Tata Steel Supply Chain Innovation

Tata Steel Nederland runs an integrated chain from one site: ore and coal arrive by sea, two coking plants and two blast furnaces make 6.7 Mt of liquid steel a year, the strip is finished on site and in twenty downstream plants for automotive, packaging and construction (Tata Steel Nederland, 2026). Innovation means replacing the coal-based ironmaking step; three conditions force it.

Carbon pricing is the first. Steel emissions come from chemistry, not fuel (Bataille et al., 2018); Tata Steel Nederland emitted about 11 Mt in 2025/26, 1.66 tonnes of CO₂e (scopes 1 and 2) per tonne of crude steel (Tata Steel Nederland, 2026). Free ETS allowances for steel are withdrawn in 2026–2034 and the CBAM entered its definitive phase on 1 January 2026 (European Parliament and Council, 2023a; 2023b). The response is to replace blast furnace 7 and coking plant 2 by direct-reduced iron with an electric arc furnace (DRI/EAF), cutting at least 5.4 Mt a year (±0.6 Mt in the JLoI) in phase one, half of actual emissions (licensed maximum 12.6 Mt) (Tata Steel Nederland, 2026). However, that response is financed from the cash flow the carbon cost erodes; carbon pricing alone does not finance first commercial plants (Vogl, Åhman and Nilsson, 2021).

Health and legal pressure form the second: RIVM (2023) linked deposition from the works to health risks in the IJmond, and the Public Prosecution Service has summoned Tata Steel IJmuiden for releasing harmful substances (Openbaar Ministerie, 2026). The response is the earlier closure of coking plant 2 (Tweede Kamer, 2025). However, legal costs compete with the transition for funds.

The third works the other way: the EU Steel and Metals Action Plan treats primary steel as strategic for infrastructure and defence (European Commission, 2025), a case for keeping IJmuiden open; strategic importance, however, brings political backing, not capital.

The purpose is explicit and since September 2025 written into a Joint Letter of Intent (JLoI) with the state, non-binding until the tailor-made agreement is signed. The JLoI sets the DRI/EAF route, first units around 2030, and climate neutrality by 2045 at the latest (Tata Steel Nederland, 2025; Tweede Kamer, 2025). The state offers up to €2 billion plus €200 million for the fuel switch (Tweede Kamer, 2026); phase one costs €4–6.5 billion, of which €2.3–4.0 billion is planned as external debt (Tata Steel Nederland, 2026). In direction, P and C are aligned. However, the company lost €206 million in 2025/26, operating cash flow of €473 million covers €251 million of capital expenditure, not a programme twenty times larger, and the auditor flags a material uncertainty about going concern because the Environment Agency and the province announced in April 2026 their intention to revoke the permits of the coking and gas plants (Tata Steel Nederland, 2026). The purpose is fundable only with debt the parent must guarantee and aid the Commission must approve, under a licence that may be withdrawn first.

The misalignment is time and money: the JLoI aimed at a binding agreement by 30 September 2026, but the date slipped to 1 March 2027 because technical, financial and legal terms are unsettled and the province wants the coking and gas plants closed earlier (Rijksoverheid, 2026); the union adds that the parent has not committed (FNV, 2026).

The transition serves SDG 9, 12 and 13, with SDG 3 decisive locally (United Nations, 2015). SDG 13: the 5.4 Mt cut in phase 1. SDG 3: the earlier closure of coking plant 2. SDG 12: recycled content rising from 21 per cent (17 in 2019) to 30 per cent once the EAF runs (Tata Steel Nederland, 2026). SDG 9: Zeremis, sold before the route exists (Tata Steel Nederland, 2024). Circularity enters through feedstock and energy: the EAF melts scrap with direct-reduced iron, a make-use-remake loop (Ellen MacArthur Foundation, 2013; Geissdoerfer et al., 2017), and the power plants bought from Vattenfall on 1 January 2026 recover process gases as electricity in-house (Tata Steel Nederland, 2026). Copper and tin in post-consumer scrap limit its share in automotive grades, so circularity reduces but does not remove the need for DRI (Daehn, Cabrera Serrenho and Allwood, 2017).

C and P show a transition forced by regulation, health and politics, with a purpose pointing the same way; the binding issue is financing and timing (Section 2).

Innovation at Tata Steel Nederland is a subsidiary function within a listed Indian group, which sets the pace of decisions. IJmuiden has a research centre, the product record behind Zeremis, and a management team that designed the DRI/EAF route and contracted Danieli and Tenova (Eurometal, 2024). The capital decision, however, belongs to the board in Mumbai, which has not signed the agreement (FNV, 2026).

This is the ambidexterity problem: the blast furnaces fund the site, the new route consumes that funding, and under weak margins firms default to exploitation (March, 1991; O’Reilly and Tushman, 2008). Zeremis shows the pattern: launched in 2022 on the existing route by allocating reductions already achieved, which needed no new furnace (Tata Steel Nederland, 2022). The SCALE restructuring removes about 1,200 FTE from a workforce of 11,636, effective 1 April 2026 (Tata Steel Nederland, 2026). The site scores ORL 5 of 9: willing and competent, but without a funded mandate.

The primary technology behind embodied Zeremis is hydrogen-ready direct reduction with an electric arc furnace. HYBRIT delivered the first fossil-free sponge iron in 2021 (HYBRIT, 2021) and full-scale plants are under construction in Sweden and Germany (SteelOrbis, 2023; SSAB, 2024): TRL 8 of 9 (scale after Mankins, 1995; author’s assessment). MRL is 6 of 10: shaft furnaces are mature on gas, but hydrogen operation, DR-grade pellets and gigawatt electrolysers are not routine (Vogl, Åhman and Nilsson, 2018). Reducing a tonne of iron needs at least 54 kg of hydrogen (Patisson and Mirgaux, 2020), at €5–8 per kilogram in Europe (IEA, 2024) against the €2–3 at which hydrogen steel competes (Vogl, Åhman and Nilsson, 2018). The levels sum to 19, the Pilot and Monitor zone (Appendix G).

Each digital technology maps onto a constraint. Automation is mature across the works; AI and IoT are operational: IJmuiden is a World Economic Forum Lighthouse site for advanced analytics and the company names digitalisation as an enabler of the transition (Tata Steel Nederland, 2026); next come scrap sorting by sensor, against the copper limit, and a digital twin of the EAF for variable power. Blockchain is not deployed: the function it would serve, immutable proof of origin and carbon content per coil (Saberi et al., 2019), is demanded by the CBAM, sanctions and Zeremis audits and met by an allocation system, not a ledger (Tata Steel Nederland, 2024). The digital layer makes Zeremis sellable, not producible.

Figure 2.1 draws the four loops that explain why a ready route is not yet built (Sterman, 2000). B1, balancing, is the policy intention: a higher carbon price raises the furnace cost and the incentive, and the new route then lowers emissions and the bill. R1, reinforcing, is the unintended consequence of that policy: the same bill cuts cash flow and the capital for the DRI unit, delays the decision and prolongs the years in which the bill grows; as allowances fall to 2034, R1 strengthens with time (^t). R2, reinforcing, is the Zeremis loop: binding contracts with a premium improve the business case, bring the decision forward and raise the green volume that attracts more contracts. B2, balancing, is political: a delayed decision deepens the crisis, which raises state support and brings the decision nearer, after an approval delay.

Three constraints hold the decision back, capital inside R1 and energy and feedstock on the path from decision to first coil, each with a leverage point in Meadows’ (2008) hierarchy: parameters lowest, stock-and-flow structures higher, rules and decision rights higher still. Capital: phase one costs €4–6.5 billion against a €2 billion state cap; operating cash flow was €473 million last year against €2.3–4.0 billion of debt still to be raised (Tata Steel Nederland, 2026). The company can carry the transition only on debt the parent must guarantee, so the leverage point is a rule on who decides and bears risk: a ring-fenced vehicle with the state as co-investor, structural ambidexterity in O’Reilly and Tushman’s (2008) sense. Energy: the EAF needs about 3 per cent of the offshore wind power expected by 2032 and green hydrogen at scale is absent (Tata Steel Nederland, 2023); the leverage point is a rule, priority grid allocation. Feedstock: automotive-grade Zeremis tolerates limited scrap and needs premium pellets (Material Economics, 2019); the leverage point is a parameter: pellet offtake and scrap pre-sorting. Capital is decisive: whoever carries the risk while R1 runs decides whether B1 closes.

Can the organisation carry the transition? Technically yes, financially no: capability, engineering and product are in place, and Zeremis proves customers buy the outcome. The binding constraint is Resources, and the decision that releases them lies outside the Dutch organisation.

The sustainability initiatives are responses to constraints, not one plan. The first was the route itself: carbon capture on the blast furnaces (Everest) was dropped in 2021 for hydrogen under pressure from unions, residents and the state (Argus Media, 2021), and the HIsarna pilot, which still needs coal, was not scaled up; only direct reduction removes the coking plant, the main source of local exposure, so health and climate pointed the same way. Capture returns in the JLoI only as an interim measure on the gas-fired DRI plant (Tweede Kamer, 2025). The second is Zeremis Carbon Lite, launched in July 2022 at a 30 per cent allocated reduction verified by DNV, 90 per cent by 2024, a bridge until embodied green steel exists (Tata Steel Nederland, 2022; 2024). The third is the health package in the state agreement: earlier closure of coking plant 2, covered raw-material storage and a DeNOx unit (Tata Steel Nederland, 2023; Tweede Kamer, 2025). However, each response was triggered from outside, landscape pressure on an incumbent in Geels’ (2002) terms, an organisation that adapts rather than leads; Zeremis, a market initiative, is the exception and the roadmap builds on it.

Five parties the company needs but does not control surround IJmuiden, its parent among them (Table 3.1). The state and the Commission finance: up to €2 billion plus €200 million for the fuel switch, state-aid approval pending and the binding agreement due by 1 March 2027 (Tweede Kamer, 2026; Rijksoverheid, 2026). The infrastructure partners, TenneT for the grid and the hydrogen developers in the North Sea Canal area, set when the EAF runs green; the process-gas plants are Tata’s own since January 2026 (Tata Steel Nederland, 2026). The customers are the automotive chain: GEDIA shows Zeremis offtake exists, but volumes and premiums are not public (Tata Steel Nederland, 2024). Upstream, every ore, coal and scrap input must carry proof of melt and processing origin to exclude Russian iron (Council of the European Union, 2014; Belastingdienst/Douane, 2024); Zeremis customers ask the same certificates.

Table 3.1. Partners IJmuiden depends on: what the company needs, who decides and the status at November 2026.

Value in such a transition lies in relationships rather than assets, because no partner can close the loop alone (Dyer and Singh, 1998; Powell, Koput and Smith-Doerr, 1996). On the public record Tata Steel collaborates well bilaterally, with engineering suppliers and single customers, but weakly as a system: state, grid operator, hydrogen developers and customers each negotiate with Tata, not each other. The Growing with Green Steel consortium of thirty companies and universities, funded by the National Growth Fund, is the exception, but pools knowledge, not capital or offtake (Tata Steel Nederland, 2026). Peers show the difference (Appendix G): thyssenkrupp decided with €2 billion of approved public funding (SteelOrbis, 2023), SSAB with its HYBRIT partners and its own balance sheet (SSAB, 2024); ArcelorMittal cancelled its German DRI/EAF projects in 2025 for lack of a hydrogen and offtake case (Argus Media, 2025).

The four loops of Figure 2.1 connect environment and company; R1, the carbon-cost loop, dominates. Table F.5 puts it in euros: at €80 per tonne the bill on 11 Mt rises from €22 million in 2026 to €427 million in 2030 and €880 million in 2034, passing last year’s capital expenditure in 2030 and operating cash flow in 2031; with phase 1 running from 2030 the 2034 bill is €448 million. R2, the Zeremis loop, alone finances the transition from the market, not the state, and it is weak because the embodied product does not yet exist and the company reports the financial effect of Carbon Lite as not material (Tata Steel Nederland, 2026); the premium itself is affordable, as full pass-through adds under 1 per cent to a car’s price (Rootzén and Johnsson, 2016). At reported green-steel premiums of €100–150 per tonne (CRU, 2024; Muslemani et al., 2021), two contracts of 0.25 Mt a year would add €50–75 million, a quarter to over a third of last year’s loss and an eighth to a sixth of the 2030 carbon bill, enough in the author’s judgement for one point on Resources (Table F.2 states the condition; the sixth point comes from state and parent). B2, the political loop, raises support as the crisis deepens, the offer grew from €2 billion in the JLoI to €2 billion plus €200 million by April 2026 (Tweede Kamer, 2026), but each increase lengthens approval. All share one delay, the years between decision and first green coil, in which R1 runs unopposed; hence Section 4 puts the decision, not the hydrogen, first.

Table F.5. The carbon-cost loop in euros: ETS bill at €80/t as free allowances are withdrawn (CBAM factor, Directive (EU) 2023/959); reference points 2025/26: operating cash flow €473m, capex €251m. Lower bound (free allocation assumed equal to 2025 emissions). Author’s calculation.

The environment helps in direction and hinders in time: every partner wants the same outcome but acts in sequence, each waiting for another, and the company cannot force the order; Section 4 turns that into a score and a roadmap.

The six dimensions and the time factor are scored 0–10; Appendix F gives the evidence and why each score is not higher. Conditions and Purpose score 8: pressure and targets are unambiguous, but the JLoI is not yet binding and the parent has not committed capital; P stays 8 in 2030 because phase 2 tests that commitment again. Structure, Environment and Networks score 6: decision rights lie abroad, approval and infrastructure keep their own clocks and coordination is bilateral. Resources score 4, the binding constraint. Time scores 5: the technology sits in the Pilot and Monitor zone and every year without a decision lowers it.

Innovation capacity is I = (C × P × S × R × E × N) × t / 10,000,000 (Papadopoulou, 2026); today (8 × 8 × 6 × 4 × 6 × 6) × 5 / 10,000,000 = 2.76 per cent, and one low score pulls the product down. With the roadmap executed (R 6, S 7, E 7, N 7, t 7) the 2030 value is 9.22 per cent, mostly from relieving Resources. The weighted timing model scores 6.75, Pilot Soon (Appendix G); the gas-fired first unit (TRL 9, MRL 8) is held back only by the same ORL of 5.

Table F.2 · Innovation capacity scorecard

Three scenarios (Table F.4) test the strategy on two trends, the hydrogen price, above €3 per kilogram in Europe before 2030 (IEA, 2024), and the carbon price as free allowances end (European Parliament and Council, 2023b); hydrogen direct reduction is the route the reviews identify (Rissman et al., 2020), with the first full-scale plants under construction (Eurometal, 2024). Optimistic: hydrogen below €3 by 2030, carbon above €100 per tonne, a €150 Zeremis premium accepted (CRU, 2024) and approval in 2027; capacity 16.06 per cent, the route essential. Baseline: hydrogen at €4–5, carbon at €80–100, a partial premium and a late grid; the route essential but phased, gas first, at 9.22 per cent. Pessimistic: hydrogen above €6, no premium, stalled state aid and imports undercutting despite the CBAM; purpose weakens (P 7) as the targets hang on the agreement; capacity 0.63 per cent, and primary steelmaking ends.

One configuration is the least-regret choice in all three: a gas-fired, hydrogen-ready DRI/EAF unit, built now, converted later: a real option, the right to switch fuel when hydrogen is cheap (Dixit and Pindyck, 1994). Under baseline trends but without a decision the 2030 score falls to 0.72 per cent (R 3, S 5, E 5, N 5, t 3); the pessimistic case is lower as purpose erodes with the aid. Five recommendations follow. First, take the investment decision within twelve months, with a kill threshold: no signed agreement by 1 March 2027 means re-scoping to a scrap-based EAF. Second, convert Zeremis intentions into binding offtake contracts, two of 0.25 Mt a year at €100–150 per tonne indexed to the carbon price, so R2 carries an eighth to a sixth of the 2030 bill. Third, build a traceable scrap and pellet network, which sanctions, the CBAM and audits already demand. Fourth, negotiate grid capacity and hydrogen supply as one package with the agreement. Fifth, ring-fence the transition in a vehicle with the state as co-investor. The first, fourth and fifth sharpen the JLoI; the second and third are the author’s own, the levers the company controls. Each has a side effect the KPIs must watch: a scrap network tightens the EU scrap market and raises copper in the mix; state co-investment shifts risk to the taxpayer; a 2030 EAF loads a grid TenneT rations.

Table F.4 · Scenarios

The roadmap follows the JLoI; KPIs are in Appendix F (Tata Steel Nederland, 2025; Tweede Kamer, 2025). Phase 0, twelve months: sign the agreement, take the investment decision, close two Zeremis contracts, file the grid request. Phase 1, to 2030: build DRI unit 1 and EAF 1 on gas with interim carbon capture, close coking plant 2 and blast furnace 7, raise recycled content to 30 per cent and ship the first embodied coils, cutting at least 5.4 Mt. Phase 2, from 2030: switch the DRI plant to hydrogen or biomethane once sufficiently available, from 2033 in the JLoI. Phase 3, to 2045: replace coking plant 1 and blast furnace 6, self-financed, to climate neutrality. Luxembourg is the historical comparison for the time factor: the 1975 crisis to the last blast furnace in 1997 took twenty-two years, the exit negotiated by state, employers and unions (ArcelorMittal, 2023); IJmuiden has eight years to the end of free allocation, which is why t cannot score above 5 while no agreement is signed.

1. Phase 0Next 12 months

   Sign the binding agreement; final investment decision on DRI unit 1 + EAF 1; two binding Zeremis contracts; grid request filed as critical path.

2. Phase 12026–2030

   Build DRI 1 and EAF 1 on natural gas with CCS; close coking plant 2 and blast furnace 7; recycled content 21 → 30%; first embodied green coils; −5 Mt (40%) from 2030.

3. Phase 2From 2030

   Switch the DRI plant to hydrogen or biomethane as soon as either is sufficiently available; the JLoI places this from 2033 onward.

4. Phase 3To 2045

   Replace coking plant 1 and blast furnace 6, self-financed; climate neutral by 2045 at the latest.

Table F.3 · Roadmap KPIs

The score is 2.76 per cent today, 9.22 with the roadmap, 0.72 without it, limited by Resources: ready in intent, not in means. The path that closes the gap is a gas-fired, hydrogen-ready first phase, jointly financed and decided before free allowances fall further.

This reflection follows Gibbs’ (1988) cycle. I wrote this report in the last three weeks of the module: I chose Tata Steel on 21 October, in week 8, drafted Section 1 on 21 October and rewrote it in the week of 26 October, drafted Sections 2 to 5 on 2–4 November, and ran AI-based examiner-style reviews on 4, 5 and 6 November (Appendix B). Three Professional Development goals framed the work: a distinction-level standard of evidence, strategic systems thinking, and a weekly OKR check-in. The defining moment was the lecture of 26 October, when the AB InBev example was shown: every dimension scored with evidence, multiplied into one innovation capacity, every number recomputable. Until then my draft had treated scoring as a summary; from that day it became the spine, and I rebuilt Section 4 around the product formula, the timing model and the scenarios.

My strongest feeling was frustration, in week 8. Research Skills had trained me to pose a main question and sub-questions, and the strategic-report format seemed to leave my analysis without an anchor. I wrote several sets and discarded them all. I asked no one; I reread the AB InBev example as a reader rather than a student and saw that it needed no question, because the framework organised the argument. The frustration turned when I settled on one guiding question, answered at the end of each section: the same discipline without a research paper. A second, quieter feeling was unease at working with generative AI drafts: fluent text made it harder, not easier, to see where a figure was unverified. That settled only once I had the Annual Report and could check the figures line by line.

The systems analysis went well: drawing the carbon-cost loop early gave every section a mechanism to return to, and the TRL/MRL/ORL exercise located the binding constraint before I wrote the roadmap, the skill my PDP targeted. Source discipline went less well: figures from news articles entered my draft before I checked their primary sources, and I completed the source log only in the final week. My OKR check-in tracked output but not verification, a gap in the goal itself.

The frustration of week 8 was the learning outcome: a strategic analysis does not answer a question posed in advance; it builds a judgement from evidence, with the framework as the organising device. In KSA terms, my knowledge of global issues and technology was there from the start, my systems-thinking skill developed only when I drew and scored the loops, and forward thinking appeared when the scenarios forced me to one least-regret choice. Of my three Inner Development Goals (Jordan, 2021), complexity awareness grew most: discovering that the carbon price both drives and delays the transition replaced my linear view of policy with a systemic one. Openness was tested when the AI examiner-style review of 4 November graded my draft a 5; accepting that and rewriting was harder than the analysis. I recognise the pattern from an earlier group project in which I took the lead too late: choosing the company in week 8 was the same reflex, adapting first and leading late. The distinction goal pushed me towards thoroughness in analysis and haste in sourcing; the second needs method, not ambition.

The report would have been stronger had I opened the source log in week 2 and benchmarked the scores against a peer before finalising them; the benchmark came only in the last week and could test the scores, not set them. Curiosity, in the IDG sense, paid: the Luxembourg comparison and the Zeremis focus came from questions outside the brief.

In block 2 I apply four actions, each with a check. Sources first, then writing: I open the source log the day the case is chosen and record every figure with its primary source before it enters the text, with a verification line in my OKR check-in (all figures sourced by mid-block). I benchmark every readiness score against one peer before finalising it. I choose the case and draw the loops in week 2 and revisit them in week 6 (complexity awareness). I ask one classmate for a ten-minute read of every draft (openness and resilience).