Scala Biodesign

The design engine is the product. ScalaOS takes a protein and proposes a redesigned version, and selling or partnering on that is the whole business. The company is explicit that the engine runs on three things together, evolutionary data, atomistic physics and AI, so the learned model is one part of the method rather than all of it. The product does not exist without the engine, and the engine is not a model alone.

The platform proposes designs, and the customer then expresses and tests them in its own lab, which is where the checking happens. Inside the software itself there is no described control: where a person reviews or overrides a proposed design, what the system does when one fails at the bench, and which steps run without a human are not set out. The testing is real but it sits in the customer's workflow rather than being a control ScalaOS defines.

The method is named and described in general terms. ScalaOS is said to combine evolutionary data, atomistic physics and AI to introduce many beneficial mutations in a single round, which conveys the approach without naming a model, giving an architecture or parameter count, or characterizing the training data. No version or update practice ties a given design to a dated system.

The platform is presented as the company's own cloud software, and no outside model provider is named as part of it. Beyond that nothing is established: no compute or hosting provider is named, no third party components are listed, there is no subprocessor list, and nothing describes what governs a customer's protein data once it enters ScalaOS.

There is industrial and academic validation, short of peer reviewed platform papers. Scala names users including BASF, Boehringer Ingelheim, Teva, Repsol and the Weizmann Institute, points to more than a hundred scientific publications associated with the underlying method, and publishes case results such as a sevenfold increase in an enzyme's conversion rate in one design round and a stabilized vaccine antigen.

What is missing is a peer reviewed evaluation cited directly for the platform and, for therapeutic uses, a clinical outcome: the evidence is named customers and case studies rather than a controlled result a reader can check in full.

A customer uploading its own proteins to ScalaOS would want to know what happens to them: whether the sequences and structures are retained, used to train the engine, or kept separate from other customers' data. Scala does not describe this. No patient records are involved, so protected health information is not in play, but the handling of a customer's uploaded proteins is left unstated.

ScalaOS takes in protein sequences and structures, not patient records, so United States health privacy rules do not apply to what it processes. Scala states no business associate position and publishes no privacy document reaching the platform service, so a customer uploading proteins has nothing on this to read.

ScalaOS is a cloud platform that takes in customers' proprietary proteins, so a customer's security reviewer would look for an independent attestation: a SOC 2 report, an ISO 27001 or HITRUST certification, or a trust portal. Scala offers none, and there is no security page, so the way the hosted environment is secured is not attested anywhere public.

No device claim is made, and the platform is scoped to match. ScalaOS redesigns proteins at the research and development stage across industrial and therapeutic uses, so any regulatory position belongs to the products its customers build rather than to the design software. There is no clearance to hold and none is claimed.

What Scala shows about the engine is a set of successful redesigns in its case studies. Governance would call for something broader, an evaluation of how often the platform succeeds or fails across a range of proteins, a benchmark, an audit, or a responsible AI policy. A handful of wins is not that measure, and none of those governance artifacts is published.

A customer that expresses and tests a ScalaOS design carries the cost when it does not work, so the natural questions are how often designs succeed and what the company stands behind. The case studies report wins rather than a success rate with a method and a denominator across attempts, and no guarantee, indemnity or correction route comes with a design, so the risk stays with the customer running the experiments.

The product is reached as its own cloud application, not something a customer folds into existing systems, so there is no clinical record surface, as a protein design tool would not need one. Nor is there a research side link: no electronic lab notebook, laboratory information management system or research data platform is named as connected, and no connector or application programming interface is offered.

The software is offered as a single cloud platform, so there is a hosted model but no stated options or isolation to evaluate. Location is implied rather than committed: nothing published names the region a customer's protein data rests in or describes tenant separation between customers sharing the platform.

Scala sells in two ways, access to the ScalaOS software or a partnership on a design project, and points a buyer to a demo request or its scientists for either. What it does not carry is a number: no rate card, tier or unit of charge appears, so the routes to buy are clear while the cost of either is not.

Coverage is named across several protein classes, with case studies behind part of it. ScalaOS is applied to enzymes, therapeutic antibodies, other therapeutic proteins and antigens, and to specific properties within each, expression, stability, catalytic activity, developability and humanization, and the company shows worked examples such as an enzyme conversion improvement and a stabilized vaccine antigen. What is not established is validated breadth across every claimed class, since the examples cover a few cases rather than the full range.