Contrarian Corpus
short seller full deck initial thesis
2022-02-16 · 88 pages

C3.ai, Inc. AI

C3.ai is a serially rebranded, money-losing AI story-stock propped up by an unraveling Baker Hughes JV and aggressive accounting; Spruce Point sees 40-50% downside to $12.85-$15.40.

N 5 Narrative
V 3 Visual
C 3 Craft
Unlock to download PDF Spruce Point research ↗

Thesis

Spruce Point argues C3.ai rode the 2020 AI-hype IPO window with a business that has pivoted three times (C3 Energy to C3 IoT to C3.ai) and depends on a shaky Baker Hughes joint venture representing over 30% of sales, now amended three times with shrinking minimums and a suspicious $16m 'sales commission' from C3 to BH. A forensic review flags 99%+ gross-margin related-party revenue, unbilled-receivables anomalies, a revolving door of three CFOs since IPO (newest with a domestic-violence arrest), and marquee partnerships with Microsoft, Google, Intel and HPE that Spruce Point's channel checks say are essentially fictitious. Governance concerns include Condoleezza Rice and Mike McCaffery's Makena Capital ties to the Siebel foundation, an endowed Berkeley chair funding a director, and $596m of CEO stock sales at ~$65. Valuing C3 as a Baker Hughes run-off plus discounted 'other customers' yields a $12.85-$15.40 price target, 40-50% below the $25.69 reference.

SCQA

Situation

C3.ai IPO'd in December 2020 at $42, surged to $177 on AI-hype, and markets a 'best-of-breed' enterprise AI platform backed by Tom Siebel, Microsoft, Baker Hughes, Google, Intel and HPE.

Complication

Forensic review finds the company is on its third rebrand, its largest customer (Baker Hughes, 30% of sales) has renegotiated three times with shrinking minimums, marquee alliances are hollow, and three CFOs have churned since IPO amid aggressive accounting.

Resolution

Investors should sell; C3 must clarify related-party flows (BH sales commission, Makena Capital, DTI lease), adopt a fraud clawback policy, and disclose executive departures via 8-K rather than quietly removing bios.

Reward

Valuing C3 as a run-off of Baker Hughes revenue plus discounted 'other customers' implies a $12.85-$15.40 price target — 40% to 50% downside from the $25.69 reference price over the next 12 months.

The three reasons

  1. 1

    Baker Hughes JV (30% of sales) has been renegotiated three times and is propping up revenue

  2. 2

    Aggressive accounting: $16m 'sales commission' to BH and 99%+ gross-margin related-party revenue

  3. 3

    Board stacked with Siebel allies (Makena Capital, endowed chair) with undisclosed conflicts

Primary demands

  • Sell C3.ai stock (Strong Sell opinion)
  • C3 should clarify murky financial relationships with Baker Hughes, Google, and directors
  • Disclose nature of related-party transactions and executive departures via 8-K filings
  • Adopt a formal clawback policy for fraud
  • Reassess Board objectivity given Makena Capital and Siebel Foundation entanglements

KPIs cited

Price target (downside)
$12.85 - $15.40, implying 40%-50% downside from $25.69
Baker Hughes % of sales
Over 30% of FY21 sales; largest customer
Revenue per salesperson
~$5,100 across 41,000 alliance salespeople — implies hollow partnerships
Cumulative tech platform investment
C3 claims $1.0bn invested; Spruce Point estimates $500-$785m
Total Addressable Market
C3 revised TAM from $170bn to $271bn during IPO process
C3 market share
Implied 0.12% — inconsistent with claimed market size
CEO stock sales 2021
Tom Siebel sold $596m at avg $65.17; Edward Abbo sold $44m at avg $69.44
BH sales commission
$16m paid to BH on ~$11m of eligible revenue (could be as low as $2.6m)
Related-party gross margin
99%+ on BH professional-services revenue; 100% on some lines
Non-BH gross margin
Mid-60s% range — low for enterprise software
Revenue growth FY18
60% actual vs 88% claimed (Deloitte-era discrepancy)
CFO count since IPO
Three CFOs; David Barter walked from 900,000 options at $17.10 strike (~$20m intrinsic)
Siebel voting control
Siebel and allies own 95% of super-voting Class B shares, 71% voting control
Dual-class dilutive adjustment
34.4m dilutive options at $6.39 plus 6m RSUs = 31.2m additional shares

Pattern membership

Where this document fits across the library's 12 rhetorical / structural patterns. Orange cells are present in this deck; neutral cells are not.

Precedents cited

  • SEC v. Siebel Systems (2002 Reg FD violation)
  • Kior (Condoleezza Rice board role, SEC fraud charges)
  • Shutterfly non-reliance opinion (Patricia House)
  • KB Home non-reliance opinion (Mike McCaffery)
  • Telenav revenue restatement (new CFO Adeel Manzoor)
  • GE / Harry Markopolos whistleblower report (Baker Hughes management)
  • Spruce Point's prior shorts: iRobot, Bazaarvoice, Lightspeed

Composition what's on the 88 slides

Visual + textual elements counted across every slide in this deck. Hover a box for what that element is; click to see every slide in the corpus that uses it.

Chart types used in this deck

Slide gallery ·

All 88
No slide inventory yet

Pass-2 extraction may still be in progress for this deck.

Notes

Classic Spruce Point short-report template: creative photo-illustration cover (cockroaches, rat, cookie jar, 'Sell C3.ai' graffiti) followed by dense institutional-grade body slides. Stacks multiple thesis types — fraud/accounting concerns (primary), governance/board conflicts, management-change (CFO revolving door) — which is why thesis_types lists three. No stake disclosed (short report, not an activist long); author is the firm (Ben Axler's Spruce Point) with no individual signature on the cover. Precedents section on p.3 uses Spruce Point's own prior wins (iRobot, Bazaarvoice, Lightspeed) as credibility priors — a distinctive short-seller rhetorical move. Heavy use of comparison tables (claim vs reality) and call-out boxes.