Basis
B2B SaaS sold to accounting/CPA firms; seat- and workflow/agent-based subscriptions layered on autonomous 'long-horizon' AI agents that execute firm workflows under human review. Private, VC-funded — not yet profitability- or IPO-disclosed.
Only the Series B valuation ($1.15B) is publicly disclosed; seed and Series A post-money were not disclosed and are plotted as conservative round-size estimates to show trajectory, not as reported marks. Total raised ~$138M.
Earnings, margins, COGS & capex
Private and pre-revenue-disclosure. Basis has raised ~$138M across seed ($3.6M, 2023, Better Tomorrow Ventures, with BoxGroup and Abstract Ventures), Series A ($34M, Dec 2024, Khosla Ventures) and Series B ($100M, Feb 2026, Accel-led at $1.15B post). No revenue, ARR, margin, or burn figures are public. Traction is reported in adoption terms rather than dollars: deployed at ~30% of the top 25 US accounting firms and ~20% of the top 150, with press-cited productivity lifts of 20-50% in targeted workflows.
Revenue trend
Margins
software-typical high gross margin expected, pressured by per-agent LLM inference cost
negative expected — venture scale-up phase
COGS structure
Not disclosed. Dominant variable cost is foundation-model inference/API spend (Basis buys from 'leading foundation labs' rather than training its own models) plus hosting; customer success/implementation with accounting firms adds services cost.
Capex
Not disclosed; minimal physical capex. Capital is deployed into engineering/ML headcount and go-to-market, per stated use of the Series B.
Latest earnings
n/a
n/a — no public guidance
- Post-money valuation
- $1.15B (Feb 2026)
- Total raised
- ~$138M
- Top-25 US firm penetration
- ~30%
- Top-150 US firm penetration
- ~20%
- Reported productivity lift
- 20-50% in targeted workflows
Growth drivers
- Land-and-expand across the top-150 US accounting firms — deepening from bookkeeping/statement prep into higher-value tax and audit agents
- Agent scope expansion — complex partnership (1065) tax returns, multi-state filings, audit workpapers — each new workflow is a new revenue surface
- Structural CPA labor shortage + Big Four/mid-market margin pressure driving firm demand for automation
- Frontier-model capability gains (longer-horizon reasoning) directly increasing the share of work agents can complete autonomously
- Blue-chip investor + operator network (Accel, GV, Khosla, Nat Friedman/NFDG, Keith Rabois, Lloyd Blankfein) accelerating enterprise firm access
Bull & bear
Basis is an early leader in the highest-ROI vertical for AI agents — professional-services labor — with real enterprise traction, a differentiated autonomous architecture, and a cap table that opens doors. If agents keep absorbing more of the tax/audit workflow, Basis compounds into a category-defining system-of-work.
- Accounting/tax/audit is a near-ideal agent use case: structured, rule-bound, high-value, chronically understaffed — automation demand is secular, not cyclical
- ~30% top-25-firm penetration in under three years shows the product clears a very high enterprise trust bar; expansion within those logos is high-margin
- Every frontier-model improvement is a free tailwind — the automatable share of each engagement rises without Basis raising its own R&D
- Vertical depth (ontologies, evals, firm integrations, partnership-tax agents) is a widening moat horizontal AI tools and copilots can't shortcut
- World-class backers (Accel, GV, Khosla) + operators (Blankfein, Friedman, Rabois) signal both conviction and privileged access to enterprise buyers and later capital
A $1.15B valuation on undisclosed revenue prices in years of flawless execution in a market where incumbents own distribution, rivals are multiplying, and a single autonomous-agent error in a regulated deliverable can be existential. The moat may be thinner than the round implies.
- No public revenue/ARR — the valuation is an adoption-and-narrative multiple; if monetization or net retention disappoints, the markdown risk is large
- Gross-margin dependence on third-party model inference caps software-like economics and is outside Basis's control
- Distribution disadvantage vs Intuit/Thomson Reuters/Wolters Kluwer, who can ship 'good-enough' agents into their installed base at near-zero switching cost
- Crowded, well-capitalized field (Truewind, Digits, Puzzle, Numeric, Rillet, Zeni, Pennylane) invites price compression and blurs differentiation
- Regulatory/liability tail on autonomous tax filings and audit workpapers is severe — a public failure could freeze the conservative buyer base and invite PCAOB/AICPA constraints
What it is worth
Private last-round mark (post-money) cross-checked vs vertical-AI comps; no public multiples possible without disclosed revenue.
<$0.7B down-round / markdown
if monetization lags, incumbents commoditize agents, or a high-profile accuracy/liability failure chills the conservative buyer base.
~$1.15B holding
with the next round gated on demonstrating revenue scale and net retention to justify the mark.
>$3B in a 12-24 month up-round
if Basis proves durable ARR growth, expands into tax/audit billings, and defends against incumbents — path to a category-leader IPO.
$1.15B post-money (Feb 2026 Series B) is a narrative/adoption valuation — undisclosed ARR means any implied revenue multiple is speculative. Comparable vertical-AI unicorns have priced at very high forward-ARR multiples (often 30-100x+ on early ARR), so the mark is defensible only if ARR is growing fast off a real base and net retention is strong. Range below reflects execution scenarios, not a public-market price.
SWOT
Strengths
- Genuine agentic architecture built for autonomous, long-horizon execution — differentiated from incumbents' 'copilot/assist' AI bolt-ons
- Strong early enterprise proof — already inside ~30% of the top-25 US accounting firms (named customers include UHY, Boulay, Clark Nuber, MarksNelson, Pinion), a hard channel to penetrate
- Top-tier cap table and operator network (Accel, GV, Khosla, Nat Friedman/NFDG, Keith Rabois, Lloyd Blankfein) — capital + distribution + credibility
- Vertical focus (accounting/tax/audit) yields deep domain ontologies and evals that a horizontal AI tool cannot easily replicate
Weaknesses
- No disclosed revenue/ARR or path-to-profit — $1.15B valuation rests on adoption metrics, not proven monetization
- Gross margin exposed to foundation-model inference costs it does not control (buys models, doesn't own them)
- Small, young company (founded 2023) selling into conservative, risk-averse, regulated buyers where a single accuracy failure is reputationally costly
- Accuracy/liability tail — an autonomous agent error in a filed return or audit opinion carries professional-liability and regulatory consequences the vendor must design around
Opportunities
- Large TAM — US accounting/tax/audit services is a multi-hundred-billion-dollar labor market ripe for automation amid a CPA shortage
- Move up-value from bookkeeping into tax and audit, where billing rates (and automation ROI) are far higher
- International expansion into other regulatory regimes once US workflows are proven
- Potential to become the system-of-work/agent layer that sits atop legacy tax/audit software (Thomson Reuters, Wolters Kluwer, Intuit)
Threats
- Incumbents with distribution — Intuit, Thomson Reuters, Wolters Kluwer, Sage — embedding agents natively into the tools firms already run
- Foundation-model providers (OpenAI, Anthropic, Google) moving down-stack into vertical agents, or firms building in-house on cheap frontier models
- Well-funded startup rivals (Truewind, Digits, Puzzle, Numeric, Rillet, Zeni, plus Europe's Pennylane) compressing pricing and land-grab
- Regulatory / professional-standards scrutiny (AICPA, PCAOB, IRS) of AI in attest and tax work; a high-profile agent error could chill adoption
Moats, dependencies & bottlenecks
Moats
evals, firm-specific workflows) Moderate-Strong Hard for horizontal AI tools to replicate; erodes if foundation labs commoditize vertical agents.
Deep integration into a firm's workflow is sticky, but early-stage and reversible if accuracy slips.
GV, Khosla, Blankfein, Friedman, Rabois) Accelerates access and capital; not a product moat.
If Basis accrues labeled outcome data across firms, evals/accuracy compound — but data-use terms with firms are undisclosed.
Dependencies
Core technology / COGS Buys frontier models rather than owning them — exposed to pricing, rate limits, and these labs entering vertical agents.
Revenue concentration / channel Narrow, conservative buyer set; adoption gates on trust and professional-standards comfort.
Rules on AI in attest/tax work could constrain how autonomous agents may operate.
Pre-profit; future scale depends on continued access to growth capital at supportive valuations.
Advantages
- First-mover with true autonomous (not copilot) agents in a high-value vertical
- Proven ability to land the hardest enterprise logos (top-25 firms)
- Capital-rich ($138M raised) and well-networked for the next phase
- Focus — a single vertical lets it out-specialize horizontal AI and out-modernize legacy incumbents
Weaknesses
- Undisclosed revenue leaves valuation unanchored to fundamentals
- No control over its core model layer (margin + roadmap risk)
- Young org selling into slow, risk-averse buyers
- Intensifying competition from both incumbents-with-distribution and funded startups
Bottlenecks
- Accuracy/reliability threshold for regulated deliverables — the last few percent of autonomy is the hardest and the most liability-laden
- Model inference cost per complex engagement (e.g. multi-state partnership returns) vs price the firm will pay
- Enterprise sales cycles and change-management inside conservative, partner-owned firms
- Talent — scarce engineers who understand both frontier ML and professional accounting/audit standards
Top signals & trends
Top signals
Tier-1 investors underwriting the category leader; unicorn status <3 years from founding.
Real enterprise adoption + quantified ROI, though self-reported.
Valuation rests on adoption narrative; monetization unproven publicly.
Distribution-advantaged competitors closing the capability gap.
Operator credibility + enterprise access signal.
Trends
Basis is directly levered to this shift; core thesis of the round.
Firms actively seek automation to offset scarce/expensive talent.
Expands automatable scope, but also lowers barriers for rivals and firms to build in-house.
Could slow adoption or impose guardrails on autonomy.
More capital and competition; likely M&A as the field consolidates.
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
Foundation-model provider — Basis buys frontier models from leading labs; OpenAI widely cited as a core model source.
Foundation-model provider (Claude); part of the 'leading foundation labs' Basis relies on for agent reasoning.
Potential model + cloud infrastructure supplier; GV is also an investor.
Compute and hosting layer for agent orchestration.
Large/national CPA firms — the beachhead segment; named logos include UHY, Boulay, Clark Nuber, MarksNelson, Pinion.
Private partnerships; the ultimate labor pool Basis-style automation targets (relationship status not individually disclosed).
Expansion tier where the CPA-shortage pain is acute.
QuickBooks/TurboTax distribution + native AI ('Intuit Assist'); largest incumbent threat, though augmentation-focused rather than fully autonomous.
Owns core tax/audit software (UltraTax, Checkpoint) used by CPA firms; embedding AI agents into that installed base.
CCH Axcess tax/audit suite with AI features; distribution-advantaged incumbent competing for the same firm workflows.
Sage Intacct + 'Sage Copilot' AI in accounting; mid-market incumbent (also LSE: SGE).
ERP incumbents adding AI agents; overlap more on corporate finance than CPA-firm attest work.
AP/AR + spend automation adjacent to bookkeeping workflows; potential agent expansion.
Private AI-native accounting startups attacking bookkeeping, close, and controllership; nearest venture peers.
Well-funded French AI-accounting player (raised a large 2025 round at multi-billion-euro valuation); geographic + product overlap as the market globalizes.