Felicis Ventures

Menlo Park-based multi-stage venture firm.

Snapshot

• Founded: 2006 by Aydin Senkut
• HQ: Menlo Park, CA
• AUM: $3B+
• Most recent fund: FV X — $900M, closed Jun 2025; the firm’s largest fund ever
• Stage: Multi-stage
• Stated thesis: “AI that speaks your industry’s language” + AI infra + foundation models. 70%+ of active portfolio is AI-native.
• Cap-table conflict status: Verified clean across all 8 Init Intelligence competitors as of 2026-05-05.

Cap-table & thesis fit

Felicis’s services-as-software thesis (“AI sells outcomes via labor”) is reflected in Mercor, Paraform, Assort Health, DoNotPay, n8n — a coherent “agents do the work” portfolio. No portfolio conflict with any of the 8 Init Intelligence competitors. The portfolio carries no AI-ITSM exposure specifically.

Lead deal partners

Partner	Thesis fit	Portfolio anchor
James Detweiler	Sourced/led/co-led 25+ companies; closest single partner to AI-services-as-software pattern	Mercor, Paraform, Assort Health
Aydin Senkut	Founder/MP; explicit framing on “AI + robotics + agentic workflows reshaping the economy”	SkildAI, n8n, Notion (legacy)
Viviana Faga	GP, GTM/enterprise — right partner for an enterprise-GTM-driven pitch	Runway, MotherDuck, Datology, DeepJudge, Supabase

(Note: Niki Pezeshki transitioned to venture partner Apr 2025 — no longer leading new deals.)

Relevant portfolio (services-as-software + AI agents)

• Mercor — Series B → Series C $350M — James Detweiler — AI-powered recruiting / labor marketplace
• Paraform — Series A $20M Jun 2025 — James Detweiler — AI-augmented recruiter marketplace
• Assort Health — Series B — James Detweiler — “specialty-specific agentic AI” for patient experience
• n8n — Series A — Aydin Senkut — “backbone for agentic AI automation”
• Letta — seed — agent framework / memory
• Browser Use — seed — agent infra (browser automation)
• Notion — seed (legacy) — Aydin Senkut — productivity SaaS being AI-rewired
• DoNotPay — seed — legal services-as-software
• Komodor — Series B — Kubernetes mgmt — IT-ops-adjacent infra
• Weights & Biases — Series A — ML/AI eval infra
• Runlayer — Andrew Berman’s new co. (ex-Zapier, Ravenna advisor); $11M Nov 2025 led by Felicis + Khosla/Rabois

Related

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

---

Fundraise intel (2026-05-25)

Appended for the investor-intel build. The content above (clean-lane analysis, 2026-05-05) is preserved verbatim; this block reconciles it with fresh primary-source research and the CRM. Felicis is now an active fundraise target, not just a clean-lane candidate — driver is the Philip Bogdanov → James Detweiler intro (5/13), which Sazzad is pursuing.

Firm Profile (refreshed)

Menlo Park multi-stage firm, founded 2006 by Aydin Senkut. ~$3B+AUM\ast \ast ;latestfund\ast \ast FVX=$900M (closed Jun 2025), the firm’s largest. 70%+ of the active portfolio is AI-native. Structure: partnership (not solo-GP); votes shares alongside founders; commits 1% on top of every first check for founders to invest in themselves; markets fast, conviction-led decisions (a speed advantage for founders). Invests across stages (seed → growth) but the AI franchise skews seed / Series A / Series B. Self-described as “among the top 8 VC firms” with 8 mega-unicorn ($5B+) backings at Seed/A/B in the last decade. Notable AI portfolio: Mercor, Skild AI, Browser Use, Supabase, Weights & Biases, poolside, Runway, Notion, n8n, DeepJudge.

Partners (deal-relevant)

• James Detweiler — General Partner, AI lead (promoted Feb 2026). The target partner. Thesis: “root node” AI — AI-enabled labor markets + robotics foundation models. Deals: Mercor (Series B lead), Skild AI (brought at Series A), DeepJudge (Series A lead), Predibase (Series A lead), Operand (seed), Paraform, Ricursive. → deep page: James Detweiler.
• Aydin Senkut — Founder / Managing Partner; agentic-economy framing.
• Sundeep Peechu — Managing Partner; co-led Mercor Series B with Detweiler.
• Viviana Faga — GP, enterprise/GTM; co-led DeepJudge Series A with Detweiler; also Runway, MotherDuck, Supabase.
• Peter Deng — GP (ex-OpenAI/Meta/Uber product); AI product/consumer.
• (Niki Pezeshki → venture partner Apr 2025, no longer leading new deals — per existing content above.)

Recent deal sketch (Pass-1; full rows in CSV)

• Mercor — Series B $100M@$2B, Feb 2025 — Felicis LED (Detweiler + Peechu); largest check ($50M) of the prior fund. [Existing investors Benchmark/GC/DST followed.]
• DeepJudge — Series A ~$41.2M, Nov 2025 — Felicis LED (Detweiler + Faga), w/ Coatue.
• Skild AI — Series A $300M@$1.5B, Jul 2024 — Felicis participated (Detweiler sourced); lead = Lightspeed/Coatue/SoftBank/Bezos. Now $14B (Jan 2026).
• Operand — Seed, May 2025 — Felicis w/ YC + SV Angel — consulting-as-software.
• Predibase — Series A, May 2023 — Felicis LED (Detweiler + Pezeshki). Full table → /Users/sazzad14/seed-market-scan/firms/felicis.csv.

Best data channels for this firm

Felicis’s own blog (felicis.com/blog/<co> and /insight/<co>) is the single best source — every deal has a partner-bylined investment memo that names the deal partner(s) explicitly (rare and gold for partner attribution). Cross-check amount/valuation with TechCrunch / Fortune / Bloomberg / Law.com. The Felicis team page + LinkedIn confirm titles. NFX Signal (signal.nfx.com/investors/james-detweiler) for per-partner lead history. The Venture Codex spine understates Felicis’s role on co-led rounds (per playbook §5) — trust the firm’s own memo for led/co-led/participated.

Fundraise Relationship (init.inc)

• Status: Active fundraise target. Detweiler owns the AI-labor lane (Mercor, Operand, Paraform) AND wrote “The Agentic Web,” whose first adoption wave names SRE + support agents — init.inc’s IT/IT-ops wedge. Check-size: Detweiler’s ~$7.8Msweetspo{t}^{[}single-source:CRM]\approx init.in{c}^{\prime }s$7M target. Felicis leads and decides fast.
• Conversation log (newest first):
  • 2026-05-13 — Philip Bogdanov intro’d James Detweiler. Sazzad pursuing this path. (CRM)
• Mutual connections / warm path: Philip Bogdanov → James Detweiler (live edge, intro made). Secondary: Mercor (Peechu) / DeepJudge (Faga) give a multi-partner IC audience.
• Personal & rapport notes (firm): founder-first, conviction culture; “show up early with conviction” is how Detweiler describes earning founder trust. Fast IC; 1%-on-top founder-investment quirk.
• Live stage: tracked in the CRM (Google Sheet).

Cross-check vs existing wiki (reconciliation)

• ✅ Thesis read agrees: the clean-lane analysis above already names Detweiler as “the right partner” and Felicis’s services-as-software thesis as the strongest fit — fresh research confirms this and adds the partner-level detail.
• ✅ Fund facts agree: FV X $900M/Jun2025,$3B+ AUM, 70%+ AI-native — all corroborated.
• ⚠️ Skild AI partner attribution: the table above attributes Skild AI to Aydin Senkut. Felicis’s own promotion post + Detweiler’s bio say Detweiler brought Skild AI to Felicis at the Series A on his early robotics-FM conviction. → Detweiler is the deal partner of record for Skild AI (Senkut, as founder/MP, may have co-sponsored). Existing table left unedited (out of scope); flag stands. (detail)
• ⚠️ Mercor round label: existing text says “Series B → Series C $350M."Primarysources:\ast \ast FelicisLEDthe$100M Series B at $2B(Feb2025)\ast \ast ;alaterSeriesCwaspursuedat$10B. The “$350M” figure isn’t corroborated as the Felicis-led round — likely a later/other round. Minor; verify in Pass 2.
• 🟢 Cap-table conflict status unchanged: “verified clean across all 8 competitors (2026-05-05)” — no new conflict surfaced; Felicis remains conflict-free for init.inc.

---

Pass 2 deal pattern (2026-05-25)

Exhaustive ~18-24mo (Nov 2024 → May 2026) pre-seed/seed pull, triangulated ≥2 sources/row (Felicis blog memos + press + firm bylines). Full 23-row table → /Users/sazzad14/seed-market-scan/firms/felicis.csv. This block is additive; nothing above is edited.

Seed lead-rate & deal counts (in-window)

Felicis is a prolific, conviction-led seed lead. In-window seed/pre-seed rounds = 11 (Entire, Sooth, Browser Use, Operand, Fifth Door, Runlayer, Observo AI, Extra/BuildForever, humans&, Periodic Labs, Smack). Of these, Felicis LED or CO-LED 8 of 11 (~73%) — Led: Entire, Sooth, Browser Use, Operand, Extra; Co-led: Fifth Door, Runlayer, Observo AI. Participated-only on the three mega-rounds where it couldn’t be lead (humans& $480M,PeriodicLabs$300M, Smack $32M combined). This matches Felicis’s own public claim of “led or co-led 83% of bets” (Crunchbase, Haskaraman profile) and “93% of the last fund into seed/Series A.” Counting Series A/B too, 22 in-window Felicis rounds were enumerated.

Check / stage profile

• Classic (non-frontier-lab) seeds cluster $9.5M–$20M (Extra $9.5M,Runlayer$11M, Observo $15M,BrowserUse$17M, Fifth Door $20M)—medianclassicseed\approx \ast \ast$15M**. Operand is the small outlier ($3.1M).
• Frontier-lab / AI-talent seeds are a separate universe ($50MSooth,$60M Entire, $300MPeriodicLabs,$480M humans&). The $60MEntireand$50M Sooth are Felicis-LED mega-seeds — Felicis writes the lead check even at nine figures.
• Felicis leads Series A too (Paraform $20M,DeepJudge$41.2M, Artemis $70M-combined-A,Mercor$100M Series B). Sweet spot for a lead seed/A check spans roughly $10M–$60M.

Cadence

Very steady — roughly one announced seed every ~3-6 weeks across the window, accelerating into 2026 (Entire Feb, Smack Mar, Extra/Artemis/Sooth Apr 2026 within weeks of each other). FV X ($900M, Jun 2025) is actively deploying; pace is high.

Sector mix

AI-native across the board (consistent with “70%+ AI-native”). Concentrations: AI-labor / services-as-software (Operand, Mercor, Paraform — init.inc’s closest analogs); agent infra (Browser Use, Runlayer, Letta); security / IT-ops-adjacent (Observo AI, Artemis, Runlayer, Smack); dev tools (Entire); frontier science labs (Periodic Labs, Sooth, Ricursive, Smack); consumer AI (Extra, Fifth Door). The security + observability + agent-governance cluster (Observo, Artemis, Runlayer) is the nearest neighborhood to init.inc’s IT-ops wedge.

Partner attribution (from Felicis memo bylines — gold standard)

• Aydin Senkut (Founder/MP): Entire (lead), Sooth (w/ Haskaraman), Fifth Door (co-lead), humans&, Fractile, Assort Health (w/ Myers).
• James Detweiler (GP, AI lead): Extra (w/ Peter Deng), Ricursive (w/ Delcambre), Mercor (w/ Peechu), Paraform, DeepJudge (w/ Faga), Operand (sourced), Skild (brought). Still the init.inc target partner. ⚠️ refinement vs Pass-1: Periodic Labs byline is Peter Deng, not Detweiler — Detweiler’s lane is AI-labor/robotics-FM/EDA, less the science labs.
• Peter Deng (GP, ex-OpenAI): Extra (w/ Detweiler), Periodic Labs. Consumer + frontier-lab AI.
• Feyza Haskaraman (Partner, NEW — ex-Menlo; AI infra/cybersecurity/applications): Sooth (w/ Senkut). Worth tracking as a second relevant door (cybersecurity/infra remit overlaps init.inc).
• Jake Storm + Nancy Wang (investing team): Runlayer, Observo AI, Artemis, Smack — i.e. the security/agent-infra cluster is run by Storm/Wang, not Detweiler. Relevant if init.inc’s wedge reads as “security/IT-ops” rather than “AI labor.”
• Viviana Faga (GP, GTM): co-led DeepJudge. Astasia Myers (GP): co-bylined Assort Health. Sundeep Peechu (MP): co-led Mercor.

Notable seeds (init.inc-relevant)

• Operand ($3.1M seed, Felicis LED, May 2025) — consulting-as-software / “managed outcomes”: the single closest thesis twin to init.inc. Detweiler-sourced.
• Runlayer ($11M co-led, Nov 2025) — MCP/agent-governance security; founder ex-Zapier; IT-ops-adjacent.
• Observo AI ($15M co-led, Jan 2025) — agentic data-pipeline/observability; closest to the IT-ops buyer.
• Browser Use ($17M Felicis LED, Mar 2025) — agent web-automation infra.

Best data channels (refined)

Felicis’s own blog memos remain the single best source — every deal carries a partner-bylined memo that names the deal partner(s) explicitly (this is how the attribution above was built; rare and high-signal). Pair with TechCrunch / Bloomberg / BusinessWire / finsmes for amount + valuation + co-investors. The blog index is JS-paginated (only ~latest few render) → use Exa neural search --include-domains felicis.com with varied query phrasings to enumerate the full post set. ⚠️ Felicis frames participations as “our seed investment in X” even when it didn’t lead (humans&, Periodic Labs) — cross-check lead vs participate against press, never trust the memo’s framing alone for role.

⚠️ Pass-2 corrections to prior content (existing text left unedited)

• Operand: Pass-1 CSV said “participated/amount unknown.” Primary sources → Felicis LED, $3.1M (May 2025). Corrected in CSV.
• Assort Health: existing table & fundraise block imply a Felicis/Detweiler-led Series B. Primary sources → Series B $76M LED BY LIGHTSPEED; Felicis participated (Sep 2025); memo byline = Senkut + Myers.
• **Mercor “$350M"\ast \ast :thepriorflagstands—theFelicis-ledroundisthe\ast \ast$100M Series B @ $2B(Feb2025)\ast \ast ;the$350M/~$10B figure is a later round, not Felicis-led.
• Skild AI partner-of-record = Detweiler (confirmed again via Felicis promotion post), not Senkut.

---

Pass 3 — Pre-seed/seed economics + Viviana Faga’s seed activity (2026-06-02)

Additive; nothing above edited. Focused pull answering “how much does Felicis (and Viviana Faga specifically) invest at pre-seed/seed.” Double-verified (Felicis primary — About/Fund-X pages + memos — cross-checked vs Fortune/TechCrunch/Crunchbase News/finsmes). Driver: Viviana Faga intro in progress via Ryan Carlson (5/31).

Firm-level stage & check economics

• Stages listed: Pre-Seed, Seed, Series A, Series B, Growth Equity. ⚠️ But of the verified 2024–26 deals, none are pre-seed — observed activity is seed and up. “Pre-seed” is a stated capability, not a demonstrated recent pattern; don’t assume a pre-seed motion.
• Two distinct (don’t merge) stage stats:
  • 94% of investments — by count — are Seed or Series A (Felicis “About” page, 2026; corroborated by “in fund nine, 94%…”).
  • >93% of FV X capital went into seed & Series A rounds Felicis led or co-led (FV X announcement). Prior fund (FV IX): 87% of deployed capital into led/co-led rounds.
• Check range (firm-level): Felicis publishes no standard check. The real spread, per Fortune at the FV X close: largest-ever check = $50M(Mercor);smallest=$100K. Third-party aggregators put the typical check around $1M–$40M (superscout, vcsheet) — rough; the firm sizes to the round and leads/co-leads.
• AUM: $4.1B\ast \ast perFelicis"About"(2026,post-FVX)—anincreasefromthe\ast \ast$3B+ recorded 2026-05-05 above (self-reported growth, not a correction). FV X = $900M (closed Jun 2025), the firm’s largest fund and actively deploying.
• 1% Founder Development Pledge — non-dilutive capital on top of every first check (coaching/therapy) — confirmed again.

Viviana Faga’s seed activity (the “how much does she invest” question)

• Per-partner checks are not disclosed. Aggregator estimate of her entry check: $500K–$15M, Seed→Series A ^[single-source: Evalyze].
• Two Felicis-LED seeds carry her byline — these are her clearest seed-lead data points (full round table on her page):
  • Knox — $6.5M seed, Felicis LED (Faga + Nancy Wang), Jul 2025. FedRAMP-as-a-service / federal managed cloud (ex-Adobe federal cloud operator). (TechCrunch + Felicis memo + thesaasnews — strongly verified.)
  • Dreambase — $3.7M seed, Felicis LED (Faga byline), Apr 2026. AI-native analytics for the Supabase/Postgres ecosystem; Supabase CFO/CTO/COO co-invested. (Crunchbase News + finsmes + Felicis memo — strongly verified.)
• Across stages she (co-)leads $3.7Mseeds\to$52.5M Series B (MotherDuck) in data / cloud-infra / enterprise-SaaS / AI. Note both seeds are angel/multi-investor rounds, so Felicis’s slice is below the round total.

⚠️ Correction to the Pass-2 seed enumeration

• Pass-2 (above) counted 11 in-window seed/pre-seed rounds, Felicis led/co-led 8 of 11 (~73%). That set missed at least two Felicis-LED seeds inside the same window — Knox ($6.5M,Jul2025)andDreambase($3.7M, Apr 2026), both Viviana-led. So the in-window seed count and lead-rate are undercounts (the firm’s own claim is “led or co-led 83% of bets”). Not recomputed to a new precise % — the rest of the 11-row enumeration wasn’t re-audited, so a new fraction would falsely imply completeness. Both missed rows added to felicis.csv.