# Zapata AI SPAC Financial Model Zapata Computing is an Industrial Generative AI software company (quantum-inspired algorithms + LLM compression + the Orquestra® full-stack platform) merging with Andretti Acquisition Corp. via a SPAC de-SPAC transaction. - Canonical: https://finamodel.com/startups/zapata-ai-spac - Excel download: https://finamodel.com/startup-models/zapata-ai-spac.xlsx - Category: AI/ML - Model type: SPAC / De-SPAC - Funding round: SPAC - Funding: $331M - Founded: 2024 - Geography: US-headquartered (Boston / Harvard spin-out 2017); global enterprise customers and partners. [DECK] Slide 4, 25. - Customer: B2B ## About the company Zapata AI is an industrial generative-AI company modelled through a SPAC transaction. The case combines an operating forecast for software and services with transaction financing, public-market valuation, ownership dilution, and the cash available after redemptions. Its commercial model requires adoption scenarios for enterprise software and services, alongside deep-tech R&D and compute spending. The transaction model must capture SPAC trust proceeds, sponsor and founder ownership, warrants, bridge financing, and pro forma capitalisation at closing. The forecast links revenue, gross margin, headcount, R&D, compute, and operating burn to transaction cash and runway. It includes sources and uses, redemption sensitivity, dilution, pro forma balance sheet, ownership, valuation multiples, and scenario-based financing requirements after the de-SPAC. ## What's included - 5-year monthly revenue build with stage-appropriate growth assumptions - Full P&L, headcount plan, and operating-expense schedule - Cash-flow statement, runway, and burn-rate tracking - Valuation via exit multiple with a DCF cross-check - Returns analysis with MOIC and IRR - Unit economics including CAC, LTV, payback, and cohort retention ## Product & value proposition - Two core products sold under the Orquestra® platform: Slide 4, 18, 19 - **Zapata AI Prose™** - LLM compression and fine-tuning for enterprise text applications; compressed models require 300x fewer tokens to match uncompressed model performance Slide 13. - **Zapata AI Sense™** - Quantum-inspired generative models for complex numerical/mathematical problems (optimization, simulation, derivative pricing). - Key performance claims (all from internal Zapata data): - 8,400x speedup vs Monte Carlo simulation for multi-asset options pricing Slide 14. - Up to 10x–1,000x speed-up on LLMs and other large models Slide 26. - Compressed GPT2-XL outperforms GPT2-Small (same model size) in F1 accuracy Slide 13. - GEO (Generator Enhanced Optimization) tied or outperformed state-of-the-art solvers in 71% of configurations for BMW manufacturing scheduling Slide 17. - Platform: Orquestra® is hardware- and cloud-agnostic; supports CPU/GPU/TPU/QPU; integrates with AWS, Azure, GCP, IBM Cloud, and on-prem. Slide 18. ## Market - TAM: $1.3T (Generative AI Software and adjacencies by 2032) Slide 21 - source: Bloomberg Intelligence, June 2023. - SOM: $366B by 2032 ($280B Generative AI Software + $86B Generative AI IT Services) Slide 21. - McKinsey estimate: 63 Gen AI use cases across 16 business functions could deliver $2.6T–$4.4T annual P&L impact for enterprise Slide 21. - Key verticals targeted: Automotive, Chemicals & Materials, Defense, Energy & Utilities, Finance, Logistics, Pharma Slide 23. ## Revenue model - Model: Recurring subscription; contracts recognized ratably. Slide 22. - Contract length: 6+ month agreements with intent for multi-year subscriptions. Slide 22. - Bundled offering: software license + scientific algorithm expertise (not pure SaaS; includes professional services component). Slide 22. - Two sales motions: Slide 22 1. Direct - C-level relationships, global sales force. 2. Partner ecosystem - consulting firms (Top 5 Global Consultancy), cloud/software (Microsoft Azure, NVIDIA), hardware (IBM, IonQ), academia (MIT, University of Toronto). - Land-and-expand: initial application foothold → expand ARPA (Average Revenue Per Account). Slide 22. ## Traction & metrics - Customers (named): Andretti Autosport, BASF, BBVA, bp, DARPA, top 5 food & bev company (unnamed), leading Clinical Research Organization (unnamed). Slides 4, 20. - Partners (named): AWS, NVIDIA, IBM, IonQ, Microsoft Azure, top 5 global consultancy (unnamed). Slide 4. - Total funding raised to date: $64M Slide 25. - Headcount (as of Aug 28, 2023): 60 employees total; 41 scientists & engineers; 24 PhDs. Slide 25. - IP: 50 US patent families & applications (100+ worldwide as of June 2023); ranked #5 most active in quantum-inspired AI patents (2020–2021, EPO). Slide 12. - Revenue, ARR, growth rates, customer count, churn, NRR: **Not in deck.** - DARPA: 2 awards to benchmark utility of quantum computing. Slide 20. - Gartner Cool Vendor recognition. Slide 23. ## Competition / moat - Moat framing: Slides 9, 11, 12, 19 - Proprietary quantum-inspired generative AI algorithms (not reliant on quantum hardware - works on classical hardware today). - Globally ranked IP portfolio: #5 in quantum-inspired AI patents 2020–2021; portfolio mix: Algorithms & Software 38%, Generative AI 30%, Differential Equations & Optimization 22%, Hardware Optimization 10%. - Founded by Harvard quantum computing PhDs; 24 PhDs on staff; ~85K total academic citations (includes founder Alan Aspuru-Guzik's ~57.5K citations). - Hardware/cloud agnostic platform vs. Big Tech's locked-in, one-size-fits-all LLMs. - Named competitive positioning: positioned against OpenAI/ChatGPT-style general-purpose LLMs as "Industrial" specialist. Slide 9. - Direct competitors: Not named in deck. ## Team & funding ask / use of funds **Zapata AI leadership:** Slide 25 - Christopher Savoie, Ph.D. - CEO; inventor of NLU behind Apple's Siri; formerly Nissan, Verizon. - Yudong Cao, Ph.D. - CTO; 10 years AI & quantum; 30 patents; 2.4K+ citations; Harvard/Purdue. - Mimi Flanagan - CFO; IHS Markit, Jenzabar, Qlik, Wyndham. - Tim Stanley - VP Global Sales; 25+ years enterprise sales. - Nicole Fitchpatric - General Counsel. **Andretti Acquisition Corp. (SPAC) leadership:** Slide 5 - Bill Sandbrook (Chairman & Co-CEO): Former Chairman/CEO U.S. Concrete (25x market cap, 24x EBITDA, 35+ acquisitions). - Michael Andretti (Co-CEO): IndyCar Champion, Founder/CEO Andretti Autosport. - Matt Brown (President & CFO): Former CEO Rocky Mountain Industrials, former CFO Forterra, former CFO U.S. Concrete. **SPAC:** Andretti Acquisition Corp. raised $230M oversubscribed SPAC IPO (NYSE: WNNR). Slide 5. **Transaction / Use of Funds:** Slide 24 - Sources: Cash in Trust $50M + Bridge Financing $20M + Zapata Rollover $200M = $270M total. - Uses: Cash to Balance Sheet $48M + Transaction Expenses $12M + Zapata Rollover $200M + Growth Capital $10M = $270M. - $48M going to balance sheet (operating runway). - $10M earmarked as "Growth Capital." - Bridge: convertible notes at 15% discount to DeSPAC; up to $14.5M new + $5.5M existing Zapata convertible debt. **Investors (pre-SPAC):** Ahren, Alumni Ventures, BASF, Bosch, Comcast Ventures, The Engine, Honeywell, ITOCHU, Merck, Pillar, Pitango, Prelude Ventures. $64M total raised. Slide 25. --- ## Recommended financial model - **Archetype + why:** SPAC / de-SPAC deal model - not an operating forecast model. This deck is a de-SPAC investor presentation, not a growth-stage pitch. The appropriate financial model is a **SPAC / de-SPAC transaction model** covering: (1) sources & uses and pro-forma capitalization, (2) dilution waterfall across redemption scenarios, (3) post-close balance sheet bridge, and (4) a lightweight operating scenario bridge (revenue ramp to illustrate when the company could hit key milestones). If the goal is an operating model for Zapata AI as a standalone company, a **B2B SaaS / enterprise software ARR model** is the secondary recommendation - but all revenue numbers would be assumed, as the deck discloses zero actual financials. - **Forecast horizon & granularity:** - Deal model: point-in-time (at close) + 1-year post-close sensitivity. - Operating bridge (if built): Annual, 5 years (2024–2028), monthly for Year 1. - **Key drivers & assumptions:** *SPAC deal model:* - SPAC trust size: $84.2M (7.9M shares × $10.66/share) Slide 24 note. - Assumed redemption rate: 40% base case Slide 24 note → $50M cash in trust after redemption. - Bridge financing: $20M convertible note at 15% discount to DeSPAC price Slide 24. - Zapata pre-money equity value: $200M Slide 24. - Pro-forma equity value: $331M Slide 24. - Pro-forma enterprise value: $283M ($331M equity − $48M cash + $0 debt) Slide 24. - Redemption sensitivity: 0%, 20%, 40%, 60%, 80%, 100% - varies cash to balance sheet and public shareholder % - Founder share earnout / promote: 5.8M shares (17.5%) - no earnout condition disclosed; modeled as fully vested at close. - Warrant dilution: excluded per slide note; model should add a warrant schedule. *Operating bridge (supplementary):* - Initial ARR / revenue: $0–$5M at close (pre-revenue scale; no figures disclosed); set as open variable. - ARR growth rate: 100%+ YoY early years, decelerating to 50% by Year 3, given enterprise AI category growth and land-expand model. - Average contract value (ACV): $500K–$2M based on enterprise C-level direct sales and 6+ month contracts; no data in deck. - Gross margin: 60–70% - typical for enterprise AI/SaaS with significant scientific services component; lower than pure SaaS. - R&D / S&M / G&A as % of revenue: heavy burn pre-revenue; model with headcount-based opex build (60 FTEs × avg fully-loaded cost). - Cash runway: $48M balance sheet post-close Slide 24; 24–36 months at current burn, flex variable. - Customer count: 7–10 active paying customers at close (based on named accounts). - NRR / expansion: 110–130% - land-expand model stated; no data. - **Scenarios (Base / Bull / Bear - which variables flex):** - Bear: 80% redemption (less cash), slow revenue ramp ($3M ARR Year 1), high burn, Series A needed in 18 months. - Base: 40% redemption ($48M post-close cash), $5M ARR Year 1, 24-month runway. - Bull: 10% redemption (more cash), fast enterprise wins ($10M ARR Year 1), gross margins above 70%. - **Required sheets / outputs:** 1. Transaction Summary - Sources & Uses, Pro-Forma Cap Table (Zapata / Public / Founder / Bridge breakdown). 2. Redemption Sensitivity Table - cash to balance sheet, pro-forma equity value, and ownership % by redemption level (0–100%, 10% steps). 3. Dilution Schedule - warrants, bridge conversion at DeSPAC, PIK interest on bridge note. 4. Pro-Forma Balance Sheet (at close). 5. Operating Bridge (assumptions-driven) - ARR build, opex by function, EBITDA, cash burn, runway. 6. Valuation Bridge - EV/ARR implied at various revenue scenarios. ## Frequently asked questions ### Is the Zapata AI SPAC financial model free? Yes. The Zapata AI SPAC model is a free Excel download with live formulas. ### Can I change the assumptions? Yes. The workbook is editable and its live formulas recalculate when assumptions change.