Spark Capital

Multi-stage venture firm.

Snapshot

• Founded: 2005
• HQ: SF, NY, Boston
• Stage: Multi-stage
• Cap-table conflict status: Verified clean across all 8 Init Intelligence competitors as of 2026-05-05.

Why this is a clean lane

The only mention of Spark across all 8 competitor entity pages is in Alex Clayton’s career history — Clayton was at Spark Capital Growth before joining Meritech. Clayton is now a Meritech GP and a personal-angel investor in Serval’s Series A. That’s a person’s career history, not Spark on any cap table.

Lead deal partners

Partner	Profile	Recent thesis-aligned bets
Arpan Shah (GP, Early; joined Nov 2024)	Newest Spark GP. Joined from Pear VC + Sequoia Scout; ex-Robinhood founding team + sold his startup Flannel to Plaid. Spark co-led Series A in Conductor (“multi-agent interface for sending teams of AI agents out to code”). Carries Series A check authority.	Conductor (multi-agent coding infra)
Nabeel Hyatt (GP, Early)	Backs only 1-2 companies a year. Anti-thesis investor — backs non-consensus product bets. Anthropic, Discord, Oculus pattern. Personally uses Claude Code, Granola, AI agents for own workflow.	Anthropic; “Open World AI” interface thesis
Natalie Vais (GP, Early)	Cloud infra/AI/dev tools. Portfolio: Polar Signals, MotherDuck, TigerBeetle, Definite (AI-native data stack with built-in AI agents). Definite markets “AI agents that bring insights and actions to customers.”	Definite, Polar Signals, MotherDuck, TigerBeetle, ElectricSQL

Other partners: Yasmin Razavi — her own framing is “Series B+, after PMF”; Anthropic board director. Andrew Parker. Will Reed — Series B/C focus.

Soft-conflict flag

Clayton left Spark for Meritech years ago — alumni-tier proximity only. (Clayton is now a Meritech GP and a personal-angel investor in Serval’s Series A — career history, not a Spark check.)

Related

• cap-table-patterns-across-startup-competitors
• Init Intelligence

Fundraise intel (2026-05-25)

Status in CRM: LOST — soft pass / keep warm for Series A (pass email 2026-05-07). Dead-tier this round; warmest of the dead tier — explicit “keen to revisit at next round.”

Firm profile (raise lens)

Multi-stage SF / NY / Boston firm, founded 2005; 600 Montgomery St SF office. Leads at every stage — recent: Mastra $22MSeriesAlead(Apr2026,"helpeverydeveloperbuildagents"),\ast \ast Latent\ast \ast$80M A co-lead (Apr 2026), Anrok $55M Series C lead (Oct 2025). Newest GP Arpan Shah (Nov 2024) carries Series A authority + the Conductor multi-agent precedent.

Why this relationship is dead/lost

Soft group-level pass, not a partner no. Arpan Shah (GP) was intro’d by Charlie Dean 2026-05-04; met Wed 2026-05-06 and “seemed decently sold,” took it to Spark’s group/partner meeting. Pass email Thu 2026-05-07 8:18 AM: “no high conviction we’d collectively reach a yes for this round — stepping aside. Keen to revisit at the next round.” So: the partner liked it; the partnership didn’t reach conviction at this round size/stage. This is a “too early for our collective conviction” pass with an explicit Series-A re-open invitation — the best re-touch posture in the dead tier.

POC (CRM contact)

• Arpan Shah (General Partner; joined Nov 2024 from Pear VC + Sequoia Scout; ex-Robinhood founding team, sold Flannel to Plaid). He championed the deal internally once; warm path Charlie Dean is still live. Single named POC; light stub not warranted (already covered in the lead-partner table above).

Fundraise Relationship (light)

• Status: This was a stage/conviction pass, not a thesis or vision no. Yasmin Razavi is an Anthropic board director.
• Conversation log (newest first):
  • 2026-05-07 08:18 — Pass email (Arpan): “no high conviction we’d collectively reach a yes for this round — stepping aside. Keen to revisit at the next round.” → soft pass / keep warm.
  • 2026-05-06 — Met Arpan (Wed). He “seemed decently sold,” took it to Spark group meeting.
  • 2026-05-04 — Charlie Dean intro’d Sazzad to Arpan.
• Mutual connections / warm path: Charlie Dean (the introducer — still warm, the re-touch channel).
• Competitor-conflict flag: None institutionally (verified clean across all 8 competitors above). Only adjacency: Alex Clayton (now Meritech GP + Serval angel) was at Spark Capital Growth years ago — career history, not a Spark check. If re-engaged, keep the deal partner clear of any tight Clayton relationship.

Best data channels for this firm

Founder/portfolio announcements (mastra.ai/blog, founder X posts like @calcsam — most reliable, name Spark as lead) → Crunchbase News (news.crunchbase.com “Spark Capital May Be Igniting Again” tracks their AI re-acceleration) → PRNewswire for formal closes. Aggregators (LeadIQ) are useful for breadth but several Spark “seeds” there are aggregator-only and need primary confirmation (West Co flagged single-source).

Cross-check vs existing wiki

Consistent with the existing clean-lane + Arpan-Shah-as-spear-tip content above — the CRM confirms Arpan was exactly the right partner (matched the page’s pre-meeting recommendation) and that he championed it internally. ⚠️ Reconcile: the page’s “Wave-3 outreach” plan is now executed and softly passed — the live state is LOST/keep-warm-for-A, not pending Wave-3. No factual contradiction; the strategy section is now historical/realized.

Pass 2 deal pattern (2026-05-27)

Seed lead-rate / check / stage. Spark is leading in-window at Series A for vertical-AI applications (Mastra $22MA—agentframework,Apr2026\ast \ast ;\ast \ast LatentHealth$80M A co-lead with Transformation Capital — clinical AI, Apr 2026) and at mega-seed scale for AI infrastructure (RadixArk $100Mseedco-leadwithAccel—SGLang/inferenceinfra,May2026\ast \ast ,ArpanShahquoted).Smallerparticipation:\ast \ast Poke$10M seed extension (Apr 2026, joint with General Catalyst — consumer AI agents via SMS/iMessage). Median in-window classic seed/A check $22M-$80M; outliers ($100M RadixArk) are open-source-infra-driven. Pass 1 had 4 rows — 3 re-verified ✅ (Mastra, Latent upgraded ⚠️→✅, Poke upgraded ❓→✅), Anrok carries over ⚠️, West Co remains ❓ (couldn’t get cross-engine confirmation, recommend drop). 1 net-new (RadixArk).

Competitor exposure (verified clean). Pass 2 surfaced no new institutional conflict — Spark remains clean across all 8 init.inc competitors. Closest adjacency is still Alex Clayton (Meritech GP, Serval angel) as pre-Meritech Spark Capital Growth alumnus — career history only. Cross-roster co-invest signals: (1) Spark + Accel co-led RadixArk $100M (May 2026); both are init.inc firms (Accel was LOST same week as Spark). (2) Spark + General Catalyst + Founders Fund angels (Ken Howery) in Poke — all three are init.inc fundraise firms in the same cap table. Neither creates a conflict, but tracks partner-network density.

Best data channel for this firm. Founder-side announcements (mastra.ai/blog, latenthealth.com/press, RadixArk’s BusinessWire) remain best — they consistently name Spark as lead/co-lead. Arpan Shah’s own quotes and the 1947 Rise podcast (May 20, 2026) interview are the richest first-party signal of his current thesis: “fewer investments, longer-term partnership, more Series A,” and an Anthropic/Granola-style preference for “AI as enabling layer made invisible, not front-and-center.” Crunchbase News tracks Spark’s AI re-acceleration. LeadIQ aggregator (used in P1) is noisy — West Co and Anrok rows from there couldn’t be primary-confirmed in P2.

P2 reconciliation with the LOST relationship. Arpan Shah passed init.inc on 2026-05-07 (“no high conviction we’d collectively reach a yes for this round”). Three weeks earlier (Apr 9, Mastra) and four weeks later (May 5, RadixArk) he was actively writing/championing $22Mand$100M checks. The init.inc pass wasn’t his thesis (he championed it internally) and wasn’t conviction in agents broadly — Spark group-conviction filter appears biased toward open-source/infra (RadixArk) and dev-framework horizontal (Mastra) over vertical enterprise-application (init.inc’s wedge).