Find the inflection before it becomes consensus.

An LLM can summarize a filing, search for a phrase or help explore an isolated thesis. What it does not reliably create on its own is a stable research universe, consistent definitions across periods and industries, source-linked evidence records, deterministic calculations, cross-document validation, explicit treatment of missing data and a repeatable ranking process. Matterhorn uses AI for extraction and explanation, then relies on structured evidence and deterministic scoring so the result can be inspected and reproduced.

Matterhorn adopts several disciplines common to institutional research: broad-universe screening, standardized calculations, industry-specific operating metrics, source hierarchy, exception handling, theme and scenario overlays, and portfolio-exposure checks. It does not claim to duplicate every hedge fund or research desk; institutions may also have proprietary data, specialist teams, direct management access and execution infrastructure. Matterhorn is designed to make much of the repeatable analytical discipline transparent and accessible to serious self-directed investors.

No. Conventional screeners are strongest when the data is standardized, and newsletters generally deliver conclusions. Matterhorn also examines narrative and industry-specific evidence—such as backlog, RPO, design wins, installed base, contracted capacity and recurring monetization—then exposes the source passages, calculations, confidence and unresolved conflicts behind the ranking. It supports judgment rather than replacing it.

It does not force every company through one universal operating metric. Software may be evaluated through ARR, NRR and RPO; semiconductors through design wins, inventory and utilization; industrials through orders, backlog and book-to-bill; medical devices through installed base, procedures and consumables. Those industry-specific indicators feed common dimensions—compounding quality, forward demand, durability, valuation and portfolio fit—only after definitions and periods are checked for comparability.

Matterhorn does not silently turn uncertainty into a favorable score. Evidence is marked as verified, derived, estimated, unavailable, incomparable, stale or conflicted. Lower-confidence evidence is penalized or held for review, and material definition changes, restatements or source disagreements are surfaced as exceptions rather than hidden inside the final number.

Theme weights let you express a view about forces such as AI capital spending, grid construction, rates, healthcare utilization or the semiconductor cycle. They change the lens through which opportunity and portfolio fit are evaluated; they do not rewrite the underlying filings or evidence. The base evidence remains visible so you can distinguish company facts from your own macro assumptions.

The platform is designed around a weekly universe refresh, with company evidence and rankings updated as new filings and supporting documents are processed. Because the product follows business and reporting cycles rather than intraday price noise, meaningful changes may occur when new evidence arrives—not simply because another day has passed.

Matterhorn is intended to be ongoing research infrastructure. The evidence it follows—filings, backlog, margins, installed-base economics, revisions and capital allocation—develops across quarters and years. Quarterly access provides a full reporting cycle; annual access is best aligned with observing whether the evidence chain is strengthening, weakening or changing.

No. A high score means the available evidence ranks well under the selected model and assumptions. It cannot eliminate uncertainty, valuation risk, macro changes, competitive response, execution failure or loss of capital. Matterhorn is a research tool—not a performance guarantee or personalized investment, legal or tax recommendation.