Insights › Rare-earth magnet supply chain atlas › Solvay — Special Chem (Rare Earths), La Rochelle

France (plant); Belgium (group headquarters)

Solvay — Special Chem (Rare Earths), La Rochelle

Rare-earth oxide separation (not a magnet, alloy, or metal producer); supplies separated oxides, including samarium and magnet-grade neodymium-praseodymium, to downstream metallizers and magnet makers.

At a glance

Founded

1948 — Rare-earth separation at the La Rochelle site dates to 1948, operated through a chain of ownership from Rhône-Poulenc to Rhodia to Solvay. Solvay demerged from Syensqo in December 2023, retaining the rare-earths business.

Headquarters

La Rochelle, France (plant); Brussels, Belgium (group headquarters)

Ownership

Publicly listed (Euronext Brussels/Paris: SOLB); demerged from Syensqo in December 2023.

Stage

Mature. Has separated rare earths at La Rochelle since 1948.

Listing

Euronext Brussels: SOLB
Investor relations

Revenue

€4.262 billion
Reported, FY2025.
No rare-earth-specific revenue figure is disclosed; the group and Special Chem segment figures are not rare-earth-only.
Source

Capacity

Each date below links to the document the figure came from. A capacity figure with no source is not on this page.

Narrative

Solvay operates a rare-earth separation business at La Rochelle, France, with roots dating to 1948 under a chain of ownership running from Rhône-Poulenc through Rhodia to Solvay; Solvay demerged from Syensqo in December 2023 and retained the rare-earths business. The La Rochelle site separates rare-earth oxides rather than producing magnet alloy, metal, or finished magnets, and supplies separated oxides to downstream metallizers and magnet makers. Under a three-way commercial agreement announced in November 2025, Solvay separates samarium oxide that is metallized by Less Common Metals and then made into magnets by Permag. The site has approximately 4,000 tons per year of broad rare earth-based formulation capacity as of 2022, not specific to magnet-grade material, plus a magnet-grade neodymium-praseodymium separation line that began operating in April 2025; Solvay has not disclosed tonnage for this line, though trade press has estimated output at a few hundred tons annually. Samarium oxide separation began in 2025. In July 2026, Solvay approved a €15–20 million investment in a heavy rare-earth (dysprosium and terbium) separation line, targeted to begin operating in autumn 2026. Solvay does not disclose a rare-earth-specific revenue figure; FY2025 group-wide underlying net sales of €4.262 billion are provided here only as context and are not attributable to the rare-earths business alone. As of September 2026, unconfirmed, single-sourced reporting has described talks between Solvay and OneIM regarding a possible strategic partnership or investment in the La Rochelle rare-earths business; no agreement has been confirmed.

Where public sources disagree

• Investment-figure estimates for the magnet-grade NdPr line vary across secondary sources and are not fully reconciled.
• Broader 2030 capex target estimates for the rare-earths business range from over €100 million to €50–100 million in different secondary sources, distinct from the confirmed €15–20 million heavy rare-earth line investment.
• The NdPr line's tonnage output is a trade-press estimate, not a Solvay-disclosed figure.
• The heavy rare-earth line's targeted start date is described variously as "autumn 2026" and "targeted for September 2026" across sources.
• As of September 2026, a single-sourced report describes unconfirmed talks between Solvay and OneIM regarding a possible strategic partnership or investment in the La Rochelle rare-earths business; no deal has been confirmed.

Latest reports

Across these reports

Across Q3 2025 to Q2 2026 Solvay's group sales, EBITDA and free cash flow declined year on year, and rare earths appear in the results only as volume commentary (lower electronics rare earth volumes in Q1 2026) with no tonnages. On rare earths specifically, Solvay inaugurated its La Rochelle magnet-materials line on 2025-04-08, signed a Viridis feedstock LOI on 2026-06-01, approved EUR 15-20 million more for separation capacity in Q2 2026, and targets Dy/Tb separation in autumn 2026.

Stated dates

Figures are as the company reported them, linked to each report. This summarizes what changed between reports; it is not a view on the company or its shares.

Fact checks from the digest

Claims made about this company that the weekly digest has put against this atlas, newest first. The verdict and the working are here in full; the issue each came from is for subscribers.

1. narrower than claimed · Digest issue 006 · 28 September 2026

   Japan has been building its way out of this since 2011, and the heavy half is eleven months old

   “China's near monopoly” is the framing every version of this story uses, and on the light elements it is simply true. On the heavies it is narrower than it sounds — and then, looked at closely, narrower again in the other direction.

   From this atlas: Lynas is the only company separating heavy rare earths at commercial scale outside China, and it got there recently — it first separated dysprosium and terbium oxides in Malaysia in 2025, and its heavy separation expansion is still under construction at up to 5,000 t/yr of heavy rare earth feedstock. How much dysprosium and terbium that actually yields is not public: Lynas reports heavy rare earths as a group and has never broken the two out. The others are smaller or not yet running. MP Materials' heavy circuit at Mountain Pass is 200 t/yr of Dy and Tb, nameplate, under construction. Solvay's dysprosium and terbium line at La Rochelle is under construction on a €15–20 million approval with no tonnage published. Neo's line at Silmet in Estonia, commissioned in April 2026, is a small-scale demonstration.

   And the fifteen-year framing needs the same discipline. For fourteen of those years this was a light-element arrangement, because NdPr from Mount Weld was what Lynas had to sell. Sojitz's first heavy rare-earth import from Lynas was 30 October 2025 — two months before China began squeezing Japan on dysprosium. Japan's heavy alternative is therefore not a hardened fifteen-year position. It is single-sourced and about eleven months old. Better than having nothing, which is what everybody else has. Nothing like enough.

2. it stands · Digest issue 005 · 21 September 2026

   Lynas is now committed to sending 75 percent of its medium and heavy rare earths to Japan

   Largest non-Chinese heavy rare-earth producer: it stands, and the bar is lower than the phrase makes it sound. On our atlas, MP Materials' heavy circuit is 200 t/yr nameplate and under construction, and Solvay's Dy and Tb line at La Rochelle is under construction with no tonnage published. Lynas is the only one separating heavies at scale today.

   Against China the number is small, and how small is not public: Lynas reports heavy rare earths as a group and does not break out dysprosium and terbium, so the size of what has been committed is known only inside the company. Three quarters of it now goes to one country. Being the largest non-Chinese producer of something is a statement about how little of it exists outside China.

3. it stands · Digest issue 003 · 7 September 2026

   Lynas confirms it was in takeover talks

   It stands, and it is worth seeing exactly how narrow the gate is, because every word in it is load-bearing. MP Materials mines, separates NdPr and makes magnets — but its heavy circuit at Mountain Pass is 200 t/yr nameplate and under construction. Solvay separates NdPr and is building a heavy line at La Rochelle — but it does not mine. Neo separates at Silmet, runs a small-scale heavy line there and now makes magnets at Narva — but it does not mine either.

   So “integrated” and “operating” are what make the sentence true, and both of them expire the moment Mountain Pass commissions its heavy circuit.

4. narrower than claimed · Digest issue 002 · 31 August 2026

   Neo and Carester close a loop between Europe's only heavy separation plant and its only new magnet plant

   I called this the first visible attempt to source European heavy rare earths without a Chinese separation step. Our own atlas says otherwise, twice. Solvay's La Rochelle has a dysprosium and terbium line under construction on a €15 to 20 million investment, and Neo's own Silmet in Estonia already runs a small-scale heavy separation line for Dy and Tb.

   What is actually new is narrower and more useful: this is the first contracted arrangement between a European heavy separator and a European magnet plant, with the swarf going back the other way. The capability existed. The commercial link is the news.

5. contested · Digest issue 002 · 31 August 2026

   A startup comes out of stealth claiming all seventeen rare earths in one pass

   Two claims, and they are not the same. “All seventeen rare earths in a single pass” is a claim about a process. It is not a claim about a capability nobody has: Solvay’s La Rochelle already separates the full suite, today, at commercial scale, which is why it is the plant everyone is circling.

   The second claim — “the only approach that can credibly reach cost superiority relative to China” — is doing a great deal with the word only, from a company with US$3.1 million and a demonstration plant due in eleven months. Ucore, Aclara, Carester and Solvay are all pursuing routes of their own. The honest reading is that a non-solvent-extraction route would be valuable and that nobody, including this company, has published a cost comparison anyone can check.

6. it stands · Digest issue 001 · 24 August 2026

   Iluka's numbers are in the slide deck and not in the announcement

   Nobody claims a record here, but 750 t/yr of separated dysprosium and terbium invites one, so here is the comparison. On our own atlas: MP Materials’ heavy circuit at Mountain Pass is 200 t/yr nameplate, under construction. Solvay’s heavy line at La Rochelle is under construction with no tonnage published. Neo’s heavy line at Silmet is small-scale, commissioned in April 2026, with no tonnage published. Lynas separates dysprosium and terbium today, at commercial scale, but publishes no tonnage for either.

   So the 750 figure is the largest stated heavy rare-earth line outside China that I can find. It is also a design number in a slide deck for a plant that is 60 percent built, which is a different class of fact from what comes out of a running plant — even when the running plant does not publish it.

Sources

1. Company website — Solvay (2026-09)

← Back to Solvay — Special Chem (Rare Earths), La Rochelle in the atlas · the full supply chain atlas