Skip to content

Live book · scored in public · open methodology

A frontier model reading the US tape.

Most research hides its book and never keeps score. We publish the names we hold and the ones we're circling — each with the exact trigger that would prove it wrong — then resolve them in public, played-out or invalidated, and Brier-scored. Falsifiable, dated, nothing to sell.

The state of the research

Everything, at a glance.

Track record

251 resolved

Scored in public

251resolved · Brier 0.26

112 played out · 139 invalidated — each vs its own published kill level.

What changed · 7d

47 new · 51 resolved

Latest moves

The book · held + circling

28 held · 33 circling

What we hold

Held · 28

Circling · 33

Hot · momentum

10 themes accelerating

Strongest trends

orbyd momentum board — the top 5 US names by 3-month price momentum, as of 2026-08-07
# Name 3m 1m Accel
1 UTZ +85.0% +83.6% +82.8%
2 EFOR +55.3% +74.6% +85.7%
3 BXC +69.5% +73.3% +75.5%
4 HELP +113.7% +72.1% +47.9%
5 TEAM +63.6% +66.6% +68.4%

Catalysts · this week

156 this week

On the clock

LGN

Legence Corp.

fda

today

Applied Optoelectronics, Inc.

earnings

in 1d

Alto Ingredients, Inc.

earnings

in 1d

ASML Holding N.V.

earnings

in 1d

+152 more this week

Themes in motion

10 accelerating · 9 maturing

Clusters picking up

Source: orbyd nightly pipeline — computed from US market close data, as of (Europe/Berlin).

Ranked · highest conviction first

Dossiers in rotation

All 756 →

HPE

fresh

Hewlett Packard Enterprise Company

MEDIUM

Five straight up sessions took HPE to a $53.22 close on 2026-08-07 with the 2026-08-04 gap floor at $50.24 still untested; Trefis put the five-day move at +19.8% against +5.6% for the S&P 500, so this leg is no longer pure sector beta.…

Special situation ai-datacenter-infrastructure ai-enterprise-software
catalyst ·

HUM

fresh

Humana Inc.

MEDIUM

Post-print floor held: the 2026-08-05 undercut to $353.69 closed green on CVS's Aetna guide raise, and 2026-08-07 closed $385.00 (+4.84%), back above the 50-day of $374.23 and the highest close since 07-28. Morgan Stanley's 08-05 upgrade…

Compounder managed-care-health-services
analysed ·

MRCY

fresh

Mercury Systems Inc

MEDIUM

Defense-electronics turnaround still re-rating on record bookings and a promised FCF turn, but the stock has given back ~20% from its $128.45 ATH and now chops near the analyst median (~$103) into a binary ~Aug 10-18 Q4/full-year print.…

Special situation defense-aerospace cyclical-industrials
catalyst ·

MU

Micron Technology, Inc.

MEDIUM

The 2026-08-03 divergence extended: MU closed $892.67 on 2026-08-04, +21.8% over four sessions, and Counterpoint put Q2 DRAM share at Micron 25% versus SK Hynix 26%. Price remains under a falling 50-day near $958 with no confirmed base;…

Cyclical recovery ai-chips-memory
catalyst ·

OKTA

fresh

Okta, Inc.

MEDIUM

Identity-security re-rating in its second, sell-side-led leg; the upgrade wave is still building nearly two months post-print — Wells Fargo lifted its target to $150 from $100 on 2026-07-20, collapsing the low-end dispersion. Cyber theme…

Defensive cybersecurity ai-enterprise-software
catalyst ·

RXO

fresh

RXO, Inc.

MEDIUM

Freight-cycle recovery intact; sell-side has fully caught up into a $20–35 battleground — bears $20 (Goldman 07-16, Susquehanna 07-14), bulls $30–35 (BMO $35 initiation 07-14, Stifel/Truist/Citi $30). Narrative matured from mispriced to…

Compounder freight-logistics managed-care-health-services
catalyst ·

Themes · accelerating

Clusters picking up.

Themes in motion →

AI enterprise software

ACCELERATING
confidence · 48%

Broad price-target-raise flow across app software: UBS and TD Cowen lifted Atlassian (08-07/08-08), three houses raised Paycom (08-06/08-07), Benchmark and HSBC upgraded Unity (08-07), Citizens upgraded Expensify after its Q2 revenue beat (08-06/08-07), Appian raised its outlook on cloud growth (08-06), and Cloudflare drew four PT raises (08-07/08-08). GTM is the counter-case so the cluster's strength is one-month, not quarter-long. Proposed as a new id because the canonical list has no advertising or media category and these names an SSP, a performance marketer, a broadcaster, a global agency and a social platform trade on ad budgets, not on enterprise software or consumer demand. Evidence: QuinStreet jumped over 25% overnight (08-07), Scotiabank upgraded PubMatic to Sector Outperform at $21 (08-07), and Gray beat at $0.21 vs $0.02 and guided Q3 sales to $935M-$965M against $896M (08-07); WPP's H1 EPS fell to $0.02 from $0.27 (08-06), so the strength is US-adtech-led, not agency-led.

Dossiers in this theme →

Binary-catalyst biotech

ACCELERATING
confidence · 30%

A small-cap clinical-stage bid rather than one narrative: Xencor beat at $(0.29) vs $(0.79) on $51.2M vs $18.9M with three PT raises (08-05/08-06), HC Wainwright lifted Precigen to $18 (08-05), Citi raised Amylyx to $26 (08-03) against a Q2 miss with cash runway stated into 2028 (08-06), and HC Wainwright reiterated Buy on LB Pharmaceuticals at $45 (08-03). ATAI, MMED and HELP have no substantive week-of news and sit here on 3-month momentum alone (80.5, 65.0, 113.7) the low confidence is deliberate, and the cluster is too heterogeneous to support a directional claim.

Dossiers in this theme →

Consumer discretionary rotation

ACCELERATING
confidence · 40%

GigaCloud beat at $1.65 vs $0.81 and guided Q3 above consensus (08-06), Corsair surged on beat-and-raise results (08-07), and FIGS drew three PT revisions in one session (08-07). The offsets are equally concrete Utz missed on both lines (08-05), Nike was downgraded to Underweight (08-04), and ELF and HNST were named in a 'may implode in Q3' screen (08-07) so this is a name-by-name bid, not a sector rotation; UTZ and OTLY are packaged food housed here for lack of a staples cluster.

Dossiers in this theme →

The live book · held + circling

What we hold, in the open.

The live book →

Not a logo wall published a quarter late — the names we hold right now, and the ones we're circling, each with its thesis and the trigger that would prove it wrong. Live and reasoned — each name dated, with the trigger that would prove it wrong.

Discovery views

Find names by what they're doing.

Most recent journal entry

Saturday, August 8, 2026

Full entry →

Excerpt

RISK-ON the regime engine's latest close-of-day read, carried into the weekend (markets closed). The regime engine's latest read is RISK-ON, unchanged from the prior read. Nothing in the underlying inputs argues with it. VIX closed at 15.81 calm. Breadth is 65.3% of the tracked universe above its 200-EMA (637 of 975), which is healthy participation rather…

Browse by

Every output the model writes, indexed.

Understand the system

How the model thinks.

News window

30d

per candidate, daily

Earnings transcripts

4Q

synthesised per name

Archetypes

8

behaviour profiles

Context

1M

side-by-side candidate review

Common questions

What orbyd is, in plain terms.

What is orbyd?
orbyd is an AI equity-research analyst. Every trading day, frontier language models — Anthropic's Claude Opus and Sonnet — read the US stock tape and write per-ticker dossiers, a regime-tagged journal, and weekly macro and sector views. We currently track 756 names, publishing each thesis alongside the trigger that would prove it wrong, then resolving them in public. The reasoning, the names we hold, and the dated record are all public.
Does orbyd have a track record?
Yes. orbyd has resolved 251 theses in public so far (112 played out, 139 invalidated), each graded against the exact invalidation trigger it published in advance — not a return marked to a benchmark after the fact. Across the 250 scored so far the Brier score is 0.26 — 0 is perfect, 0.25 a coin flip. The full Brier-scored record is public on the track record.
What do we publish?
Everything the model thinks. Every thesis, regime call, bull and bear case, and invalidation trigger — the full reasoning — plus the names we hold and the ones we're circling, each dated and scored against the line that would prove it wrong. You see exactly what the model thinks and why, and can hold us to the triggers we published.
What is the current market regime?
The model classifies the macro backdrop into a single regime each trading day, and that call sets how aggressively we buy, how large a position can get, and how much cash we hold. The most recent call, as of 2026-08-07, is Risk-on. The full taxonomy and what each regime gates live on the regimes page; the dated history runs through the journal.
Is orbyd investment advice?
No. We publish educational research under the EU and BaFin framework — a transparent record of how a language model reads the market, not a recommendation to buy or sell anything. Always do your own work.
Is there anything to buy or sign up for?
No. There's no product, paywall, newsletter upsell or affiliate link — nothing to sell. We publish in the open as a falsifiable, dated track record. The incentive is to be right, and to be seen being right.

Built for agents to crawl

No auth. No rate limits. Canonical URLs for every resource.

JSON-LD on every page, llms.txt + llms-full.txt for context windows, JSON Feed + RSS for event-driven agents, WebMCP + OpenAPI for discovery. Cite the canonical URL.

Agent endpoints