Human-gated · three-tier hierarchy

Rules & thesis

Resolution order: macro thesis → general category rules → company-specific rules (win on conflict) → your Level-3 manual overrides (always win). Claude may draft rule revisions; you approve.

Read the methodology →
Valuation inputs · macro assumptions

Macro assumptions

CAPM cost of equity = risk-free + β × ERP + country risk premium. The GDP ceiling hard-caps DCF terminal growth. Category β is the default; a company rule can override it. The US carries no country premium by construction — it is the mature market the ERP is already measured against.

Level 1 · DCF
Tier 1 · Macro thesis

RadTech

synthesized · Sep 25, 2026

Across RadTech, the forward-looking signals are best characterized as cautious-to-hedged, with profitability uncertainty as the one theme confirmed by every company in the set: ARAY expects net losses in the future and cannot assure profitability; ATNM expects continued net losses and says it lacks resources to run required late-stage trials independently; SRTS reported a $7.7 million 2025 net loss and seeks a 2026 return to profitability while flagging uncertainties; and MIR cites a return to profitability but explicitly warns about sustaining positive net income. A second shared pattern is external policy/trade/regulatory pressure—ARAY cites tariffs and China sales declines plus a new Financing Transaction risk section and IEEPA-related tariff refund claims; MIR adds new tariff/global-trade, SMR-market, and AI risks; ATNM cites FDA regulatory uncertainty, the 2025 government shutdown, proposed FDA reductions in force, and healthcare reform; SRTS cites reimbursement-code-driven demand but also customer-concentration, inflation, and cyber/AI competitive risks—though the specific mechanisms are company-specific. Shared growth priorities include international expansion and recurring-revenue/services or software (ARAY and SRTS; MIR and SRTS), partnerships/acquisitions (ARAY, ATNM, MIR), and cost/efficiency actions (ARAY, ATNM, SRTS, MIR), while ATNM's cGMP New York facility and cyclotron Ac-225 production, MIR's SMR/data-center nuclear tailwind, and SRTS's Sensus Financial Services/Sensus Link launch are more company-specific signals rather than category-wide confirmations.

MedTech

synthesized · Sep 28, 2026

Across the provided MedTech set, the forward-looking picture is cautiously innovation-oriented but lacks shared financial visibility: ISRG gives no capex or risk-factor changes and only broad confidence in synthesizing biology, computing, imaging, algorithms, and robotics, plus a Quintuple Aim focus on outcomes, experience, access, and lower total cost per episode; TWST offers no revenue, margin, or EPS guidance and warns it may never achieve profitability, while detailing platform priorities in DNA/RNA/protein synthesis, NGS sample prep, biopharma services, moving up the value chain, and AI/mRNA/proteins. The only cross-company thread is a platform/technology-convergence and value-or-access orientation, confirmed by both ISRG (biology plus computing, imaging, algorithms, robotics; Quintuple Aim) and TWST (DNA synthesis plus AI; affordable synthetic DNA and global expansion), though ISRG is too thin to corroborate most themes; TWST-specific signals include tariffs, the federal research grant halt, AI regulatory uncertainty, the LDT Final Rule rescission, and explicit profitability/loss caution. No category-wide capex, margin, or guidance pattern is supported by more than one company.

Pharma

synthesized · Sep 28, 2026

With only MRNA represented in the provided Pharma content, there is no multi-company corroboration to support a category-wide macro thesis; every forward-looking signal here is company-specific to MRNA. MRNA management is cautiously optimistic on revenue growth in 2026, platform potential, geographic expansion, new launches, cost efficiency, and prioritized pipeline readouts, while explicitly expecting future losses after 2025 and 2024 net losses and flagging execution and commercial risks in a competitive market. Its capex and expansion—new commercial drug product manufacturing and packaging at the MTC campus, completed Marlborough facility for intismeran clinical supply, and fully licensed international facilities in the UK, Canada, and Australia—confirm a company-level push toward manufacturing scale and global reach, not a validated sector-wide trend. Any pharma macro read from this alone would be overreach.

Technology

No thesis yet — run the thesis synthesis job or write one.

Energy

No thesis yet — run the thesis synthesis job or write one.

Claude-drafted · awaiting your approval

Rule-change proposals

  • rule-book-to-billJul 18, 2026
    Current

    Flag when book-to-bill falls below 1.0. Escalate on a second consecutive quarter below 1.0.

    Proposed

    Flag when book-to-bill falls below 1.0 AND backlog declines sequentially. A sub-1.0 quarter with rising backlog is timing, not deterioration, and should stay at 'watch' rather than escalate.

    Reasoning

    Review of the last two ARAY escalations: both sub-1.0 quarters were followed by in-line revenue because backlog held. The rule over-escalated on timing noise. Adding the backlog condition would have kept both at 'watch' and avoided two false-elevated flags.

    Triggered by: Position review after ARAY flag did not translate into a revenue miss.

Tier 2 · General rules (by category)

Leverage discipline

allv3

Flag when net leverage exceeds 2.5x. Escalate above 3.5x or when a refinancing falls inside 12 months.

rule-leverage
Open flags raised (3)
  • 688329.SHNet leverage ~3.66x (net debt CNY 61.2m / TTM EBITDA CNY 16.7m) exceeds the 2.5x flag line and breaches the 3.5x escalation threshold. Gross debt/EBITDA is ~11.0x. Note the small EBITDA denominator flatters the ratio; no debt maturity schedule was provided, so the refinancing-within-12-months escalation cannot be assessed.
  • ARAYNet leverage (~6.2x, roughly $107M net debt against TTM EBITDA of -$18.5M) is far above both the 2.5x trigger and the 3.5x escalation threshold; with TTM EBITDA negative, leverage is effectively non-positive and coverage is impaired. The term loan carries a ~23.9% effective rate and the company has relied on covenant amendments, keeping refinancing/covenant risk live.
  • MIRDebt definitions conflict. Terminal facts show total debt of $444.8M against $418.7M cash — net leverage of roughly 0.14x, comfortably inside the 2.5x threshold. However the Q1 2026 10-Q discloses $1.23B of total debt including $775M of convertible notes; on that basis net debt of ~$811M against $189.6M TTM EBITDA implies ~4.3x, beyond the 3.5x escalation trigger. The $444.8M figure appears to track the term loan only (the 10-K references '$450 million in term loans plus convertible notes'). Reconcile the convert treatment — whether cash-settled, equity-classified, or simply excluded from the data feed — before treating this balance sheet as low-leverage; on the filing's own numbers this is a refinancing-sensitive story, not a fortress.

Last updated Jul 22, 2026

Order momentum

RadTechv2

Flag when book-to-bill falls below 1.0. Escalate on a second consecutive quarter below 1.0.

rule-book-to-bill

Last updated Jul 22, 2026

Revenue concentration

allv4

Flag when a single product or customer exceeds 40% of revenue, or when a credible competitive threat targets that concentration.

rule-concentration
Open flags raised (5)
  • SRTSSingle-customer concentration far exceeds 40%: largest customer was 52% of FY2025 revenue (73% in 2024), and Q1 2026 booked no sales to that historically large U.S. customer; two U.S. customers were 25% and 24% of Q1 revenue. With FY2025 revenue down 34% to $27.5M and Q1 2026 down 59% to $3.4M at a 29.4% gross margin, the re-rating case depends on CMS SRT reimbursement (effective 1/1/2026) reviving demand from a very narrow buyer base — a credible threat to the concentration rather than a diversification.
  • MRNARevenue is dominated by the COVID vaccine franchise, well past the 40% single-product threshold, and that concentrated base is under a credible competitive and demand threat: Q2 2026 product sales fell 18% YoY and the franchise is in a seasonal trough. Diversification hinges on mNEXSPIKE's European ramp and the mRNA-1010 August 2026 PDUFA; until those scale, concentration risk is the dominant vulnerability.
  • ISRGRevenue remains heavily concentrated in the da Vinci robotic-surgery platform, and a credible competitive threat now targets that concentration: J&J received FDA authorization for its Ottava robot. Near-term impact is limited by ISRG's installed base, $3.4B backlog, and da Vinci 5 ramp (1,464 systems), but competitive share pressure warrants monitoring.
  • SPCXTwo customers account for roughly 38% of Q2 revenue — below the 40% single-customer threshold, so the rule has not fired, but the margin of headroom is thin and management itself names concentration a key watch item. AI Solutions & Infrastructure is ~28% of Q2 revenue, also under threshold. Re-test on the next quarter; a single customer crossing 40% or a credible competitive threat to the AI or launch franchise would escalate this.
  • ARAYDisclosure-integrity concern adjacent to concentration: the RPO/backlog restatement from ~$818M/$866M to ~$59M removes previously claimed order visibility, and the FY2025 10-K flags competitive threats and tariff/IT-security risk. Reported remaining performance obligations of $59.1M exclude open orders, so forward revenue concentration is opaque rather than demonstrably below the 40% threshold.

Last updated Jul 22, 2026

Cash runway (pre-revenue)

Pharmav1

For pre-revenue names, flag when estimated cash runway is under 24 months ahead of a value-inflection readout.

rule-runway
Open flags raised (3)
  • DRTSEstimated cash runway of ~3-4 months ($12.2M cash vs. -$42.3M TTM EBITDA) is far below the 24-month threshold for a pre-revenue name ahead of a value-inflection readout.
  • ATNMPre-revenue radiopharma with cash of $36.5M at 6/30/26 against ~$5.9M/quarter operating burn implies roughly 18 months of runway — under the 24-month threshold — and it must span the Actimab-A basket readout (2H:2026/1H:2027), Iomab-ACT SCD data, and the 2H:2026 cGMP plant. Burn will likely rise if registrational trials are funded, and management is candid about needing partners or additional capital; a financing is effectively required before the runway runs out.
  • 2675.HKCash of ~CNY86.0m against a ~CNY62m EBITDA loss and ~CNY89m net loss implies roughly 1.0–1.4 years of runway, well under the 24-month threshold. No value-inflection readout is specified in the facts, but the pending RMB-share issuance proposal signals the company is already arranging to extend its funding. Treat near-term financing as the key swing factor for solvency and dilution.

Last updated Jul 22, 2026

Customer capex cycle

MedTechv1

Flag when leading indicators of customer capex (research/biopharma funding) turn negative for two consecutive periods.

rule-capex-cycle
Open flags raised (1)
  • ARAYLeading indicators of customer capex have turned negative for consecutive periods — US hospital/radiotherapy capital-spending softness, longer installation cycles and budget constraints are cited across the FY2025 10-K, the amended 10-Q and the Q2 FY2026 10-Q, with management guiding continued adverse impact through FY2026. Note the rule is scoped to MedTech while ARAY is tagged RadTech, but the underlying customer-capex dynamic applies.

Last updated Jul 22, 2026

New filing

allv1

Raise an informational flag whenever a new EDGAR filing is detected and ingested.

rule-filing
Open flags raised (11)
  • MIRNew EDGAR filing detected and ingested: Form 8-K filed 2026-07-02 disclosing a compensatory/secondment arrangement for Loïc Eloy (President, Nuclear & Safety Group) — $415,000 base salary, 50% target bonus, 12-month initial U.S. secondment extendable to 36 months, plus customary expatriation benefits. No financial or operational updates were provided; no impact on estimates.
  • SRTSNew EDGAR filings detected and ingested: Q1 2026 Form 10-Q (59% revenue decline, $2.6M net loss, covenant default and Fifth Third facility termination notice), FY2025 Form 10-K ($7.7M net loss on 34% revenue decline), Form 8-K furnishing the Q2 2026 earnings press release (Exhibit 99.1), and Form 8-K reporting the new $15M City National Bank of Florida revolver (matures 2027-06-05, SOFR+3%, 1.50x minimum DSCR, $2.23M cash collateral and blanket lien).
  • ISRGNew EDGAR filings detected and ingested: Q2 2026 10-Q (revenue +18.5% to ~$2.9B, net income $818M), Q1 2026 10-Q (+23% revenue, +48% operating income), an 8-K referencing the July 16, 2026 Q2 press release, and the FY2025 10-K. Informational only.
  • ATNMNew EDGAR filings detected and ingested: an 8-K delisting-risk notification (NYSE American non-compliance with the $4.0M minimum stockholders' equity standard; equity ~$2.3M at 3/31/26; remediation plan due 6/26/26, compliance deadline 11/27/27, trading under '.BC'), the Q1 2026 10-Q (net loss improved to $5.5M on cost controls), the 10-K (2025 net loss $33.9M, accumulated deficit $409.7M), and a prior 10-Q ($53.4M cash at 9/30/25 on $19.3M nine-month burn). No guidance, backlog, capex, or capital-allocation detail was provided in the 8-K.
  • SPCXNew EDGAR filing detected and ingested: 8-K dated August 4, 2026 (Item 2.02 earnings release notice, Exhibit 99.1 referenced but not included in available text). No incremental financial or qualitative content extractable.
  • SPCXNew EDGAR filing detected and ingested: Q2 2026 10-Q (first since the June 2026 IPO). Discloses $7,814M quarterly revenue, $(143)M operating loss, $93.5B cash, $39.4B debt, $41.9B accumulated deficit.
  • TWSTNew EDGAR filings detected and ingested: an 8-K covering the upsized common stock offering (3,125,000 shares plus 468,750 option shares at $96.00, ~$327.1M net proceeds, closing on or about 2026-08-06 under shelf Form S-3), plus fiscal Q2 2026 and Q3 2026 10-Qs. The 8-K is a purely financing event with no operating, guidance, backlog or use-of-proceeds disclosure; the November 17, 2025 Form 10-K remains the substantive risk source.
  • 688329.SHNo EDGAR filing detected or ingested (filing summaries empty); this issuer reports via the Shanghai exchange. Six announcements are queued/not yet ingested, including the 2025 audit report, the 2025 profit distribution plan, the related-party transaction budget for 2026, and the legal opinion on cancellation of unvested restricted shares — the underlying figures here may lag the latest disclosure.
  • MRNANew EDGAR filings detected and ingested for MRNA, including the H1 2026 10-Q and associated 8-K. The 10-Q carries substantive content (net loss ~$2.1B, $950M Q3 2026 settlement payment, $1.5B Ares facility with $600M drawn, $500M minimum cash covenant); the 8-Ks are press-release-only shells with no standalone MD&A.
  • 2675.HKNew filing signals detected and queued: 2026 INTERIM REPORT, INTERIM RESULTS ANNOUNCEMENT (six months ended June 30, 2026), and an INSIDE INFORMATION PRELIMINARY PROPOSAL FOR THE PROPOSED ISSUE OF RMB SHARES. These are HKEX items rather than EDGAR submissions; the PHIP (1st submission) and OC Announcement were rejected and not ingested. Informational only — no financial data from these has been incorporated.
  • ARAYNew EDGAR filings detected and ingested: 8-K (convertible notes repaid at maturity 6/1/2026), Q2 FY2026 10-Q, FY2025 10-K/A, and an amended 10-Q restating the RPO footnote. No operational or strategic update in the 8-K.

Last updated Jul 22, 2026

Balance-sheet distress

allreliabilityv1

Flag Stage 1 as unreliable when net leverage is high AND margins are thin or negative — the mechanical peer multiple overstates fair value for distressed names. Present a distressed range, not a single number.

max net leverage 4.5×min EBITDA margin 8%
rule-distress-all
Open flags raised (2)
  • ARAYNet leverage is high AND margins are thin/negative (27.7% gross margin, -12.2% net margin, -$49.2M TTM net income). The Stage 1 baseline is unreliable: a mechanical peer EV/EBITDA multiple overstates fair value for a distressed name. Report a distressed range, not a single point estimate.
  • 688329.SHHigh net leverage (~3.66x) combined with thin margins (3.8% EBITDA, 4.5% net) means the Stage 1 peer-multiple output overstates fair value — treat it as unreliable and present a distressed haircut range rather than a single point estimate. The fact set contains no peer EV/EBITDA multiple or share price, so the range can only be framed qualitatively, not numerically.

Last updated Jul 22, 2026

Tier 3 · Company-specific rules
ARAY1 rule

ARAY balance-sheet distress

RadTechARAYreliabilityv1

ARAY carries ~6.2x net leverage on thin, negative-net-income results. Treat the Stage 1 baseline as distress-flagged with a haircut range; the peer EV/EBITDA multiple is not a reliable point estimate here.

max net leverage 4×min EBITDA margin 10%
rule-distress-aray
Open flags raised (1)
  • ARAYARAY-specific distress condition confirmed: ~6.2x net leverage on thin, negative-net-income results. Treat the Stage 1 baseline as distress-flagged with a haircut range; the peer EV/EBITDA multiple is not a reliable point estimate. Fair value should be driven by covenant compliance, liquidity runway and refinancing terms rather than a multiple.

Last updated Jul 22, 2026

ISRG1 rule

ISRG durable compounder

MedTechISRGv1

ISRG compounds revenue ~20%/yr with a long robotic-surgery penetration runway. Use the full 10-year explicit horizon and a terminal growth at the GDP ceiling, rather than fading its growth to the terminal rate inside the generic window.

rule-compounder-isrg

Last updated Jul 30, 2026