18 kilometers. That's the length of cabling needed to run a single train car — a figure that sums up the industrial complexity of the sector. At the Tech For Industry Show, Mehdi Belahcen, Cloud Compute and Infrastructure Architecture Director (Alstom) and Patrice Bessiere, IT Director Solution Delivery (Exide Technologies) shared, alongside Clément Abecassis, Snr Manager, DBP and OCI Sales (Oracle France), two transformation journeys from opposite industries — rail mobility and energy storage — but with surprisingly convergent challenges.

At Exide — a battery manufacturer for data centers (UPS), forklifts (Motion division) and automotive (OEM and aftermarket, under the Fulmen and Tudor brands) — the migration from a legacy ERP to a market ERP had failed three or four times before. This time, several factors converged: an aging, hard-to-maintain IT team, a change in ownership, and a desire to rely on a standard ERP supported by external companies rather than solely on internal knowledge. After narrowing 10 ERPs down to 3, then to 1, Oracle Fusion Cloud won out for the depth of its manufacturing coverage — supported by its integrator (EVO6, now Mastech).

At Alstom — rolling stock (around 40% of the business), services and software, the latter two growing fast — it was the market that forced the transformation. "Five M&A deals in ten years": digital services needed the elasticity and agility to absorb massive internal and external growth. The encounter with Oracle, historically a database vendor, opened the door to adopting its cloud (OCI) for strategic use cases.

Exide's approach: adopt the standard, staying "as vanilla as possible" — while still incorporating the accounting and management specifics unique to each country across Europe and Asia-Pacific. The architecture isn't a monolith but an assembly of Oracle modules: Fusion at the core, EPM (formerly Hyperion) for budgeting and consolidation, WMS for warehouses, and OIC (Oracle Integration Cloud) to integrate legacy systems — all on a 4-year migration plan, from the Nordics and Poland (June 2024) to every country by 2028.

The scope broadened: lacking an equivalent to their on-premise BI solution (IBM Cognos) within the Fusion suite, Exide adopted Oracle Analytics Cloud — analytics being seen as just as critical as the ERP for a company undergoing major change. And beyond the ERP, a genuine data center exit strategy is taking shape, migrating to OCI the applications from a data center that notably hosts an AS/400 — for cost and ROI reasons.

Alstom's business is "very cash-sensitive": years go by between order and payment on delivery. A striking illustration: when homologating the first TGV in the United States, authorities required additional simulations (crosswind resistance, rail expansion depending on temperature). By scaling HPC infrastructure in the cloud, calculations that would have taken a year to a year and a half were completed in a few weeks — reassuring the client, closing the project, and getting paid nearly a year earlier. A case directly overseen by the CEO.

Another challenge: for an operator, "the more the train runs, the more money it makes" — a breakdown mid-line is the worst-case scenario. Alstom developed Hub, a predictive maintenance platform that uses sensor data to anticipate failures and plan replacements at night or on weekends. For the Riyadh metro, the client required that its data remain in Saudi Arabia: OCI's local presence made it possible to build, test and deliver the solution in a few weeks — a concrete illustration of the data residency challenge.

At Exide, the shift from batch to real time transformed analytics: traditional reporting is gone, replaced by dashboards issued every morning for executives and operations teams — now enriched with AI, at the almost daily pace of new Oracle Analytics features, which certified users request faster than IT can deploy them.

The rail world, "designed to be national," makes every train unique — a Thalys train combines three different trains. As a result, tenders run to hundreds, even thousands of pages. Alstom uses AI (on OCI) to analyze these requirements, identify what has already been delivered, and reuse existing work rather than starting from a blank page — a crucial issue for delivering on time. On the development side, with close to 7,000 developers, AI coding assistance adds 10 to 15% extra efficiency. And exploration continues toward agentic capabilities.

At Exide, analytics AI generates automatic commentary on report data and feeds a series of agents with a predictive approach (a KPI turns red → alert, email, meeting). However, modeling and automatic report creation remain under human validation: "AI can be wrong," and specialists keep control over what must be validated by the business.

None of these transformations happened alone. Alstom relied on DXC (infrastructure and cloud architecture redesign from the start), Capgemini (SOC / cybersecurity) and British Telecom (network) — coordinating this expertise was decisive for the speed of adoption. Exide, for its part, is advancing with Mastech (formerly EVO6) on the ERP implementation.

For Patrice Bessiere (Exide): having completed the migration to OCI, exited the legacy data center (AS/400 included), and deployed AI across the PaaS and SaaS layers — while keeping specialized applications well integrated (such as Blue Yonder for logistics). For Mehdi Belahcen (Alstom): supporting the business, measured by OTD (On Time Delivery) — delivering on time to bring in cash and absorb rail's "monstrous" growth. Two journeys, two industries, one shared goal: gaining agility, making better use of data, and creating value.

Watch the full panel, along with the other Tech For Industry Show replays, on our dedicated page.