SAASPOCALYPSEverdict #ZOOM-D2D3
scanned 2026.05.04 · 14:33
subject of investigation

zoom.us

unified communications & video conferencing platform
verdictCONTESTED
wedge score
45
/100
wedge thesis

the door is vertical depth: Zoom is a horizontal platform with shallow features everywhere, and any single vertical (telehealth, education, dev-team standups) can be out-featured by a focused niche product at a fraction of the price.

real walls — pick your flank·ship in 8 weeks·run for $47.00 + usage
the doordata
wedge

where the walls are.

methodology →
the door

no proprietary corpus — they're running on off-the-shelf data.

watch out

their capital wall is real — ongoing capex puts a floor under any clone.

capital
7.0/10
investment the incumbent had to make
why this scorehigh confidenceZoom's real capital moat is enterprise compliance infrastructure — HIPAA BAAs, SOC 2 Type II, FERPA, FedRAMP — plus...

Zoom's real capital moat is enterprise compliance infrastructure — HIPAA BAAs, SOC 2 Type II, FERPA, FedRAMP — plus the legal/audit teams, dedicated enterprise implementation staff, and global data-center/SFU infrastructure that underpins SLAs. The report explicitly flags compliance posture as a 'nightmare' requiring months and legal spend. Hardware endpoints (Zoom Rooms), carrier integrations (Zoom Phone), and payments risk from large enterprise contracts add further non-software spend. An indie builder can replicate the software layer cheaply but cannot replicate the compliance org or the trust infrastructure quickly.

  • Report flags HIPAA, SOC 2, FERPA as 'nightmare' tier — months of legal spend to establish
  • Zoom Rooms hardware ecosystem and Zoom Phone carrier integrations represent non-software capital commitments
  • Enterprise procurement inertia cited as a primary moat — implies large dedicated sales/implementation teams
technical
5.0/10
depth of the underlying engineering
why this scorehigh confidenceThe report is explicit that the hard technical problem — reliable multi-party video at scale (SFU, adaptive bitrate,...

The report is explicit that the hard technical problem — reliable multi-party video at scale (SFU, adaptive bitrate, packet loss recovery) — is already solved by Daily.co and Livekit, which an indie builder can drop in. The AI Companion features are described as 'one-prompt wrappers around Whisper and GPT-4.' The remaining medium-difficulty work (AI notes pipeline, recording storage) is real but tractable. Zoom's technical depth is real at the platform level but is largely commoditized by WebRTC-as-a-service providers, leaving only the integration breadth (phone, hardware, calendar, SSO) as genuine engineering depth.

  • Report explicitly states SFU/adaptive bitrate complexity is 'why you use Daily or Livekit instead of rolling your own' — commodity abstraction available
  • AI Companion described as 'one-prompt wrappers around Whisper and GPT-4' — no proprietary AI depth
  • AI meeting notes pipeline rated 'medium' difficulty, not hard
network
5.0/10
users compound users
why this scoremedium confidenceZoom has a meaningful but asymmetric network effect: the 'join a Zoom' link is ubiquitous and guests don't need...

Zoom has a meaningful but asymmetric network effect: the 'join a Zoom' link is ubiquitous and guests don't need accounts, which creates soft viral pull. The app marketplace (Zoom Apps) and ISV partner ecosystem add some multi-sided value. However, the core video call is not a true marketplace or social graph — participants don't accumulate value by being on Zoom vs. a competitor. The network effect is primarily brand/habit-driven rather than structural liquidity. A vertical niche product can bootstrap its own network within a single organization or community without needing Zoom's breadth.

  • 'Zoom' is a verb — brand recognition creates soft network pull for external meeting links
  • Zoom App Marketplace provides ISV/partner ecosystem lock-in for enterprise buyers
  • No true marketplace liquidity or social graph — calls are ephemeral, not accumulating network value
switching
6.0/10
stickiness of customer data + workflow
why this scorehigh confidenceEnterprise switching costs are real and multi-layered: SSO/SAML configurations, Zoom Phone number porting, Zoom Rooms...

Enterprise switching costs are real and multi-layered: SSO/SAML configurations, Zoom Phone number porting, Zoom Rooms hardware investments, calendar/CRM integrations, recorded meeting archives, and IT-approved procurement processes all create friction. For SMB/prosumer users, switching is easy. The report's own framing — 'enterprise procurement inertia' as the primary moat — confirms that switching pain is concentrated in the enterprise segment where Zoom earns its highest ARPU. A vertical niche attacker targeting SMB or a single department faces low switching costs from that segment.

  • Enterprise procurement inertia explicitly cited as primary moat in the report
  • Zoom Phone number porting and Zoom Rooms hardware create hard switching costs beyond software
  • SSO/SAML, directory sync, and IT-approved vendor status create approval-chain friction
datadoor
3.0/10
proprietary data accumulates over time
why this scoremedium confidenceZoom has accumulated behavioral data (meeting patterns, engagement signals, feature usage) and potentially a large...

Zoom has accumulated behavioral data (meeting patterns, engagement signals, feature usage) and potentially a large corpus of transcribed meetings, but there is no public evidence of a proprietary training dataset or a data flywheel that meaningfully differentiates their AI features. The report describes their AI Companion as wrappers around commodity models (Whisper, GPT-4), which implies no proprietary model trained on Zoom-specific data. Meeting content is user-owned and privacy-sensitive, limiting Zoom's ability to exploit it as a training corpus. The data moat is weak relative to the platform's scale.

  • AI Companion described as 'one-prompt wrappers around Whisper and GPT-4' — no evidence of proprietary model trained on Zoom data
  • Meeting content is privacy-sensitive and user-owned — Zoom cannot freely exploit it as a training corpus
  • No public evidence of a proprietary behavioral data flywheel driving product differentiation
regulatory
7.0/10
real licenses, not SOC 2 theater
why this scorehigh confidenceThe regulatory moat is real but vertical-specific. Zoom holds HIPAA BAA capability, SOC 2 Type II, FERPA compliance,...

The regulatory moat is real but vertical-specific. Zoom holds HIPAA BAA capability, SOC 2 Type II, FERPA compliance, FedRAMP authorization (in progress/achieved for government), and PCI considerations for payment flows. These are not trivial — the report explicitly rates enterprise compliance as 'nightmare' tier requiring months and legal spend. However, this moat only applies in regulated verticals (telehealth, education, government, finance). A horizontal indie builder attacking SMB video conferencing faces no regulatory barrier. The moat is high where it exists but doesn't cover the full addressable market.

  • HIPAA BAA, SOC 2 Type II, FERPA, and FedRAMP compliance cited as 'nightmare' tier in the report — months of legal spend
  • Telehealth, education, and finance verticals (the high-ARPU niches) all require compliance posture Zoom already holds
  • HIPAA BAA requires legal agreements, audit processes, and breach notification infrastructure — not just a checkbox
distribution
3.9/10
brand SERP grip, knowledge graph, news flow
take

the blunt take.

Zoom is the Microsoft Office of video calls — it does everything, owns nothing deeply, and charges enterprise prices for features that are table stakes. The wedge isn't cloning Zoom; it's carving off one vertical and going 10x deeper than they ever will.

Their moat is brand recognition and enterprise procurement inertia, not technical depth. The AI Companion features they're now charging for are one-prompt wrappers around Whisper and GPT-4. A focused team can ship a better async-meeting-notes tool, a better telehealth room, or a better dev-standup bot in weeks — and undercut Zoom's per-seat pricing on day one.

cost

cost of competing.

what they charge
Zoom Pro (per user)
$15.99
/ user/mo
Business plan is $21.99/user/mo; AI Companion included on paid plans
annual:$191.88
what running yours costs
01 · Vercel Pro (edge functions for signaling)$20.00
02 · Daily.co or Livekit Cloud (WebRTC infra, starter tier)$0.00
03 · Supabase Pro (rooms, users, recordings metadata)$25.00
04 · OpenAI Whisper + GPT-4 (transcription + AI notes)??? — scales with usage
05 · Cloudflare R2 (recording storage)$1.00
06 · Resend (notifications)$0.00
07 · Domain$1.00
08 · Sentry free tier$0.00
TOTAL / mo$47.00 + usage
▸ break-even:immediately for solo users — Zoom Pro is $15.99/mo vs ~$53/mo self-run, but at 4+ seats your build is cheaper than their per-seat billing
build

what you're up against.

2 weeks WebRTC plumbing · 2 weeks vertical-specific UX · 2 weeks AI meeting notes pipeline · 2 weeks auth, billing, polish
easy
medium
hard
nightmare
01
easy
Auth + room creation CRUD
Standard OAuth + Supabase row-level security. Half a day.
02
easy
Embedding a WebRTC SDK (Daily/Livekit)
Both have React SDKs with working examples. Not your problem to solve.
03
medium
AI meeting notes pipeline
Pipe audio to Whisper, chunk transcript, summarize with GPT-4. Latency and cost management are the real work.
04
medium
Recording storage + playback
R2 for storage, signed URLs for playback. Seek/scrub UX takes longer than the upload logic.
05
hard
Reliable multi-party video at scale
SFU architecture, adaptive bitrate, packet loss recovery — this is why you use Daily or Livekit instead of rolling your own.
06
nightmare
Enterprise procurement & compliance (HIPAA, SOC 2, FERPA)
The vertical niches with real money (telehealth, education, finance) all require compliance posture that takes months and legal spend to establish.
stack

their position.

detected signals· measured
cdnCloudflare
recommended stack · inferred
inferNext.js 15 + ReactinferLivekit Cloud or Daily.co (WebRTC SFU)inferSupabase (Postgres + auth + storage)inferOpenAI Whisper + GPT-4 (transcription/notes)inferCloudflare R2 + CDN
rivals

who else has tried this.

option A
Jitsi Meet (self-host)
fully open-source WebRTC stack, Docker-deployable, no per-seat fees. Lacks AI features but covers 90% of meeting use cases.
option B
Google Meet (free tier)
free for up to 60 minutes, already in every Google Workspace account. Hard to beat on price-to-effort ratio.
option C
Daily.co or Whereby embed
drop-in WebRTC SDK — skip the infra entirely and build only the vertical-specific layer on top.
compare

similar scans.

same shape - different moat
ready to wedge in?
Get the wedge plan. Cancel some plans.
▸ generated with love, by a heartless robotverdict v2.5 · saaspocalypse.dev