← Corpus / chroma-decks / exploration

Discrepancy Reconciliation — Founder Deck + Call Notes vs MemoPop

Side-by-side reconciliation of every fact-bearing claim where the founder's Series A deck (2026-05-11), the founder call notes (2026-04-29), live primary sources (2026-05-12), and the MemoPop-generated investment memo (ChromaDB v0.0.7, 2026-05-03) intersect or disagree. The call notes substantially close the open questions from v0.0.1.0; what remains is the disclosure-level decision on the Ask slide and a small handful of memo-only-but-keepable category-framing claims.

Path
explorations/Discrepancy-Reconciliation__Founder-Deck-vs-MemoPop.md
Authors
Michael Staton
Augmented with
Claude Code on Claude Opus 4.7 (1M context)
Tags
Reconciliation · Chroma · Enhanced-v1 · DD-Citation-Discipline · Fact-Check · Founder-vs-Memo · Call-Notes-Integrated

Discrepancy Reconciliation — Founder Deck + Call Notes vs MemoPop

Walks Proto slide by slide against the founder’s call notes (the new ground truth), MemoPop’s research, and live primary sources. As of v0.0.2.0, founder call notes from the 2026-04-29 conversation supersede most prior escalations. The MemoPop memo is now relegated to supporting category framing — it never sources a specific Chroma number that ships on Enhanced-v1.

What changed in v0.0.2.0

  • Founder call notes (2026-04-29) ingested as a new primary input. Notes live at ~/content-md/lossless/ChromaDB.md (outside the repo).
  • Five rows previously escalated-pending are now resolved from the call notes alone: round size + stage (5), lead investor (6), team headcount framing (7 — defer; founder OK with current 12-person slide), syndicate naming (12), use-of-funds split direction (founder anchors in ingestion-agent leadership + outbound enterprise engine).
  • Two rows materially clarified: the “Series B Oct 2025 $18M” claim in the MemoPop capital-efficiency section is confirmed hallucinated — there is no Series B. Funding history is: pre-seed (rolling notes, May 2022) → priced seed $18M at $75M premoney (April 2023, Quiet leads) → current Series A $12M at $120M post (Quiet leads again).
  • New material the deck doesn’t currently surface but the founder confirmed in conversation: Context-1 flagship model, enterprise ASP $300K/year, customer breadth (Paramount, Qualcomm, Slack, Notion, Krea AI, Nebius, CoreWeave), $4M cash + ARR, >9 months runway without raise, board composition (Jeff + Alex Kvame from Quiet).
  • Enhanced-v1 narrative restructured around these facts; Market slide retained as analyst framing because management will screen the deck before send.

Inputs

  • Founder call notes (NEW, primary): ~/content-md/lossless/ChromaDB.md — notes from a 2026-04-29 call with Jeff Huber. Includes financing history, current round details, customer list, ASP, runway, board, and product-cadence quotes.
  • Founder deck: corpus/management-supplied/2026-05-11_Chroma-Series-A_MS-Resort.pdf — 14 pages, captured 2026-05-11. The artifact Jeff presents from live.
  • Live primary sources checked 2026-05-12:
    • GitHub chroma-core/chroma: 27,915 stars · 2,241 forks · created 2022-10-05
    • PyPI chromadb: 13,780,462 monthly downloads · 3,710,469 weekly · 597,664 daily
    • npm chromadb: 765,383 monthly · 185,136 weekly · ~25K daily
  • MemoPop memo: corpus/memos/memopop-generated/ChromaDB-v0.0.7/ — generated 2026-05-03. Fact-check pass classifies 39 of its claims as unverifiable, 3 as contradicted, 5 as needing correction. Useful for directional category framing only.

Headline finding

The founder’s call notes are now the operative source. They confirm the deck’s numbers, supersede the memo on funding history, and add concrete material (ASP, cash position, runway, customer breadth, board composition, the Context-1 flagship) that the deck is silent on. Most of the work in this reconciliation is no longer “memo vs founder” — it’s “what of the new founder material should ship on Enhanced-v1.”

The MemoPop memo retreats to supporting research. It can still be cited (in data-source attributes) for the Market slide’s analyst-consensus framing, but it never sources a specific Chroma claim that ships.

Funding picture, fully resolved

Pre-seed (May 2022)Seed (April 2023)Series A (current, May 2026)
VehicleRolling convertible notesPriced round (notes converted)Priced round
Total at this stage(notes, undisclosed)$18M$12M target ($10M committed, $2M available)
Pre-moneyn/a$75M$108M (implied from $120M post − $12M)
Post-moneyn/a$93M$120M
% soldn/a19.3%10%
Lead(none — operator angels)Quiet Capital (Astasia Myers)Quiet Capital again (Astasia Myers) — second lead
BoardJeff Huber + Alex Kvame (Quiet)
Institutional participationQuiet, Bloomberg Beta, Air Street Capital, AIX VenturesQuiet $3M, Bloomberg Beta $250K, AIX Ventures $250K, insider syndicate $4.5M
Operator angels (named)Anthony Goldbloom (Kaggle), Nat Friedman (ex-GitHub CEO)+ Naval Ravikant, Max + Jack Altman, Guillermo Rauch (Vercel), Amjad Masad (Replit), Akshay Kothari (Notion), Spencer Kimball (CockroachDB), Jordan Tigani (MotherDuck)“all institutional insiders participating”

Cumulative funded post-Series-A: $30M. Modest by category. (Pinecone ~$138M / $750M valuation; Weaviate ~$68M / $200M+ valuation per the call-notes comparable table.)

The MemoPop memo’s “$36M total ($18M seed + $18M Series B Oct 2025)” framing is wrong. The $18M Series B was sourced from a single weak source (SalesTools AI) and never happened. The capital-efficiency section’s runway, gross-margin, and unit-economics inferences are downstream of that error and should not be cited.

Reconciliation table

Legend for Type: contradiction | unit-conflation | different-thing | sequential | gap | framing. Legend for Status: resolved | founder-confirmed-on-call | escalated-pending | screening-by-management (i.e. on the slide but flagged for Jeff’s review).

#SlideClaimFounderMemoLive (2026-05-12)TypeRecommendationStatus
103 TractionGitHub stars27k (call notes: 27.7K)12,000+27,915contradictionUse founder. Live confirms. Memo is ~18 months stale.resolved
203 TractionDownloads14M monthly15M+ weekly npmPyPI 13.78M/mo, npm 0.76M/mounit-conflationUse founder (“14M monthly downloads”). Combined PyPI+npm monthly ≈ 14.5M, matches. Memo’s “15M weekly npm” was a units error. Don’t add “PyPI” qualifier — that would be a new conflation.resolved
303 TractionCloud teams50K”500K+ MAU”n/adifferent-thingUse founder (50K cloud teams). Memo’s MAU is unverified and measures the OSS user base — different denominator, doesn’t contradict. Don’t add MAU.resolved
403 TractionCloud run rate$2.4M(silent)n/agapUse founder.resolved
515 AskRound size & stage”raising $12M” / “Series A""$18M seed (Apr ‘23) + $18M Series B (Oct ‘25)” → $36M totaln/asequential (memo also wrong about Series B)Use founder. The $18M Series B is hallucinated — drop entirely. Funding history is pre-seed → seed $18M @ $75M premoney → current Series A $12M @ $120M post. Quiet leads both priced rounds.resolved (call-notes confirmed)
615 AskLead investor (current round)(deck is silent)(silent for current)n/agapQuiet Capital again. Disclosure level on the slide is a separate decision (max disclosure / middle / conservative — see §“Open disclosure decisions” below).founder-confirmed-on-call
710 TeamHeadcount12 on slide”~25 employees” (unverified, capital-efficiency section)n/adifferent-thingUse founder. The 12 on the slide are the visible team / leadership the deck wants to surface; total HC isn’t load-bearing. No “12 of N” sublede unless founder asks for it.resolved
809 CompetitionPinecone share / framing(positioned bottom-left on matrix)“30% overall market share, managed-only”n/aframingUse founder positioning. Memo’s specific share number is unverified per fact-check. The matrix already encodes the strategic argument; no share number ships.resolved
909 CompetitionChroma market share(orange dot, top-right; no number)“25% OSS / 15% overall”n/aframingOmit any specific share number. Founder didn’t claim, memo unverified. The orange-dot positioning carries the leadership claim visually; that’s enough.resolved
1009 CompetitionCompetitor listOpensearch · Pinecone · Turbopuffer · Glean · “Labs” · In-housePinecone · Weaviate · Qdrant · Zilliz/Milvusn/aframing (different-thing — different question)Use founder. His list answers “where do we sit on the open/platform axis”; memo’s lists vector-DB peers. Both can be true, the founder’s framing is the strategic one.resolved
1104 Market (NEW)TAM / CAGR(silent)“$113B → $250B; ~45% CAGR; $28B SAM”n/agap (memo-only addition)Keep slide. Numbers ship as analyst-cited framing (cite IDC + Gartner directionally, no fabricated report URL). All numbers visible on the slide flagged in data-source-status as memo-borrowed-directionally. Management screens before send.screening-by-management
1211 Backed by (NEW)Capital syndicate(deck silent)partial list (capital-syndicate section)n/agapUse call notes (full + correct list). Lead callout = Quiet Capital (Astasia Myers, both rounds). Institutional grid: Bloomberg Beta, Air Street Capital, AIX Ventures. Operator-angels grid: Kimball, Tigani, Rauch, Masad, Kothari, Goldbloom (Kaggle), Naval, Max + Jack Altman. Pre-seed mention: Goldbloom + Nat Friedman.founder-confirmed-on-call
1307 Case StudyxAI quote / Grok integrationfull slide content(silent on xAI specifically)n/a (verifiable on Chroma blog)gapUse founder. Add an “Also powering” logo strip below from call notes: Paramount, Qualcomm, Slack, Notion, Krea, Nebius, CoreWeave.resolved
1412 RoadmapToken throughput target”400 → 10k tok/s”(silent)n/agapUse founder.resolved
1512 RoadmapSingle-index size target”250M → 100B”(silent)n/agapUse founder.resolved
1603 Traction”Authored Context Rot report”claim on slide(silent)verifiable on trychroma.com/bloggapUse founder.resolved
1702 Opening”Agents will automate knowledge work…“headline(no equivalent)n/aframingUse founder.resolved
1804 Bottleneck”Bottleneck is no longer reasoning failures, it’s context and memory failures”headline(memo frames as “AI-native search infrastructure”)n/aframingUse founder.resolved
19(omit)Gartner Visionary status(not asserted)“Gartner 2025 Vector DB Magic Quadrant Visionary”unverifiedaspirationalOmit. Memo asserts, fact-check tags unverifiable, founder didn’t claim it.resolved
20(omit)Forrester developer-preference(not asserted)“40% of AI developers prefer ChromaDB for prototyping”unverifiedaspirationalOmit. Same.resolved
2111 Business Model (NEW LINE)Enterprise ASP(deck silent)(memo silent)n/agap (call-notes-only addition)Add: “Enterprise ASP — $300K/year” as a small footnote on the Enterprise tier. Founder-confirmed on call.founder-confirmed-on-call
2213 Capital Efficiency (NEW SLIDE)Comparable funding(deck silent)(silent)publicly verifiable per call-notes tablegap (call-notes-derived addition)Add new slide: Pinecone $138M / $750M · Weaviate $68M / $200M+ · Chroma $30M post-A / $120M. Story: more brand, more adoption, less capital.founder-confirmed-on-call (table is from call notes; competitor numbers verifiable)
2312 Roadmap (REFRAMED)Product cadencethree-stream chart (Sync · Database · Agent)(silent)n/aframingReframe with founder’s words from call: “Frontier models — done. Search agent — done. Ingestion agent — next.” Same three streams; sharper headers.founder-confirmed-on-call
2407 Case Study (NEW STRIP)“Also powering” customer logos(deck shows 13 logos on slide 3)(memo names Replicate + LangChain only)n/agap (call-notes addition)Add small “Also powering” logo strip below xAI block: Paramount, Qualcomm, Slack, Notion, Krea, Nebius, CoreWeave. Crawl-fetch-ingest for the SVGs.founder-confirmed-on-call
25(DROPPED)Gross margin footnote on slide 11 (Business Model)(founder silent)“75–85% gross margins on managed Cloud”unverifiedgap (memo-only inference)Drop. No founder gross margin or unit economics provided. Even framed as “infra-software benchmark,” it implies Chroma has those margins.resolved
26(DROPPED)“24–36 months runway” line on slide 14 (Use of Funds)“9+ mo without raise; $4M cash + ARR""24–36 months runway”n/agap (memo-only inference)Drop the line. Use of Funds describes what the money does, not how long it lasts. Founder’s actual runway figure is on a different fact basis (pre-raise) and shouldn’t be extrapolated.resolved
27(DROPPED)“MongoDB / Elastic / Confluent precedent” footnote on slide 11(founder silent)implied by capital-efficiency-section comparablesn/aframing (memo-only)Drop. Category-history claim Chroma didn’t make; not strong enough to stand without founder validation.resolved
2815 Ask (NEW LINE)Fast facts: cash + AR + runway + board”9+ mo without raise; $4M cash + ARR” (call); founder confirms 2026-05-12: $4M cash + AR · 9–12 months runway without raise (current burn, no cuts) · Board: Jeff Huber + Alex Kvame (Quiet)(silent)n/agap (call-notes addition + founder confirmation)Add as small muted footnote on the Ask slide. Pro-VC convention: cash on hand + AR = runway without raise is the standard first question. Surfacing it preempts the question and signals raising-from-strength. The “current burn, no cuts” qualifier is load-bearing — pro VCs occasionally see inflated runway figures that assume hypothetical cuts; stating “no cuts” upfront signals the founder is being straight.founder-confirmed-on-call (with 2026-05-12 disclosure-level confirmation)

Open disclosure decisions (Ask slide only)

The Ask slide (#15) is the one place where founder discretion is most consequential. Three postures, ranked from most to least disclosure:

  1. Maximum disclosure“Series A · $12M at $120M post · Quiet Capital leading again (their second lead) · $10M committed by insiders · $2M available.” Best for a tightly-pre-qualified LP list; signals momentum and insider conviction.
  2. Middle“Series A · $12M · led by Quiet Capital · $10M committed by insiders.” Drops post-money (which anchors negotiations). Good default if the deck might leak or be forwarded.
  3. Conservative (current Proto)“raising $12M” with no further detail. Leaves the conversation entirely for the meeting.

Default recommendation: Middle. Names the lead (which strengthens the deck), shows insider commitment (which signals momentum), but doesn’t anchor valuation.

Founder picks; reflect in slide 15 on the next iteration.

Discipline notes for Enhanced-v1

The founder is the source. The call notes + the deck are the primary inputs. The MemoPop memo is now background research — it can be cited in data-source attributes for the Market slide’s analyst framing, but it does not source a specific Chroma claim that ships.

Citation discipline. Every fact-bearing line in Enhanced-v1 carries data-source and data-source-status attributes. Status values used:

  • founder-deck — appears in the founder’s PDF
  • founder-call-2026-04-29 — from the founder call notes
  • live-verified — verified against a primary source (GitHub API, PyPI, etc.) at data-source-checked
  • memo-borrowed-directionally — analyst framing only (only used on the Market slide, screened by management before send)
  • analyst-consensus — cited industry analyst (IDC/Gartner/Forrester) without claiming a specific report URL

Management screens before send. Per user instruction (2026-05-12), Enhanced-v1 in its current form goes to Jeff for review before any external distribution. The Market slide and any analyst-framed numbers are flagged for his veto.

Cross-references

  • Plan: [[../plans/Author-Enhanced-v1-from-MemoPop-Research]]
  • Enhanced-v1 narrative: [[../../../../context-v/narratives/ChromaDB_Deck-Outline__Enhanced-v1]]
  • Proto narrative: [[../../../../context-v/narratives/ChromaDB_Deck-Outline__Proto]]
  • Founder deck: corpus/management-supplied/2026-05-11_Chroma-Series-A_MS-Resort.pdf
  • Founder call notes (primary source as of v0.0.2.0): ~/content-md/lossless/ChromaDB.md (outside repo; user’s content vault)
  • Memo (consolidated): corpus/memos/memopop-generated/ChromaDB-v0.0.7/7-ChromaDB-v0.0.7.md
  • Memo (densest fact reference): corpus/memos/memopop-generated/ChromaDB-v0.0.7/8-one-pager-content.json
  • Memo (fact-check audit): corpus/memos/memopop-generated/ChromaDB-v0.0.7/4-fact-check-verified.md
  • DD-citation spec: [[../../../../context-v/specs/Dididecks-AI-DD-Ready-Citation-and-Source-Access]]