Through-lines, tensions, and strategic implications from the 2026 conference program. Drawn from a topic-and-format survey of all 65 conference presentations and deep extraction of the 52 decks containing extractable slide content.
Each finding is supported by named-deck citations in the through-line that follows. Triangulated where speakers converged; flagged where they did not.
It surfaced in nineteen of sixty-five decks (29%), anchored five sponsored product theaters, and is no longer a question of whether but of how to govern. URAC has launched a Health Care AI Accreditation; eight state-level AI healthcare laws are now enacted or pending; on the OpenAI GDPval benchmark, AI moved from a 12% to a 47.6% win rate against industry experts on healthcare tasks in fifteen months.
43% of plan sponsors are now willing to accept substantially less in rebate guarantees in exchange for the UM controls they actually want (PSG, n=228). Employers Health and CVS Caremark show what aggressive exclusion delivers — Humira utilization dropped 88.5%, Stelara 90.2% — but therapeutic substitution leaks (44.7% of post-Stelara new starts shift to IL-23, not biosimilars).
IRA, OBBB, MFN, PBM reform, 340B reform, state AI laws, and DTP/AKS guidance combine into a coherent regime. MFP applies for as many as three years post-LOE; 15 major biopharma companies have voluntarily engaged in MFN pricing talks; ISPOR Europe 2025 modeling suggests aligning US WAC to the OECD basket floor would yield a ~68% average reduction in oncology.
US drug spend is modeled at ~$740B by 2030 with specialty driving 80% of growth (Carelon panel). Cell and gene therapy SP sales projected 20x growth from $1.5B (2025) to $31.2B (2030) — Cardinal Health. Cell therapy's center of gravity is migrating from oncology into autoimmune; multiple first-in-class entries hit FDA timelines in 2026.
Plans rate SPs only 5.9 of 10 on whether they differ meaningfully in patient care (PSG). Decks with quantified outcomes — Mayo's $1,620 per-Rx loss teardown, Shields/UMass's $2,649 PMPM medical cost reduction, NCODA's <1% MIP medication abandonment — are staking the differentiation case. AI-native operators are pricing toward the marginal-cost floor of fulfillment.
Each through-line draws on multiple anchor decks. Citations preserved for follow-up.
Nineteen of sixty-five decks engage with AI directly. Three sponsored product theaters center their pitches on AI deployment (CareTria × 100ms, Risa Labs, Doceree's co-pay.com). URAC's Health Care AI Accreditation moves the field from "does this work" to "is this auditable." Operationally: Mandolin/Amber cut time-to-fill from 57h to 18h; Risa Labs reports a 99.5% date-of-service authorization rate on $600M+ in Part B drugs; co-pay.com collapses prior-auth processing from 15 minutes to under 2; CVS Specialty Expedite reports 94,000 prescribers EHR-connected with 50%+ of PAs requiring no provider outreach.
Macro context (CareTria × 100ms): healthcare is the second-fastest-growing enterprise AI sector at 8x YoY; healthcare AI spend grew 20x in patient engagement, 10x in prior auth, 5x in payer ops year-over-year (Menlo Ventures); hyperscaler capex has reached $602B in 2026E. The accountability layer is forming faster than commonly recognized: eight state-level AI healthcare laws are enacted or pending (Barclay Damon / Prescryptive).
Implication. Operators treating AI as a 2027 question are competing against operators mid-deployment. Manufacturers whose access functions are not yet using agentic AI are competing against peers who have already redesigned cost-to-serve. The strategic question shifts from should we adopt to can our AI pass an accreditation, and how do we redeploy the staff capacity AI creates.
PSG's national survey of 228 plan sponsors shows the #1 stated goal remains managing specialty trend (43% rank it first), but the #1 stated challenge is now strategy for coverage of expanded indications — i.e., the pipeline is outrunning coverage decision-making. 43% are willing to accept substantially less in rebates to implement desired UM. 26% don't support PA relaxation for any drug class. Concerns concentrate in oncology, hemophilia, and MS.
Employers Health / CVS Caremark provide the cleanest exclusion proof points: Humira utilization down 88.5%, spend down 83.9%, after exclusion in a single quarter; Stelara utilization down 90.2%, spend down 90.7%. But therapeutic substitution leaks — 44.7% of post-Stelara new starts shift to IL-23 (Skyrizi, Tremfya); 26.1% to IL-17 — meaning biosimilar capture is partially offset by class jumping into still-branded mechanisms. One in eight patients switches back to Humira (Truveta).
MedImpact projects 80% biosimilar conversion by 2029 → 25% reduction in payer specialty spend ($17 PMPM, $68.70 → $51.80). Wholesaler economics: biosimilars are 7% of Big Three wholesaler revenue but 25% of total gross profits — a structural reason wholesalers haven't aggressively pushed direct-purchase models.
The conference's most quotable data point. PSG's specialty-generic pricing teardown — published in PSG's 2026 Trends in Specialty Drug Benefits Report (downloadable via the QR code on the deck): PBMs charge $1,000–$9,000 for generic dasatinib in PSG's Artemetrx book; Mark Cuban Cost Plus charges $259 — roughly a 35x markup. For generic teriflunomide, the spread is $16 (Mark Cuban) vs. up to $6,000 (PBM range).
Ten of sixty-five decks engage IRA directly. Roughly half the corpus engages with at least one of the seven concurrent pressures (IRA, OBBB, MFN, PBM reform, 340B reform, state AI laws, DTP/AKS guidance). Clarivate's "Swallowed by Reform" framed five interlocking levers; Deloitte's IRA panel emphasized 19-team mobilization and MFP applying three years post-LOE.
Three named MFN-style demonstrations are now public: GUARD (Part D brands), GLOBE (Part B buy-and-bill), GENEROUS (Medicaid international reference pricing — notable because it would require global net price reporting to CMS for the first time). 15 major biopharma companies have voluntarily engaged in MFN pricing talks; the Repatha withdrawal from Denmark has been linked in Scandinavian trade press to MFN exposure.
The contrarian read came from Klick Health: of the 15 IPAY 2027 negotiated drugs, only 4 have IRA rebates that outperform existing market discounts. Zero IPAY 2026 drugs outperformed market last cycle. The implication: headline negotiated prices may be less economically punishing on Part D than fear modeling assumed; Part B effectuation in IPAY 2028 is the harder shift.
The DTP angle: IntegriChain frames "the Great Repricing" — three diverging models (commercial, government, self-pay/DTP). TrumpRx now hosts 78 products, 10 specialty: Humira at $950 (-86%), Amjevita at $299 (-80%), ustekinumab biosimilars at $360–$575 (>90% discount). The February 2026 OIG DTC Bulletin establishes a potential AKS safe harbor for DTP cash programs serving government patients.
Walgreens' pipeline reference deck catalogues ~80 named investigational assets across 30+ manufacturers. Capital concentrates in three places: obesity (5 Phase III GLP-1/multi-receptor assets racing through 2026–27 — Lilly fielding three); oncology ($252B → $441B by 2029, with ADC, bispecific, and oral SERD growth); rare/orphan (projected 21% of pharma sales by 2032; orphan drugs were 72% of FDA approvals in 2024).
Charles River Associates documents a structural shift: rare diseases with more than two approved treatments grew from 16 in 2015 to 101 in 2025 — payer formulary management has fully arrived. Annual direct medical cost of rare disease ~$400B in the US — equal to cancer, heart failure, and Alzheimer's. Excessive price premiums (>20%) yield disadvantaged access in competitive classes.
Two structural shifts beyond raw concentration. Cell and gene therapies are migrating from ultra-rare into large prevalent indications — Vertex's zimislecel (T1D), apitegromab (SMA), Orca-T (heme malignancies), the first CRISPR cardiomyopathy program. A growing CAR-T cluster targets autoimmune disease (lupus, myositis, MG, stiff person syndrome) — cell therapy's center of gravity is shifting from oncology. Multiple first-in-class entries hit 2026: oral PCSK9 (enlicitide), oral orexin-2 agonist (oveporexton), psilocybin for MDD (COMP360), antisense for chronic Hep B (bepirovirsen), FVIIIa mimetic (denecimig).
PSG's plan-sponsor data is the leading indicator: SPs rated 5.9 of 10 on whether they differ meaningfully in patient care, with employers slightly more willing (6.1) and health plans less so (5.5). The decks that responded most aggressively divide into three positioning camps:
Geography. Klick Health: 290 specialty drugs operate in limited-distribution networks averaging just 5 pharmacies, with 28% restricted to a single pharmacy. Top three PBM-owned SPs hold ~65% market share at $140.9B in 2022 dispensing volume. Health-system-owned SPs are now >50% of accredited specialty pharmacy providers — but not of revenue. The middle of the market is hollowing out.
Five real divides where reasonable experts on stage took different positions. These are where executive judgment matters most.
Pro-DTP camp: Wheel, IntegriChain, Doceree, Cost Certainty/Turquoise, Bayvrio, Avalere. Contrarian voice: HealthDyne ("DTP doesn't work for utilization controls, high-touch/specialty therapy, ongoing monitoring, multiple stakeholders, adherence-critical conditions" — essentially the entire SP disease portfolio). HealthDyne's 2x2 framework places DTP only in the high-affordability-pressure / low-therapy-complexity quadrant. TrumpRx's ~1M weekly visits demonstrate demand; conversion is unknown.
Synthesis. DTP is a structural feature of the 2026 commercial model but not the answer for complex specialty therapies. For specialty-focused brands, a hybrid posture — DTP for low-complexity patient segments, high-touch SP for the rest — is the durable model.
Most operational AI decks frame AI as augmentation (Mandolin/Amber: automation freed staff for clinical interventions, which rose 10x). The replacement framing emerged from the CareTria × 100ms theater ("I would rather invest $1M in AI than hire 20 more reps") and Travis's "Will AI Revolutionize Pharmacy Before It Kills Us" (CUT vs. CREATE strategic choice; cites median p(doom) of 5–10% across 2,778 AI researchers, with 48% giving >10% probability to AI-driven catastrophe).
Synthesis. In operational practice, augmentation is the dominant 2026 deployment pattern. Competitively, AI-native operators are pricing toward marginal-cost floors that incumbents cannot match without similar redesign. The realistic strategic frame is Travis's CUT-vs-CREATE choice.
Existential framing: Clarivate (5 interlocking levers, ~68% modeled OECD-floor reduction in oncology), Deloitte (3 years post-LOE, 19-team mobilization), IntegriChain ("Great Repricing"). Contrarian: Klick Health ("only 4 of 15 IPAY 2027 MFPs outperformed existing market discounts; zero IPAY 2026 drugs outperformed market last cycle").
Synthesis. Both can be correct. Part D negotiation has so far yielded discounts that, while painful, do not represent revenue destruction. Part B transition (IPAY 2028) is the harder shift — and MFN (especially GLOBE for Part B and GENEROUS for Medicaid) is the wild card.
Provider-side: McDermott Will & Schulte's "Infusion in Flux" documents margin compression on IRA-negotiated Part B drugs. Mayo shows traditional cross-subsidies (commercial-to-government, drug margin to office visits) eroding under CGT economics. Payer-side: 35% of plans have site-of-care programs (PSG); CVS Infusion Care stakes out the alternate-site positioning. White bagging is being increasingly hemmed in by state laws on disclosure, consent, and provider right-to-refuse.
Synthesis. Unresolved, will resolve over years. Manufacturers planning Part B launches in 2026–2028 should model both maintained-current-site and forced-shift scenarios.
Most decks accept PBM reform is happening (FTC-ESI settlement, CAA of 2026, DOL fiduciary rule). The question is what survives. Klick frames vertical integration as a structural advantage that PBM-owned SPs reinforce through favorable internal reimbursement. PSG's specialty-generic pricing teardown is the rebuttal evidence. CenterWell / Mark Cuban / Humana / Risk Strategies panel: "Transparency is not the end goal, it is the entry fee."
Synthesis. Rebate-and-spread economics are eroding. The likely 2027–2028 model: outcomes-based contracting at scale (enabled by AI smart-contract infrastructure), low-WAC formulary preferences capturing the savings rebates used to capture, and continued vertical integration — with Mark Cuban Cost Plus, Amazon Pharmacy, and manufacturer-direct programs as the disintermediating market pressure.
A theme-prevalence read across the 65-deck program. Captures where the program directed attention.
Notable absence. Value-based contracting in the dedicated VBC sense barely surfaced as a primary frame. Touched in the Specialty Contracting Crystal Ball panel and the Alnylam rare-disease panel, but the honest conclusion is that VBC remains rhetorical: cited universally as the future, operationalized rarely, with the AI-smart-contract substrate as the enabling infrastructure that has not yet arrived at scale. Two further notable absences: deeper engagement on health equity, and any meaningful program engagement with global pricing strategy beyond MFN reactivity.
Citable specifics that reset internal mental models. Each carries an anchor deck for follow-up.
35× PBM markup
PBMs charge $1,000–$9,000 for generic dasatinib in PSG's Artemetrx book. Mark Cuban Cost Plus charges $259. From PSG's 2026 Trends in Specialty Drug Benefits Report.
PSG$23B / 80% never converts
Pharma annual co-pay and patient-support spend; 80% never converts to activated savings.
Doceree$400B annual
Direct medical cost of rare disease in the US — equal to cancer, heart failure, and Alzheimer's.
Charles River Associates41M of 96M
Abandoned prescriptions in 2024 had patient OOP under $10. Cost uncertainty, not cost itself, drives abandonment.
Turquoise / IQVIA2M weekly
ChatGPT health-insurance and pricing messages. Patients are self-researching upstream of the visit.
OpenAI / Turquoise10–20× overstated
ChatGPT estimated $2,000–$4,000 OOP for an example product where actual was $200.
Turquoise / Eli LillyRepatha → Denmark
Withdrawal from Denmark linked in Scandinavian trade press to MFN exposure — first observable cross-border launch decision attributable to MFN.
Clarivate19 of top 100
Drugs by 2022 net sales were orphan-only. Median net sales ~$2B.
Charles River Associates90% / 68%
Of rare-disease patients/caregivers feel personally responsible for pushing to receive tailored care; 68% in "communication limbo."
PANTHERx / Morning ConsultBimzelx +1,223%
PMPM growth year-over-year. New IL-17 entrants are reshaping autoimmune economics quarterly.
Employers Health / CVS−$1,620 / Rx
Mayo Clinic's pharmacy formulary leadership argued non-340B health system SPs face this per-Rx loss — what they characterized as the "Island of no 340B."
Mayo Clinic12% → 47.6%
AI-vs-expert win rate on real healthcare tasks (OpenAI GDPval), in 15 months.
100ms / OpenAI4 of 15
IPAY 2027 MFPs that actually outperform existing market discounts. IRA Part D may be less punishing than fear modeling assumed.
Klick Health / AAF~1M weekly
TrumpRx visits — strong demand, but conversion is unknown.
HealthDyne / Word & Brown290 / avg 5
Specialty drugs in limited-distribution networks averaging 5 pharmacies. 28% restricted to a single pharmacy.
Klick Health400% of CPI
Pharmacy cost trend in 2025, rising to 525% of CPI by 2027. Specialty ~75% of pipeline.
CenterWell / Humana / AONCMS sponsored fireside
Wednesday morning anchor: a CMS official appearing with ADVI Health. Slide artifact was 2 pages (title and close only); we have no record of substantive content. The signal is the programming choice, not what was said.
ADVI Health × CMS8 state AI laws
Enacted or pending state-level AI healthcare laws — California, Illinois, Maine, Nevada, Utah, plus pending New York.
Barclay Damon / PrescryptiveDrawn from Walgreens' pipeline reference deck (152 normalized rows) plus rare-disease and CGT cross-references.
One-paragraph read: Obesity is now visibly crowded with five Phase III GLP-1/multi-receptor assets (retatrutide, eloralintide, MariTide, CagriSema, survodutide); oncology remains the volume play ($252B → $441B by 2029); rare/orphan reaches 21% of pharmaceutical sales by 2032 with the PRV program reauthorized through 2029; cell and gene therapies are migrating from ultra-rare into prevalent indications (T1D, ATTR-CM, AMD, RP) with autoimmune emerging as the new center of gravity for CAR-T (lupus, myositis, MG, stiff person syndrome); multiple first-in-class entries hit 2026 (oral PCSK9, oral orexin-2 for narcolepsy, psilocybin for MDD, antisense for chronic Hep B, FVIIIa mimetic). Charles River Associates documents that excessive price premiums (>20%) yield disadvantaged access in competitive rare-disease classes.
Five product theaters across the program. Three of five centered on AI — reinforcing the AI through-line.
| Sponsor | Topic | Editorial signal |
|---|---|---|
| Doceree | DTP / patient affordability | Reframes the $23B co-pay category around therapy-start outcomes. |
| 100ms × CareTria | AI patient access operations | Two-hour interactive workshop — the longest single audience-contact investment of the program. |
| Risa Labs | AI prior authorization | 99.5% date-of-service authorization claim; $600M+ Part B drug volume processed. |
| Bayvrio | AI specialty marketplace | Category-creation pitch — "marketplace" architecture against bundled vendor stacks. |
| ADVI Health × CMS | Regulatory fireside | Editorially the most striking — CMS official on a sponsored stage. (We make no claim about session content; no record retained.) |
Five questions worth carrying into the next ninety days.
URAC's Health Care AI Accreditation makes this auditable. Mandolin/Amber and CVS Specialty Expedite show the operational floor. Is your strategy defined for a CFO and chief compliance officer, or only for an IT roadmap?
PSG's 43% (substantially less rebate for desired UM) is the leading indicator. MedImpact projects $17 PMPM savings under 80% biosimilar conversion by 2029. Does your launch pricing assume a rebate-defended formulary position?
IRA + OBBB + MFN + PBM reform + 340B reform + state AI laws + DTP/AKS guidance is the regime. Klick's "only 4 of 15 IPAY 2027 MFPs outperformed market" is a useful corrective to existential framing on Part D, but Part B IPAY 2028 effectuation and MFN-driven OECD-floor exposure are real.
Obesity approaches saturation by 2030; oncology remains the volume play; autoimmune is the new CGT frontier. CRA's data on competitive rare-disease classes (HAE, NMOSD, Gaucher) is the structural map for where payer management is intensifying.
PSG's 5.9 of 10 SP differentiation rating is the commodity-perception risk. CVS Specialty, Mayo, Shields/UMass, NCODA, and PANTHERx demonstrate what differentiated positioning looks like. AI-native operators are pricing toward marginal-cost floors. The risk crystallizes inside the 2026–2027 contract cycle.
Drawn from a topic-and-format survey of all 65 conference presentations and deep extraction of the 52 decks containing extractable slide content. Cross-deck triangulation supports the major findings; the contrarian voices in Part II are surfaced where reasonable experts on stage took different positions.
Limits. Spoken content from panel discussions is permanently lost from artifacts for ~25 decks. Cross-speaker disagreements within panels are not captured. Consensus framings here are built from the captured artifacts, not from total verbal content.