Eikon Therapeutics

The differentiating asset is an instrument, not a model. What Eikon has that others do not is super resolution live cell fluorescence microscopy descended from Nobel recognised single molecule imaging work, capable of tracking individual proteins moving inside living cells rather than inferring behaviour from fixed or lysed samples.

Machine learning is genuinely load bearing, because extracting protein dynamics from single molecule tracking data at scale is not tractable by hand, but it interprets what the optics produce rather than constituting the advantage. This sits close to the index's platform not algorithm precedent, graded B rather than C because the analysis layer is integral to the measurement being usable at all.

The platform is a screening and target characterisation engine feeding medicinal chemistry decisions made by scientists, with imaging measurement providing an empirical anchor at the front of the process rather than a prediction that must later be checked. Held at B because no disclosure was located on how hits are triaged, what confidence attaches to a dynamics based finding, or where human judgment gates progression, so the oversight is inferred from the workflow rather than documented.

The measurement approach is described clearly and the computational layer is not. Public materials explain super resolution fluorescence microscopy, single molecule tracking in living cells and the optical lineage behind it, which is more physical detail than most in this category provide.

What was not located in this review is any description of the machine learning architecture, any benchmark of the platform against conventional screening, or a peer reviewed methods paper establishing that dynamics based screening finds what other methods miss. For a company whose entire thesis is that it can see something others cannot, that is the disclosure a buyer most needs.

The corpus is generated in house rather than sourced from patients, which is a genuine structural answer and worth distinguishing from the vendors in this lane whose platforms rest on acquired human data. The platform operates on cellular imaging data produced by the company's own instruments in its own laboratories, so there is no acquisition to disclose, no consent framework to examine and no third party dataset whose provenance a partner would need to check.

That is a materially cleaner position than an inherited corpus and it should be credited rather than treated as an absence of information. Human subject protection in the ongoing oncology trials runs through trial protocol and site ethics oversight rather than through vendor policy, which is the correct arrangement.

What is absent is enumeration on the computational side, with no model or model family named, no hosting arrangement described and no sub processor list located, and the imaging data volumes involved in single molecule tracking are large enough that where they are processed and stored is a real question rather than a formality. The partner side is also unaddressed: nothing states whether partner supplied targets or compounds contribute to models used for others. Ask where imaging data is processed and retained, for a sub processor list, and for the partner data position.

The most important finding in this record, and it inverts how the company reads at first glance. Eikon has four candidates in clinical trials including a Phase 2/3 lead, which looks like the strongest clinical position in this category. It is not evidence about the platform.

EIK-1001, the TLR7/8 dual agonist in Phase 2/3 for advanced melanoma and stage IV non small cell lung cancer in combination with pembrolizumab, was acquired from a partner, and the two PARP inhibitors in early stage trials, EIK-1003 and EIK-1004, were licensed from Impact Therapeutics. None of the clinical assets was generated by the imaging platform.

The platform derived pipeline, described as more than 15 in house programmes, remains preclinical as of this review with no disclosed clinical candidate. This is the same failure mode the index flags elsewhere in the portfolio: clinical stage status obtained by asset acquisition does not validate the discovery engine, and a buyer assessing the platform must separate the two.

Not applicable in the provider sense and rated accordingly rather than penalized. The platform operates on cellular imaging data generated in house with no patient records in the discovery workflow. Human subject protection in the ongoing oncology trials runs through trial protocol and site institutional review board oversight rather than vendor policy.

Not applicable. The company is a clinical stage therapeutics developer rather than a business associate of covered entities handling protected health information.

Converted from Not Rated. The prior note reasoned to the right place and the retrieval had not been done. It is now done, and the answer is that the obligation exists but the document does not yet.

The company completed an initial public offering on the Nasdaq Global Select Market under the symbol EIKN, with trading beginning 5 February 2026 at 18 dollars per share for gross proceeds of approximately 381 million dollars. It is therefore a reporting issuer, and the annual report cybersecurity item that supplies a graded disclosure for the listed vendors in this category applies to it in principle. In practice its first annual report is not due until early 2027, so no such item exists yet. The registration statement carries risk factors only. A risk factor is drafted to protect an issuer from shareholder claims and evidences no controls, so it cannot stand in for the governance item that names processes, oversight and an accountable executive.

This is the second vendor in this category to sit in exactly that gap, after Generate Biomedicines, and both listed in February 2026. Recheck this row when the first annual report is filed.

No SOC 2, ISO 27001 or equivalent attestation was located and no trust centre was found. The company publishes a website privacy policy covering visitors to its site. It is a generated template rather than a drafted instrument: its opening sentence still carries an unfilled article placeholder where the entity type should read, and it states that the company collects any and all information entered on the website. It says nothing about the research platform, partner material or the compute environment.

Two similarly named companies publish their own privacy and security pages and rank on the same queries, Eikonoklastes Therapeutics and Eikon X. Confirm the domain resolves to eikontx.com before crediting anything to this company.

Real and reasonably advanced regulatory activity at the asset level, with the platform correctly not presented as a regulated device. EIK-1001 is in Phase 2/3 trials in advanced melanoma and stage IV non small cell lung cancer in combination with pembrolizumab, and EIK-1003 and EIK-1004 are in early stage trials.

The qualification that belongs alongside this grade is the one on the evidence axis: these are in licensed and acquired assets, so the regulatory record demonstrates the company's development capability rather than its discovery platform. No approval has been obtained.

Converted from Not Rated after a second search directed at publications rather than the product site. Material was found, and it changes what this row can say, though not enough to move it past C.

The platform is published. A high throughput single molecule tracking method is described in a peer reviewed venue that uses the open review model, under which the referees' assessments are posted permanently alongside the paper and the authors cannot suppress them. The reported scale is imaging on the order of a million cells per day and screening on the order of ten thousand compounds, demonstrated on a fluorescently tagged estrogen receptor. Publishing into a venue where the criticism travels with the claim ranks above a vendor generated figure on a vendor selected benchmark, and most of this category does not do it.

What the publication does is demonstrate capability. It is not a characterisation of limits, and the limits are the governance question here. The domain relevant issue for an imaging led platform is what the measurement itself can and cannot see. Single molecule tracking requires attaching a fluorescent label to the protein under observation, and independent published work establishes that the label can measurably alter the behaviour being measured, with bulkier labels sterically hindering lateral mobility and changing observed interaction frequencies. The concern is therefore specific and documented rather than hypothetical. No systematic account was located of which targets, labels or cellular contexts this platform handles poorly.

The same absence applies to the cell lines used, whose genetic backgrounds shape what any cellular screen can find, and which are not described.

What would move this row is an applicability statement: the target classes and labelling strategies where the measurement is reliable, and those where it is not.

The measurement approach is described clearly and the differentiating claim is untested, which is the whole assessment. Public materials explain super resolution fluorescence microscopy, single molecule tracking in living cells and the optical lineage behind the instrumentation, which is more physical detail than most of this category provides and is genuinely creditable: a reader can understand what is being measured and why the measurement is hard.

What was not located is any description of the machine learning architecture, any benchmark of the platform against conventional screening, or a peer reviewed methods paper establishing that dynamics based screening finds what other methods miss, and no warranty, indemnity or remediation commitment.

For a company whose entire thesis is that it can see something others cannot, a comparison against what others do is the disclosure a buyer most needs and the only test that could establish the claim. Describing a better instrument is not the same as showing it finds better molecules, and the history of drug discovery contains many technologies that measured something real and did not improve outcomes.

The comparison is also straightforward to design: run both approaches against the same target set and report what each found. Nothing indicates that has been done or published. Ask for a head to head against conventional screening on shared targets, the hit rate and its denominator, and any prospective evidence that dynamics based hits progress further.

Not applicable. This is a discovery and clinical development organization with no provider workflow surface and no EHR touchpoint.

Not applicable in the software sense. The imaging platform is capital intensive instrumentation operated in the company's own facilities and is not licensed, deployed or otherwise made available to customers, so tenancy and residency questions do not arise. The commercial model is the company's own pipeline rather than platform access.

Capital history is unusually well documented and the company has now entered public reporting. More than 1 billion dollars was raised privately since the 2019 founding, including a 517 million dollar Series B in 2022 and a Series D of roughly 351 million dollars, followed by a Nasdaq listing under EIKN in a 2026 offering reported at 381 million dollars, with a registration statement on file.

Held at B rather than A because the listing is very recent so the reporting record is short, and because the material commercial fact for anyone assessing the technology, that the clinical pipeline is in licensed rather than platform derived, has to be assembled from asset level disclosures rather than being stated plainly in the platform narrative.

Concentrated on oncology and on small molecules. Disclosed clinical work covers advanced melanoma, non small cell lung cancer and PARP inhibition in DNA repair deficient tumours, and the internal programme count of more than 15 sits behind that without disclosed indication breadth.

The underlying imaging capability is in principle applicable to any biology where protein movement matters, but no application outside oncology was located, so buyers should treat breadth as potential rather than demonstrated.