Plug in your own numbers, but the structure should hold.

Scenario 1: Mid-Market Manufacturer ($5M total migration budget)

• Big 4 SI baseline. $5M total: $3M consulting (60%), $1M SAP licenses, $1M change/training. 14-month timeline. Day rate $250 blended.
• AI-native path (Option D). $3M total: $1.5M AI platform + thin services (50% reduction on consulting line), $1M SAP licenses, $0.5M change/training (faster timeline = less change cost). 9-month timeline.
• Net savings: $2M (40% total budget reduction). Time saved: 5 months of run-rate license carrying cost on dual systems.

Scenario 2: Large Enterprise Retailer ($30M total migration budget)

• Big 4 SI baseline. $30M total: $18M consulting (60%), $7M SAP licenses, $5M change/data/integration. 20-month timeline.
• AI-native path (Option D). $20M total: $9M AI platform + services (50% reduction), $7M licenses, $4M change/data (compressed timeline). 14-month timeline.
• Net savings: $10M. Time saved: 6 months. Avoided 2028 extended-maintenance fees ≈ $1.8M/year on $20M base maintenance.

Scenario 3: Fortune 500 Multinational ($150M total migration budget)

• Big 4 SI baseline. $150M total: $85M consulting (57%), $40M SAP licenses, $25M change/integration/data. 36-month timeline. Multi-entity, multi-region, regulatory complexity.
• AI-native path (hybrid: AI tooling + Big 4 oversight). $105M total: $42M AI platform + Big 4 oversight (50% on the mechanical-code line, full Big 4 on architecture and program management), $40M licenses, $23M change/integration. 28-month timeline.
• Net savings: $45M. Time saved: 8 months. Avoided multi-year RISE-premium pricing if migration completes pre-2027.

The scenarios assume Nova's reported 50% cost-reduction transfers to your custom-code estate. It may not — that is the diligence question. But even at half the claimed effect (25% reduction on the consulting line), the savings on Scenario 3 is $20M, which more than pays for the platform cost twenty times over.

A Real-World Example: Cargill and the Custom-Code Bottleneck

Justin Kershaw was CIO of Cargill, the privately held agribusiness giant, before joining Nova as Chief Customer Officer. His public commentary on the migration problem describes the structural bottleneck cleanly: large SAP customers have, in many cases, 10–20 years of accumulated ABAP modifications, no current owner of most of them, and no realistic timeline to manually re-author the entire estate within the deadline window.

Cargill is the textbook profile: long-tenured SAP customer, deeply customized for agricultural commodity logic, multi-region, regulated, and operationally critical (a single day's outage during harvest season has measurable food-supply impact). The traditional path — Big 4 SI, 24+ month program, $50M+ budget — was the only viable option in 2022. By 2026, with agentic tooling reading and modernizing the custom estate at 5x speed, the path narrowed to "Big 4 + AI tooling" or "AI-native + Big 4 oversight on architecture only." The savings line is real. The deadline pressure makes it inevitable.

The lesson for other CIOs is the timeline of the realization. In 2024, AI for ABAP migration was a demo. In 2025, it was a pilot. In May 2026, it has reference customers in production with measurable outcomes and a $40M+ funded vendor with SAP's own venture arm on the cap table. The buyers who treat it as still-experimental in the 2026 budget cycle are going to be the ones explaining to the board, in mid-2028, why the migration ran 30% over budget while a competitor's came in on time at half the cost.

What to Do About It

For CIOs. Add Option D to your S/4HANA migration RFP this quarter. Specifically: require any responding SI to disclose what AI tooling they will use, what percentage of the custom-code work the tooling will handle, and what cost concession the tooling produces in their bid. If the answer is "we'll use our internal AI platform but pricing is the same," the bid is non-responsive. The market clearing price for SAP migration is moving — your procurement language has to move with it.

For CFOs. Ask the migration program owner for a specific line item: "AI tooling savings target." If the project plan does not have one, the project plan is a 2024 plan in a 2026 market. The default assumption should be that 25–50% of the consulting line in the original budget is now negotiable — and the negotiation is more likely to succeed if you go in with reference data (Kyndryl's 75% manual-effort reduction, Festo's 5x speed, the funding round that just validated the category).

For Business Leaders. If your enterprise is sponsoring the migration, the change-management timeline compresses with the technical timeline. A 9-month migration instead of 14 means three fewer months of dual-system overhead, three fewer months of training pause, three fewer months of integration freeze on adjacent systems. The downstream operating-rhythm benefit is larger than the direct cost saving for many businesses. Plan the change calendar to capture both.

For Boards. The S/4HANA migration is the largest discretionary IT decision most enterprises will make this decade. The 2027 deadline is real. The cost compression from agentic AI is also real. The job of the board is to ensure management is solving for both — not deferring the migration past the deadline window, and not paying 2022 consulting rates for 2026 work that should be largely automated.

The category Nova entered is not new — companies have been migrating from ECC to S/4HANA for nearly a decade. What is new is that the labor-intensive part of the work, the part that has historically dominated the budget, can now be done by software. The next 18 months will be the period in which buyers either capture that economic shift or pay a premium for not having captured it. The vendors are funded. The reference customers are public. The deadline is fixed. The decision is downstream.

---

Continue Reading

• $1.5B War: Anthropic vs OpenAI Deploy Palantir AI Playbook
• SAP Pays $645M/Year for an 18-Month-Old AI Lab. Here's Why.
• ServiceNow + Accenture: FDE Program for the 95% That Fail in Pilot
• Frontier Firms Use 3.5x More AI: Score Your Gap
• Microsoft + Accenture's 743,000 Copilot Deployment: Inside the Largest Enterprise AI Rollout

---

Sources cited in this piece: Fortune (Nova Intelligence Series A exclusive, May 5, 2026); TechCrunch (Anthropic + OpenAI joint ventures, May 4, 2026); Accenture press release (ServiceNow FDE program, May 4, 2026); Anthropic enterprise services announcement; Blackstone press release; Gartner enterprise AI agents forecast (40% by 2026); Forrester 2026 enterprise software predictions; SAP community deadline guidance (ECC end-of-support 2027, extended maintenance 2030, RISE through 2033); TJC Group analysis of post-2027 migration options; ERP Research consulting cost benchmarks 2026; Tachyon S/4HANA mid-market migration cost analysis; Pragmatic Engineer forward-deployed engineering analysis; GroovyWeb AI consulting rates 2026; Palantir Forward Deployed Software Engineer model documentation; CNBC reporting on the Anthropic/Blackstone/Goldman venture (May 4, 2026).