The Chemical Has a Name. The Buyer Needs a Match.

A label can be perfectly accurate and still tell a buyer almost nothing. Consider propylene glycol. The name points to a chemical; it does not settle the grade, purity, package size, documentation, or the plant that can reliably make it. A purchasing team that searches by name may end up comparing a pharmaceutical grade with an industrial one. The spreadsheet looks tidy. The comparison is nonsense.

This is the narrow, expensive confusion that Valdera was built to address. The San Francisco company helps manufacturers find and evaluate suppliers of chemicals, ingredients, and raw materials. Its software starts with the buyer’s actual specifications, then maps them to potential suppliers, manages requests for quotation, tracks responses, and keeps the paper trail in one place. The promise is modest enough to sound almost quaint: fewer blind calls, better options, and less time lost discovering that the promising supplier cannot make the required material.

The search begins before the product

Sruti Arulmani, Valdera’s co-founder and chief executive, traces the idea to a habit that made ordinary shopping unusually laborious. She read the ingredients, safety information, sources, and certifications behind sunscreen, paint, even mattresses. That curiosity exposed a less visible problem upstream: if consumers struggle to understand what went into a finished product, manufacturers face a still harder task deciding which inputs belong in it. She and her brother Dheev, now the company’s chief operating officer, turned that observation into a procurement business.

Dheev brought experience working with industrial and chemical clients at McKinsey. In a 2026 interview, he described beginning with data rather than an AI interface: chemical specifications, supplier catalogs, packaging, and the connections between them. That ordering matters. A general language model can discuss a chemical with pleasing confidence. It cannot reliably infer a supplier’s current capabilities from a vague public listing. Valdera’s bet is that useful automation must sit on top of category-specific records.

“Traditional sourcing methods, like Googling one supplier at a time, just aren’t enough.”Sruti Arulmani, Valdera co-founder and CEO

A database is only the first appointment

Valdera reports more than 400,000 suppliers and 1.7 million chemistries or CAS numbers in its database, spanning 130 countries. Those are company figures, not a count of qualified alternatives for any one ingredient. The distinction is essential. A manufacturer does not need a directory of everyone who can pronounce an ingredient. It needs the short list that can meet a formula, satisfy a certification, ship in the right form, and negotiate commercial terms.

The buyer workflow makes that distinction visible. A team submits a material request, with a custom form if needed. Valdera finds candidates and sends requests for quotation while letting the buyer control when its identity is shown. Quotes and supplier communications can be compared inside the platform. Certifications and documents live alongside the discussion, and buyers can track an order through execution. Suppliers, meanwhile, can sign up for free, receive matched leads, respond by email, and manage bids without publishing their inventory or prices to the world.

This is why Valdera resists calling itself a distributor or an open marketplace. It does not buy the chemicals and resell them. The buyer and supplier remain responsible for their own contract and the material itself. Valdera sells the work around the transaction: search, outreach, market context, and a cleaner way to compare what comes back. Its master services agreement calls the product a SaaS procurement solution; buyer fees sit in private order forms. The supplier side advertises free leads with no markups.

Toray found the map had blank spaces

Toray Industries offers a useful example because it was hardly starting from zero. Its procurement team had long-standing supplier relationships and substantial market knowledge. John Eustis, a senior procurement leader at Toray, said the gap lay in chemicals that had not received the same attention: specialty, often single-sourced items that the team lacked the bandwidth to study thoroughly. Valdera offered a way to see sources of supply the team had not known existed. On Valdera’s site, Eustis describes getting a global view of the market for each specific material.

That is a more interesting use case than a simple claim that AI makes buying faster. The buyer’s familiar suppliers may be excellent. The risk is mistaking familiarity for the whole market. A wider, specification-aware search can reveal a backup source, a regional option, or a price comparison grounded in the right grade. For a company trying to protect production and margin, those are not decorative discoveries.

The awkward part still belongs to people

Supplier discovery is not supplier approval. A candidate still has to pass technical review, regulatory checks, and the buyer’s own rules. Dheev has said that changing a pharmaceutical input may take years of approval, while some industrial changes take months. Valdera can make the search and comparison less laborious; it cannot vote a new ingredient through a quality committee. That limit is also a clue to where the product fits. It is most useful when a buyer has complex requirements, a fragmented supplier field, and enough purchasing volume to justify a repeatable process.

The market is crowded from two directions. General procurement suites handle broad purchasing workflows. Specialist chemical platforms such as Knowde, Covalo, and Kemiex offer discovery or commerce in overlapping categories. Valdera’s particular argument is that the chemical specification should be the center of the system, with sourcing and qualification steps built around it. The company’s data set is the part a generic interface cannot summon overnight.

In October 2024, Index Ventures led a $15 million Series A, joined by existing investors Susa Ventures, Lerer Hippeau, and BoxGroup. That is the disclosed financing behind the expansion, not a public price list or a measure of customer savings. Valdera’s site reports savings of 5 to 15 percent per request; such a figure is best read as a company claim that depends on the material and the buyer’s starting position. No algorithm can find competition where only one producer meets the specification.

The lesson is useful beyond chemicals. Before automating a purchasing decision, decide what an acceptable answer actually looks like. The name on a drum is the beginning of that definition. Grade, purity, origin, certificates, and delivery make it complete. Valdera has built a company around the interval between those two sentences, where a product’s future is often decided before anyone sees the product at all.