Pope Capture — value-capture plan

The world after AI board · 8 theses · all 8 have a real capture path · 10 Opus agents
Bottom line
The single best opportunity is the AI-power grid-stability call (#1). It is the only thesis on the board that pairs a genuinely fired data-layer signal (large-power-transformer PPI at 10.6 sigma, FDR-survived, our own ground pack, not keyword noise) with two independent capture paths a solo shop can run this week: a paid per-name discrimination note to a power-infra hedge desk, and a self-capturable pair trade (long the un-priced reactive-support pure-play AMSC, underweight the priced-for-perfection megacaps GEV/ETN) that pays off if we are right even when no buyer closes. The edge is differentiated and dated: the crowd is priced on the megawatts story while NERC's Level 3 alert, ERCOT NOGRR282, and FERC's PJM co-location order (Jan-Feb 2026 filings) force the spend one layer down into compliance hardware that is not discounted per name. The one real gap is that the named fund PM is unknown, so finding that name is the first move; until then AMSC investor relations (ir@amsc.com is public) is a reachable secondary and the pair trade needs no counterparty at all. The runner-up for fastest cash is the AI-insurers/evidence-recorder call (#2): the most reachable named payer with the most acute pricing gap (Karthik Ramakrishnan at Armilla, pricing a novel exposure with zero loss history against a hard Aug 2 2026 EU AI Act Art.12 clock), but it has no instrument so being right only pays if someone signs.

Do this week

  1. Write and send the AI-power grid-stability one-pager (#1): constraint shift from megawatts to compliance, anchored on the 10.6-sigma transformer-PPI fire and the NERC/ERCOT/FERC catalyst calendar, to one power-infra fund desk; in the same hour find the named utilities/industrials PM via LinkedIn and the 13F holders of GEV/ETN/AMSC, since that name is the blocker.
  2. Send Karthik Ramakrishnan (CEO, Armilla AI) the AI-insurer evidence-recorder one-pager (#2) with the EU AI Act Art.12 mid-2027 call, a kill-criterion, and our Brier record attached, asking for 20 minutes with whoever owns reserving on the Vanguard AI product - this is the single most acute named buyer on the board and has the hardest clock (Aug 2 2026).
  3. Find the a16z partner who led the Periodic Labs $300M seed (check the Periodic announcement byline and the Wilson Sonsini deal page), then send the 'Where the rent actually lands in materials-AI' memo (#3) to that partner and a retargeted copy to Geoffrey von Maltzahn at Lila; separately, size and place the self-capturable long-ULS leg this week since it needs no counterparty.
  4. Cold-email Kevin Flyangolts at Aclid the OSTP-framework timing forecast (#4) framed as fundraising ammunition, and in parallel pull Telesis Bio's current CEO and IR contact from the latest 8-K on SEC EDGAR so the TBIO instrument is actionable.
  5. Send Christopher Wellise at Equinix the heat/water site-risk one-pager (#6) with a named Frankfurt or Helsinki offtake-node shortlist via the public Heat Export inbound channel - lowest-cost first move on the board and likely to get a fast reply that validates the screening product.

Ranked plans

#1 low effort
AI power bottleneck moved from megawatts to grid-stability compliance; sell the per-name read and take the long-AMSC / short-GEV-ETN pair ourselves.
Who: A power/utilities hedge desk (PM name unknown, find it) plus American Superconductor (AMSC) IR as the reachable secondary
How value is captured: Two parallel paths: paid per-name discrimination note + dated catalyst calendar to a fund desk ($5-15k initial, scaling to a $10-50k/quarter retainer), AND a self-capturable pair trade (long AMSC reactive-support pure-play, underweight GEV/ETN) that needs no counterparty.
First move: Write a one-page dated brief (constraint shift + NERC Level 3 / ERCOT NOGRR282 / FERC PJM order, anchored on our 10.6-sigma transformer-PPI fire) and send it to one power-infra fund desk; in parallel find the named utilities/industrials PM via LinkedIn + 13F holders of GEV/ETN/AMSC.
Medium-high prize: $10-50k/quarter retainer if a desk bites, plus uncapped upside on the pair trade we can put on alone. Odds: medium - the data-layer proof is real and the catalysts are hard-dated, but the fund-PM name is unknown so the first sale is cold.
#2 low effort
Liability moves to the evidence recorder; AI insurers are the reachable, hungry payer pricing a novel exposure with no loss history.
Who: Armilla AI (CEO Karthik Ramakrishnan), then Munich Re aiSure (Michael von Gablenz) and Testudo
How value is captured: Paid structural-intel retainer to an underwriting/strategy desk: 3-month pilot read at $25-75k feeding their reserving, scaling to $100-250k/yr. No instrument - all cash from the retainer.
First move: Send Karthik Ramakrishnan a one-page dated call this week on what counts as admissible tamper-evident evidence under EU AI Act Art.12 by mid-2027, with kill-criterion and our Brier record, and ask for 20 minutes with whoever owns reserving on the Vanguard AI product.
Largest pilot ticket on the board ($25-75k) with the most acute named buyer and a hard Aug 2 2026 clock. Odds: medium - insurer procurement is slow and there is no instrument, so being right only pays if they sign.
#3 medium effort
Rent in materials moves from discovery to qualification; sell the read to a16z/Lila and go long the listed TIC pure-play ourselves.
Who: a16z (Periodic Labs deal partner, name unknown) and Lila Sciences (Geoffrey von Maltzahn); UL Solutions (NYSE: ULS) as the self-capturable long
How value is captured: Two paths: a 4-week fixed-fee qualification-layer teardown to the discovery funds ($15-50k), AND a long-ULS position we take ourselves on qualification/reliability rent (Element/Temasek as unlisted confirmation).
First move: Write the one-page dependency-graph memo 'Where the rent actually lands in materials-AI' and send to the a16z partner who led the Periodic $300M seed; first sub-task is to find that partner's name via the Periodic announcement byline and Wilson Sonsini deal page.
$15-50k teardown plus a clean self-capturable long on ULS that compounds with no counterparty. Odds: medium - a16z is cold and hard to reach, but the trade is ours to take regardless and the consensus is mid-flip in our favor.
#4 low effort
Programmable-biology access becomes a licensed screening layer; sell market-sizing intel to Aclid and express it via the TBIO micro-cap.
Who: Aclid (founder Kevin Flyangolts) and Telesis Bio (NASDAQ: TBIO, current CEO/IR from latest 8-K)
How value is captured: Paid regulatory/market-timing memo to the screening vendor as fundraising ammunition ($10-25k pilot, then ~$5-10k/mo), plus a self-capturable TBIO equity position (lean short if the framework mandates onboard screening faster than Telesis can ship).
First move: Cold-email Kevin Flyangolts a one-page dated forecast on whether the overdue revised OSTP framework mandates verifiable AI-resilient screening + benchtop access controls as a hard funding condition, and by what date; in parallel pull Telesis Bio's current CEO + IR from the latest 8-K on SEC EDGAR.
$10-25k pilot from a fundraising-motivated buyer, plus a clean (if binary) TBIO instrument. Odds: medium - real chance the framework lands soft/voluntary (the disqualifier), which kills both the TAM and the trade.
#5 medium effort
Autonomous-lab interoperability standards are the bottleneck before autonomy; sell the 'which layer wins' map to the VC already publishing our thesis.
Who: IQ Capital (the partner who authored the self-driving-labs-as-infrastructure piece, name unknown); NIST (Howie Joress) as the free credibility anchor
How value is captured: Quarterly intelligence retainer feeding autonomous-lab diligence ($8-25k/quarter) plus one-off vendor competitive-standards briefings ($10-40k to Artificial or Bruker/SciY). No clean instrument (BRKR a weak proxy).
First move: Identify by name the IQ Capital partner who wrote 'self-driving labs = infrastructure, not just intelligence', then send a 200-word cold email that agrees and adds the standards-layer bottleneck (cite the NIST modular-lab project and the four standard classes) with a dated kill-criterion and a free first memo; in parallel email Howie Joress at NIST for a 30-minute roadmap read.
$8-25k/quarter from a VC pre-primed to agree with us, which raises conversion odds. Offset by a 3-5 year resolution that weakens the proof window. Odds: medium.
#6 low effort
AI campuses reprice on permitted heat rejection and water rights; sell a pre-build site-risk screen to the operators and funds eating $130B of blocked build-outs.
Who: Equinix (VP Sustainability Christopher Wellise, public Heat Export inbound channel) and Blackstone BXDC (DC asset-management MD, name unknown)
How value is captured: Paid 4-6 week per-metro screening pilot + recurring site-risk intel feed to operators and capital allocators; data-license fee from Paces. No clean instrument - pure advisory.
First move: Send Christopher Wellise a one-page cold email with a dated shortlist of 5-8 district-heat offtake nodes plus discharge-permit flags for one EnEfG-exposed metro (Frankfurt or Helsinki), offering a paid 4-week pilot; in parallel identify the named Blackstone BXDC DC asset-management lead.
Most reachable payer (public inbound channel, screaming $130B pain, cheapest first move) but no instrument and it competes with Paces and Big-4 advisory, capping the prize. Odds: medium-high on getting a reply, lower on a differentiated paid pilot.
#7 medium effort
Agent authority rails are a contested standards land-grab; sell the standards-and-liability call to VCs and incumbents.
Who: Skyfire (founder Amir Sarhangi) as credibility node; Felicis Ventures and Stripe's ACP team as the real payers
How value is captured: Per-deal diligence fee on agentic-commerce term sheets ($10-25k) or quarterly intel retainer ($15-40k); incumbent strategy-desk forecast ($25-75k). No instrument (V/MA too diversified).
First move: Write the one-page dated scorecard 'Which agent-authority standard becomes the network and who eats the liability' with hard probabilities and resolution dates, send it cold to Amir Sarhangi, and in parallel identify the ACP product lead on the agentic-commerce-protocol GitHub org and the agent-infra partner at Felicis.
Good ticket if a VC bites, but our data layer is fully blind here (the edge is pure judgment) and the fight could collapse into one obvious winner within 60 days, evaporating the edge. Odds: medium-low.
#8 medium effort
Human perturbation atlases are already priced; the live edge is the narrower dispersion call on which atlas owner pharma standardizes on.
Who: A biotech crossover fund PM (13F holders of RXRX/ILMN) plus AstraZeneca (Jim Weatherall); long ILMN / short RXRX pair as self-capture
How value is captured: Research subscription/retainer to crossover funds ($30-120k/yr) or per-report scoping fees to pharma ($25-75k), plus a self-capturable market-neutral pair (long ILMN, short RXRX).
First move: Write the one-page dispersion read 'Who wins the human perturbation atlas: Recursion vs Illumina BioInsight vs Arc', anchored on Recursion-Roche >$500M, the Jan 2026 Illumina launch, and our spatial-transcriptomics HHI ~1177 patent read, and send it to RXRX IR and Jim Weatherall's office.
Theme is already capitalized (Jan 2026 Illumina launch), so the remaining edge is a subtle dispersion call and the pair risks carrying directional beta rather than the standardization signal. Lowest differentiated edge on the board. Odds: medium-low.
Full detail (named people, exact asks, kill-criteria, checkpoints) is in the local capture run + the free Ultra scaffold (63 verification tasks). Reply and I'll pull any single plan into a ready-to-send one-pager.