The investment thesis extends beyond selling CMP materials

Toyota Tsusho announced on October 21, 2026, that it invested in US semiconductor materials company ChEmpower through stock acquisition rights under a Simple Agreement for Future Equity (SAFE). The size of Toyota Tsusho's commitment and its eventual equity ownership were not disclosed.

The stated objective covers two businesses. Toyota Tsusho plans to supply ChEmpower CMP materials to chip manufacturers in Japan while exploring the collection of liquid waste from semiconductor manufacturing and the recycling of rare metals.

Supplying a new consumable is familiar territory for an industrial trading firm. Taking responsibility for what emerges after the consumable is used is a broader business ambition, not a confirmed commercial recovery operation.

Why CMP has depended on abrasive particles

Chemical mechanical planarization, or CMP, flattens a semiconductor wafer between processing stages so subsequent layers can be patterned accurately. Conventional CMP combines a polishing pad, chemical solution and a slurry containing fine abrasive particles to remove surplus material.

These particles can contribute to scratches and residue under certain conditions. Removal rates, selectivity and uniformity all matter because higher layer counts and tight interconnect structures leave less tolerance for surface defects.

Used slurry and cleaning liquids may contain solid particles, dissolved metals and chemical ingredients. Treatment of these streams creates an additional operation alongside the polishing process itself.

ChEmpower changes the chemistry of the pad, not the existence of polishing

ChEmpower's Chakra platform uses abrasive-free functional fluids paired with chemically reactive polishing pads. Its technology embeds functional chemical groups in the pad and combines these with the fluid to remove material and planarize the wafer without conventional abrasive particles.

Abrasive-free is not polishing-free. The company still sells pads and chemical materials and positions the combination for use with existing CMP platforms. It reports owning manufacturing techniques for the pads and enabling material-specific chemistry variations.

The supplier argues the design can reduce particle-induced defects and improve process stability. Public product claims are not equivalent to an independently disclosed yield gain in production at a named fab.

The first product addresses copper interconnect processing

Chakra Copper Series targets CMP of copper wiring, a key material for routing signals and power within chips. For copper polishing, manufacturers must remove intended material while protecting nearby barrier and dielectric layers.

Controlled removal rates, surface finish and selectivity are central to evaluating the design. A chemistry that works for copper cannot automatically be assumed to work for oxide or other metal processes.

ChEmpower describes a roadmap to wider material coverage. Each step would still require customer qualification rather than converting one technology demonstration into an industry-wide adoption claim.

The wastewater may become easier to process—but it does not disappear

Toyota Tsusho argues that removing abrasive particles also removes their associated disposal burden and may make water and rare-metal recycling from liquid waste easier. The composition of what enters a fab process affects what must be separated afterward.

This does not mean all wastewater, dissolved metals or process chemicals disappear. A practical recycling business would still need to meet recovery-rate, purity, treatment-cost and regulatory requirements.

It is therefore useful to distinguish a technically plausible change in the waste stream from a measured reduction in a customer's total wastewater cost. No completed recovery rate or commercial reclamation margin was announced.

Selling the input and collecting the output demand different capabilities

Toyota Tsusho cites prior experience supplying semiconductor process chemicals as a reason it can bring ChEmpower's materials to Japanese manufacturers. In parallel it wants to build a business collecting waste liquids and recovering rare metals.

Materials distribution depends on consistent specifications, supply reliability and price. Waste-liquid collection requires characterization, compliant transport, separation and purification, and a market for recovered products.

Combining these may create useful knowledge of process inputs and output chemistry, but a second revenue stream is not automatically profitable. Customer contracts, logistics costs and recovery economics must be demonstrated separately.

Five thousand pads of monthly capacity is not five thousand shipments

ChEmpower raised $18.7 million in Series A financing in April 2025. It launched Chakra in February 2026 and its CEO described pilots and trials with memory and logic semiconductor manufacturers in the United States and Asia.

The company's website currently lists US pad manufacturing capacity of roughly 5,000 pads per month. That describes its stated ability to manufacture, not documented sales of 5,000 pads each month.

Likewise, pilot evaluations and a sampling-ready product do not establish qualification as a high-volume commercial fab supplier. Follow-up evidence would need to show customer process adoption, measured yield outcomes and actual supply continuity.

The economic test covers the whole process rather than material price alone

A fab evaluates CMP materials through defect performance, removal rates, selectivity, uniformity, pad lifetime, equipment compatibility, process time and total operating cost.

Even a more expensive fluid or pad might be attractive if verified yield improvements and easier downstream waste treatment exceed the additional cost. The opposite can also occur if integration, consumables or purification are expensive.

The metrics worth tracking now are qualification and deployment in Japanese fabs, consistent reclaimed water and metal purity, recovery cost and contractual economics. The public October investment announcement did not report those outcomes.

BANSEOG VIEW

BANSEOG VIEW | A material's competitiveness can depend on what it leaves behind

Toyota Tsusho's investment concerns new CMP materials, but the business direction extends from delivering those materials to collecting and recycling the process residues. Removing abrasive particles may change both wafer surface conditions and the composition of the output stream.

Most materials businesses compete on the quality, price and supply of inputs. A different possibility is to design materials so that a customer's downstream handling problem becomes easier to solve, then build a service around that change. This requires capabilities spanning chemistry, fab qualification, distribution, waste treatment and resource recovery.

Neither the commercial gain nor operational economics have been proven in the public release. Customer validation and recovery data will determine whether this becomes a durable circular business. The strategic signal is the attempt to connect what enters a semiconductor factory with what leaves it.

SOURCES

Primary sources and references

Prepared October 22, 2026. Toyota Tsusho's Oct 21 release confirms a SAFE-type ChEmpower investment and plans to supply abrasive-free CMP materials in Japan while collecting process liquid waste and reclaiming rare metals. The 2026 investment amount, eventual equity ownership, customer orders, actual recovered material and economic performance are undisclosed. ChEmpower describes functional abrasive-free fluid and reactive Chakra pads for copper CMP; polishing and pads still exist. Its stated 5,000 pads/month factory capability is not sales or fab acceptance. Its $18.7m Series A was a separate April 2025 round, and Feb 2026 memory/logic manufacturer pilots do not prove high-volume adoption or independently measured yield improvement. Eliminating abrasive particle disposal does not eliminate all wastewater treatment. The integrated-input/output business analysis is a strategic hypothesis.