Strengths
- Groq publicly claims the LPU was pioneered in 2016 for inference.
- GroqCloud has a documented OpenAI-compatible API and model/pricing table.
- Groq publicly lists multiple global strategic partnerships.
Diligence report
Groq
Groq Public-Source Startup Diligence Report
Proceed only with detailed financial, capacity, contract, and technical benchmarking diligence because the upside depends on whether Groq can translate LPU speed/cost claims and sovereign/enterprise partnerships into durable gross margin against Nvidia and hyperscaler competition.
Company profile
Groq Public-Source Startup Diligence Report
Groq shows strong public evidence of differentiated AI inference positioning, developer-facing products, substantial funding headlines, and named global partnerships; the main diligence constraint is that economics, capacity, customer concentration, and contract terms remain private.
- Website
- groq.com
- Sector
- AI inference hardware and cloud services
- Geography
- United States with announced global data-center and sovereign-AI partnerships
- Stage
- Late-stage private unicorn
- Known aliases
- Groq, Inc., GroqCloud, LPU
- Report version
- 1.0
- Timezone
- UTC
Executive summary
Risks
- Valuation and financial quality are not publicly auditable
- Inference hardware scale-up and capital intensity
- Competition from Nvidia, hyperscalers, and specialist AI clouds
Gaps
- Audited financial statements, ARR, gross margin, burn, cash runway, and revenue concentration.
- Cap table, financing documents, liquidation preferences, and debt/data-center obligations.
- Executed customer, partner, supplier, model-license, DPA, and SLA agreements.
- Independent production benchmarks and capacity/utilization data.
Risk register
R-001: Valuation and financial quality are not publicly auditable
The public record includes funding and valuation anchors, but revenue, gross margin, burn, backlog, and customer concentration are not disclosed.
Diligence request: Request audited financial statements, revenue by channel, gross margins by workload, cash runway, and board-approved forecasts.
R-003: Competition from Nvidia, hyperscalers, and specialist AI clouds
Groq has differentiated public speed/cost claims, but large competitors control GPU supply, cloud distribution, and enterprise procurement channels.
Diligence request: Benchmark latency, throughput, price, availability, and win/loss data against named competitors.
R-002: Inference hardware scale-up and capital intensity
Groq competes in a capital-intensive inference infrastructure market where chip supply, data-center rollout, and utilization determine economics.
Diligence request: Request manufacturing, capacity, utilization, and capex commitments by region and partner.
R-005: International sovereign AI and export-control exposure
Public partnerships span Saudi Arabia, Canada, India, Australia, and U.S. government infrastructure, increasing export-control, sanctions, data-localization, and public-sector compliance complexity.
Diligence request: Map jurisdictions, license requirements, sanctions screening, data residency, and government-contract compliance obligations.
R-004: Enterprise data-processing and privacy obligations
Groq distinguishes controller processing from customer-data processor processing, but data flows, retention, and customer controls require contract review.
Diligence request: Review the Services Agreement, DPA, security reports, subprocessors, incident history, and customer opt-out/retention controls.
R-007: Dependence on third-party open models and ecosystem permissions
GroqCloud public docs emphasize hosted third-party models and OpenAI-compatible APIs, which may create model availability, licensing, and differentiation constraints.
Diligence request: Review model provider agreements, deprecation policies, and proprietary model/tool roadmap.
R-008: Public website terms are not enough to diligence enterprise liabilities
Groq website terms expressly separate cloud services from website terms, so enterprise risk cannot be assessed from public terms alone.
Diligence request: Request all current customer agreements, SLAs, indemnities, limitation-of-liability schedules, and insurance certificates.
R-006: Partner and customer concentration cannot be tested publicly
Named partnerships are visible, but revenue contribution, exclusivity, backlog, and termination provisions are not public.
Diligence request: Request revenue concentration schedules, contract terms, renewal status, and implementation milestones.
01Financial Information
Public sources verify funding and valuation headlines but not financial statements, forecasts, capitalization, debt, or revenue quality.
I.A Annual and quarterly financial information for the past three years
not publicly verifiable confidence: low
No audited financial statements, backlog, AR aging, or product/channel/geography revenue split were public in reviewed sources.
Evidence gaps
- Audited financial statements, management reporting, revenue split, backlog, and AR aging are required.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide audited statements, monthly financials, revenue by workload/customer/geography, COGS by token, backlog, and AR aging.
| signal | public evidence | verification status | private data request |
|---|---|---|---|
| Revenue/ARR | No audited revenue or ARR found in reviewed public sources | not_publicly_verifiable | Audited financials, ARR bridge, bookings, deferred revenue |
| Customer ROI / cost savings | Company-curated quote says chat speed rose 7.41x and costs fell 89% | partially_verified | Customer reference call, invoices, workload mix, gross margin |
| Price/performance inputs | Public model prices and speeds disclosed for selected hosted models | verified | COGS per token, utilization, support cost, discounts |
I.B Financial Projections
not publicly verifiable confidence: low
Funding momentum and product demand headlines suggest growth ambition, but forecasts, capex, and working-capital assumptions are private.
Evidence gaps
- Board-approved forecast, utilization assumptions, capex plan, and financing assumptions are required.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
I.C Capital Structure
not publicly verifiable confidence: low
Founder and financing headlines are public; full capitalization, preferences, debt, warrants, and employee equity remain private.
Evidence gaps
- Cap table, option pool, warrants, SAFEs/notes, debt, investor-rights agreements, and preference stack are required.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
| stakeholder | public position | evidence | diligence caveat |
|---|---|---|---|
| Jonathan Ross | Founder listed in public unicorn list | Founder named for Groq | Actual equity, vesting, and governance rights not public |
| Funding investors | Investors in 2024/2025 rounds not fully extracted from public source snapshots | Public round headlines exist | Request cap table, pro forma ownership, and investor rights |
| Debt or infrastructure financing providers | not_publicly_verifiable | No debt schedule in public sources | Request debt, leasing, and data-center obligation schedules |
I.D Other financial information
partially verified confidence: medium
Public sources provide funding, valuation, and pricing inputs but not tax/accounting policies or full financing terms.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
| date | round or event | amount | valuation | source | verification status |
|---|---|---|---|---|---|
| 2024-08-05 | Series D / funding headline | US$640M | US$2.8B headline linked by Groq newsroom | Groq newsroom / Bloomberg headline | partially_verified |
| 2025-09-17 | Funding headline | US$750M | Not stated in fetched headline | Groq newsroom | partially_verified |
| 2025-12 | Unicorn-list valuation | not_publicly_verifiable | US$20B | Wikipedia unicorn list | partially_verified |
Round documents, investor allocations, liquidation preferences, and secondary-sale terms were not public.
02Products
Groq has a clearly described product surface around LPU inference and GroqCloud, with public model speed and pricing data.
II.A Description of each product
verified confidence: high
Public sources provide limited direct evidence for description of each product; complete diligence requires company data-room support.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide independent benchmark results, product roadmap, uptime/SLA data, and deprecation policy.
| product | audience | key features | public evidence | verification status |
|---|---|---|---|---|
| LPU architecture | Enterprise and cloud AI inference buyers | Purpose-built inference chip / stack | Groq says it pioneered the LPU in 2016 | verified |
| GroqCloud API | Developers and enterprises | OpenAI-compatible API, hosted models, rate limits, billing docs | Docs overview and models page | verified |
| GroqChat / Playground | Developers and evaluators | Cloud-service usage governed separately from website terms | Terms and privacy policy reference GroqChat, Playground, GroqCloud | verified |
| model | public speed | public price | context window | diligence note |
|---|---|---|---|---|
| llama-3.1-8b-instant | 560 tokens/sec | US$0.05 input / US$0.08 output per 1M tokens | 131,072 | Compare realized margin and SLA under production load |
| llama-3.3-70b-versatile | 280 tokens/sec | US$0.59 input / US$0.79 output per 1M tokens | 131,072 | Benchmark against GPU clouds and hyperscalers |
| openai/gpt-oss-20b | 1000 tokens/sec | US$0.075 input / US$0.30 output per 1M tokens | 131,072 | Verify model license, availability, and deprecation risk |
03Customer Information
Named partnerships and customer claims are visible, but revenue concentration and contract economics are private.
III.A Top customers by application
not publicly verifiable confidence: medium
Public sources provide limited direct evidence for top customers by application; complete diligence requires company data-room support.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
III.B Strategic relationships
partially verified confidence: medium
Public sources provide limited direct evidence for strategic relationships; complete diligence requires company data-room support.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Partnership announcements may represent pilots, marketing, or infrastructure commitments rather than recurring revenue.
Follow-up questions
- For each named relationship, provide contract value, term, exclusivity, termination rights, and implementation status.
| relationship | type | public evidence | diligence gap |
|---|---|---|---|
| McLaren Formula 1 Team | Customer/official partner | Newsroom and homepage say McLaren chooses Groq for inference | Contract economics and renewal terms |
| Bell Canada AI Network | Sovereign AI / inference provider | Newsroom states Groq becomes exclusive inference provider | Exclusivity definition, SLAs, revenue commitment |
| Meta Llama API | Model/ecosystem collaboration | Newsroom states Meta and Groq collaborate on official Llama API inference | License, economics, and duration |
| Saudi / Aramco Digital / HUMAIN | Sovereign AI infrastructure | Newsroom lists Saudi expansion and Aramco data-center progress | Capex, ownership, localization, export-control approvals |
III.C Revenue by customer
not publicly verifiable confidence: medium
Public sources provide limited direct evidence for revenue by customer; complete diligence requires company data-room support.
Evidence gaps
- Revenue by named customer/partner and concentration greater than 5 percent are required.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
| account or partner | disclosed weight | public signal | verification status |
|---|---|---|---|
| McLaren | not_publicly_verifiable | Named partnership/customer story | partially_verified |
| Bell Canada | not_publicly_verifiable | Exclusive inference-provider announcement | partially_verified |
| Long tail developers | not_publicly_verifiable | Self-serve API/docs and pricing | partially_verified |
Null bars represent missing customer revenue weights that should be requested.
III.D Significant relationships severed within the last two years
not publicly verifiable confidence: medium
No severed strategic relationship was verified in reviewed public sources; this requires customer and supplier schedules.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
III.E Top suppliers
not publicly verifiable confidence: medium
Public sources provide limited direct evidence for top suppliers; complete diligence requires company data-room support.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
| dependency | role | public evidence | concentration risk |
|---|---|---|---|
| Semiconductor manufacturing and packaging | LPU production | not_publicly_verifiable in reviewed public sources | High if single-foundry or constrained-node exposure exists |
| Data-center footprint and power | Low-latency inference hosting | Groq says LPU-based stack runs in data centers worldwide; Sydney and Saudi public announcements exist | Capacity, power, and regional approvals may constrain growth |
| Open model providers | Hosted models on GroqCloud | Docs list Meta, OpenAI OSS, Alibaba/Qwen and other models | Model license/deprecation risk |
04Competition
Groq competes in a high-pressure AI inference market against GPU incumbents, hyperscalers, and specialist AI chip/cloud providers.
IV.A Competitive landscape by market segment
partially verified confidence: medium
Public sources provide limited direct evidence for competitive landscape by market segment; complete diligence requires company data-room support.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Competitors may compress inference pricing faster than Groq can scale utilization.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
| competitor | segment | product overlap | target differentiator | key diligence test |
|---|---|---|---|---|
| Nvidia / GPU cloud ecosystem | AI accelerators and inference stack | Inference compute and enterprise AI workloads | Purpose-built LPU and public speed/cost claims | Price/performance under production workloads |
| Cerebras / SambaNova / specialist AI chip clouds | Alternative AI accelerators | Inference and model-serving infrastructure | GroqCloud developer distribution and named partnerships | Availability, model coverage, gross margin |
| Hyperscaler inference services | Cloud AI platforms | Hosted models and enterprise procurement | Low-latency LPU story | Enterprise attach, cloud marketplace, SLAs |
| axis | groq public position | competitor pressure | evidence |
|---|---|---|---|
| Latency / throughput | High based on public model speeds | High | Groq docs list tokens/sec by model |
| Cost per token | Low-cost positioning with published prices | High | Homepage and model pricing |
| Distribution | Developer API plus enterprise/sovereign partnerships | Very high from hyperscalers | Docs and newsroom |
05Marketing, Sales, and Distribution
The public GTM model combines self-serve developer API, enterprise/sovereign partnerships, and high-visibility marketing relationships.
V.A Strategy and implementation
partially verified confidence: medium
Public sources provide limited direct evidence for strategy and implementation; complete diligence requires company data-room support.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
| channel | region | public evidence | gap |
|---|---|---|---|
| Self-serve developer API | Global | Groq docs quickstart, models, API, billing, rate limits | Conversion, retention, support cost |
| Enterprise / sovereign AI partnerships | Canada, Saudi Arabia, India, U.S., Australia | Newsroom partnership announcements | Contracted ARR, implementation status |
| Brand/marketing partnerships | Global | McLaren Formula 1 partnership | Attribution and customer acquisition cost |
| signal | evidence | verification status | diligence note |
|---|---|---|---|
| Homepage positioning | Fast, low-cost inference message | verified | Test NPS and win/loss versus claims |
| Developer documentation | OpenAI-compatible quickstart and model catalog | verified | Analyze activation funnel and paid conversion |
| Newsroom cadence | Multiple 2025 partnership and funding announcements | verified | Separate pipeline PR from contracted revenue |
V.B Major Customers
not publicly verifiable confidence: medium
Public sources provide limited direct evidence for major customers; complete diligence requires company data-room support.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
V.C Principal avenues for generating new business
partially verified confidence: medium
Public sources provide limited direct evidence for principal avenues for generating new business; complete diligence requires company data-room support.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
V.D Sales force productivity model
not publicly verifiable confidence: medium
Sales productivity, quotas, sales cycle, and compensation are not public despite visible developer and partnership motions.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
V.E Ability to implement marketing plan with current and projected budgets
not publicly verifiable confidence: medium
Marketing budget sufficiency cannot be verified publicly; request budget, CAC, and pipeline conversion.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
06Research and Development
R&D evidence centers on LPU architecture, GroqCloud, model support, and global infrastructure; R&D staffing and costs are private.
VI.A Description of R&D organization
not publicly verifiable confidence: medium
Public sources provide limited direct evidence for description of r&d organization; complete diligence requires company data-room support.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
| person or role | public role | source | diligence gap |
|---|---|---|---|
| Jonathan Ross | Founder listed in unicorn table | Wikipedia unicorn list | Confirm current operating role, equity, and succession plan |
| LPU engineering organization | not_publicly_verifiable by function/headcount | Public pages describe LPU but not org structure | Request R&D roster, retention, roadmap ownership |
VI.B New Product Pipeline
partially verified confidence: medium
Public sources provide limited direct evidence for new product pipeline; complete diligence requires company data-room support.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
| project | status | public evidence | risk |
|---|---|---|---|
| LPU inference platform | Commercialized / scaling | Homepage and docs describe LPU and GroqCloud | Manufacturing capacity and utilization |
| Open model support | Active model catalog | Models page lists production and preview models | Model license/deprecation and third-party dependency |
| Global data centers / sovereign AI | Announced partnerships | Sydney, Saudi, Bell, DOE announcements listed | Regulatory and capex execution |
07Management and Personnel
Founder information and careers links are public, but management roster, headcount, compensation, and turnover data require company records.
VII.A Organization Chart
not publicly verifiable confidence: medium
Public sources provide limited direct evidence for organization chart; complete diligence requires company data-room support.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
VII.B Historical and projected headcount by function and location
not publicly verifiable confidence: low
Public sources provide limited direct evidence for historical and projected headcount by function and location; complete diligence requires company data-room support.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
| function | public evidence | verification status | request |
|---|---|---|---|
| Engineering/R&D | Product requires LPU and cloud engineering; no function-level headcount disclosed | not_publicly_verifiable | HRIS export by function/location |
| Sales/partnerships | Named strategic partnerships imply enterprise GTM; no sales productivity data public | not_publicly_verifiable | Quota capacity, pipeline, CAC payback |
This chart intentionally shows absent public headcount anchors.
VII.C Senior management biographies
not publicly verifiable confidence: medium
Public sources provide limited direct evidence for senior management biographies; complete diligence requires company data-room support.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
| name | role | tenure signal | source |
|---|---|---|---|
| Jonathan Ross | Founder | Founder listed with Groq in public unicorn table | Wikipedia unicorn list |
| Full executive staff | not_publicly_verifiable in reviewed snapshots | Needs direct company confirmation | Analyst gap assessment |
VII.D Compensation arrangements
not publicly verifiable confidence: medium
Employment agreements, benefits, and compensation plans are not public.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
VII.E Incentive stock plans
not publicly verifiable confidence: medium
Equity plan documents and option schedules are not public.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
VII.F Significant employee relations problems, past or present
not publicly verifiable confidence: medium
Employee-relations issues were not verifiable from reviewed public sources.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
VII.G Personnel Turnover
not publicly verifiable confidence: medium
Public sources provide limited direct evidence for personnel turnover; complete diligence requires company data-room support.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
| category | public signal | verification status | diligence request |
|---|---|---|---|
| Executive departures | No systematic public turnover source reviewed | not_publicly_verifiable | List departures and regretted attrition for last 24 months |
| Technical retention | No retention metrics public | not_publicly_verifiable | R&D attrition, offer acceptance, key-person risk |
08Legal and Related Matters
Privacy and terms pages provide legal signals, but material contracts, litigation, insurance, export-control, and IP schedules require counsel-led diligence.
VIII.A Pending lawsuits against the Company
not publicly verifiable confidence: medium
Public sources provide limited direct evidence for pending lawsuits against the company; complete diligence requires company data-room support.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
| case | court | filed date | status | source |
|---|---|---|---|---|
| No material pending lawsuit verified in reviewed public sources | not_publicly_verifiable | not_publicly_verifiable | Requires counsel confirmation | Analyst gap assessment |
| IP, employment, commercial claims | not_publicly_verifiable | not_publicly_verifiable | No docket pull performed | Analyst gap assessment |
VIII.B Pending lawsuits initiated by Company
not publicly verifiable confidence: medium
Public sources provide limited direct evidence for pending lawsuits initiated by company; complete diligence requires company data-room support.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
| defendant | court | filed date | status | source |
|---|---|---|---|---|
| No company-initiated litigation verified in reviewed public sources | not_publicly_verifiable | not_publicly_verifiable | Requires counsel confirmation | Analyst gap assessment |
| IP enforcement matters | not_publicly_verifiable | not_publicly_verifiable | No docket pull performed | Analyst gap assessment |
VIII.C Environmental and employee safety issues and liabilities
not publicly verifiable confidence: medium
Environmental, safety, power, and data-center liabilities are not disclosed in public source snapshots.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
VIII.D Material patents, copyrights, licenses, and trademarks
not publicly verifiable confidence: medium
Public sources provide limited direct evidence for material patents, copyrights, licenses, and trademarks; complete diligence requires company data-room support.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
| asset or action | jurisdiction | public status | diligence request |
|---|---|---|---|
| LPU architecture / GroqCloud marks | U.S. and international | Company publicly uses LPU and GroqCloud brands | USPTO/WIPO patent and trademark schedule; invention assignment audit |
| Export controls / sovereign AI deployments | U.S., Canada, Saudi Arabia, India, Australia | Announced global infrastructure and public-sector relationships | Export-control memo, sanctions screening, government-contract terms |
| Data privacy compliance | U.S., EU/UK/Switzerland, Saudi Arabia and others | Privacy policy references international transfers and Data Privacy Framework safeguards | DPA, subprocessors, SCCs, audit reports |
VIII.E Insurance coverage and material exposures
not publicly verifiable confidence: medium
Public sources provide limited direct evidence for insurance coverage and material exposures; complete diligence requires company data-room support.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
VIII.F Material contracts
partially verified confidence: high
Public sources provide limited direct evidence for material contracts; complete diligence requires company data-room support.
Evidence gaps
- Executed customer agreements, strategic partner contracts, SLAs, DPAs, indemnities, and insurance policies are required.
Hidden risks
- Private contracts, customer economics, and operating metrics may differ materially from public positioning.
Follow-up questions
- Provide source documents and management explanations for this checklist section.
| contract or policy | public evidence | risk | request |
|---|---|---|---|
| Groq Services Agreement / DPA | Referenced by privacy policy and terms but not fully diligenced | Customer data, indemnity, and SLA exposure | Executed customer forms, DPA, subprocessor list |
| Strategic partnership contracts | Named relationships public | Exclusivity, capex, termination, localization | Master agreements and statements of work |
| Insurance | not_publicly_verifiable | Cyber, E&O, product, D&O, and supply-chain coverage unknown | Certificates and claims history |
VIII.G Regulatory agency problems
not publicly verifiable confidence: medium
Public sources provide limited direct evidence for regulatory agency problems; complete diligence requires company data-room support.
Evidence gaps
- Request the underlying private data-room materials, board reporting, contracts, and management reconciliations.
Hidden risks
- Global sovereign AI deployments may create regulatory obligations not visible in public announcements.
Follow-up questions
- Provide export-control, sanctions, government-contract, privacy, and data-residency compliance memos.
| risk id | risk title | severity | evidence basis |
|---|---|---|---|
| R-004 | Enterprise data-processing and privacy obligations | medium | EC-008, EC-009 |
| R-005 | International sovereign AI and export-control exposure | high | EC-007 |
| R-006 | Partner and customer concentration cannot be tested publicly | medium | EC-007, EC-011, EC-012 |
| R-007 | Dependence on third-party open models and ecosystem permissions | medium | EC-004, EC-005 |
| R-008 | Public website terms are not enough to diligence enterprise liabilities | medium | EC-009 |
Evidence
| ID | Claim | Status | Sources |
|---|---|---|---|
| EC-001 | Groq appears on a public unicorn list at a $20 billion valuation in December 2025. | partially verified medium | SRC-001 |
| EC-002 | Groq was established in 2016 and positions its LPU as purpose-built for inference. | verified high | SRC-002 |
| EC-003 | Groq claims fast, low-cost inference and a globally deployable LPU stack. | verified high | SRC-002 |
| EC-004 | GroqCloud is presented as a fast, OpenAI-compatible developer API. | verified high | SRC-003 |
| EC-005 | Groq publishes model speed, price, and rate-limit information for hosted models. | verified high | SRC-004 |
| EC-006 | Groq has public funding headlines including a $640 million raise and a later $750 million raise. | verified medium | SRC-005SRC-008 |
| EC-007 | Groq has announced multiple strategic partnerships and global infrastructure relationships. | verified medium | SRC-005 |
| EC-008 | Groq privacy policy separates customer data processed for cloud offerings from website/controller processing. | verified high | SRC-006 |
| EC-009 | Groq website terms do not govern GroqCloud customer usage. | verified high | SRC-007 |
| EC-010 | Groq public pages identify careers and policy pages but do not disclose detailed headcount by function. | not publicly verifiable low | SRC-002SRC-000 |
| EC-011 | Groq marketing includes customer ROI claims such as faster chat and lower cost. | partially verified medium | SRC-002 |
| EC-012 | Core financial statements, revenue concentration, cap table, debt, legal dockets, and customer contracts were not publicly available in the reviewed sources. | not publicly verifiable low | SRC-000 |
| ID | Publisher | Title | Accessed |
|---|---|---|---|
| SRC-000 | GitHub Copilot | Analyst public-source evidence gap assessment | 2026-06-07 |
| SRC-001 | Wikipedia | Wikipedia - List of unicorn startup companies | 2026-06-07 |
| SRC-002 | Groq | Groq homepage | 2026-06-07 |
| SRC-003 | Groq | Groq Docs - Overview | 2026-06-07 |
| SRC-004 | Groq | Groq Docs - Supported Models | 2026-06-07 |
| SRC-005 | Groq | Groq newsroom | 2026-06-07 |
| SRC-006 | Groq | Groq privacy policy | 2026-06-07 |
| SRC-007 | Groq | Groq website terms of use | 2026-06-07 |
| SRC-008 | Bloomberg | Bloomberg headline linked by Groq newsroom: AI Chip Startup Groq Gets $2.8 Billion Valuation | 2026-06-07 |
Disclaimer
This report is a public-evidence diligence snapshot, not investment advice. Important financial, legal, technical, and contractual facts remain non-public and should be verified directly with management and primary documents before any investment decision.