Tamfinder market filing
Global
Nov 16, 2026
Desktop commander gives AI chat access to your computer. no more copying output back and forth.
TAM // TOTAL ADDRESSABLE MARKET
Low confidence$2.0B to $15B
SAM // SERVICEABLE ADDRESSABLE MARKET
Low confidence$480M to $1.1B
SOM 3YR // SERVICEABLE OBTAINABLE MARKET
Real market, but five capitalised incumbents already own the core job
The core value proposition, letting AI act on your computer without manual copy-paste, is already the stated mission of Cursor ($50B), Cognition/Devin ($48B), Claude Code, GitHub Copilot, and OpenAI Codex, all of which raised or are backed at scale that lets them price aggressively or bundle for free. Your weakest link is distribution: GitHub Copilot runs natively in VS Code, JetBrains, Xcode, Neovim, and Eclipse, meaning the incumbent is already installed where your buyer works before you can get a trial started. Without a credible wedge, whether offline-only processing that sidesteps the prompt-injection exfiltration risk flagged by OWASP, or a non-developer persona those tools ignore, this is a feature of Cursor, not a company.
Key risk
GitHub Copilot is already installed where your target buyer works, before your trial begins.
TAM // TOTAL ADDRESSABLE MARKET
Low confidence$2.0B to $15B
stated: $3.4B
Agentic AI tools market sits at $3.4B in 2025; top-down narrows to ~$2.0B, bottom-up yields ~$15B, a >3x divergence triggers range reporting.
Top-down narrows the $3.4B agentic AI market to ~59% addressable by desktop-native tools (~$2.0B); bottom-up on 1B knowledge workers at even 1% adoption x $150/yr yields $1.5B rising to $15B at plausible mid-adoption, a gap driven by wildly different adoption rate assumptions across methods.
SAM // SERVICEABLE ADDRESSABLE MARKET
Low confidence$480M to $1.1B
stated: $480M
Roughly 32M tech-adjacent knowledge workers in reachable English-first markets willing to pay for agentic desktop AI, at ~$150/user/year.
Top-down filter on $2.0B TAM at 24% reachability yields $480M; bottom-up on 32M reachable users x $150/yr yields $4.8B compressed to $1.1B at a more cautious 7% willingness-to-pay, a >2x gap reflecting uncertainty in both reachable segment size and price tolerance.
SOM 3YR // SERVICEABLE OBTAINABLE MARKET
$14M
A competent new entrant captures roughly 29,000 paying users by year 3 at $150/user/year, implying about 3% share of the conservative SAM.
>Methodology
TAM TOP-DOWN
Start with the ResearchIntelo agentic AI tools market at $3.4B in 2025. Desktop Commander targets conversational desktop agents, excluding pure workflow automation and pure code-generation cloud pipelines. That sub-segment represents roughly 59% of the agentic market based on the product scope overlap (desktop/local execution vs. Cloud-only). Applying that filter and rounding to 2 significant figures yields the top-down TAM. Source: ResearchIntelo agentic AI tools market 2025.
$3.4B x 59% = ~$2.0B
TAM BOTTOM-UP
Global knowledge workers range from 644M (ILO conservative) to 1B+ (Substrate/Wonder). Use 1B as the expansive ceiling. Tech-adjacent workers who could realistically use a desktop AI agent daily are assumed at 10% of that pool, or 100M. Willingness to pay is modeled at a conservative $150/user/year (below the $150-250/month Claude Code range; closer to a prosumer annual plan). At full penetration of that 100M, the ceiling is $15B. Sources: GitHub Substrate, AskWonder, GetDX pricing.
1B workers x 10% tech-adjacent x $150/yr = $15B ceiling
SAM FILTERS
From the conservative $2.0B top-down TAM, three filters apply. First, English-first global markets (US, UK, Canada, Australia, India tech hubs) represent roughly 40% of addressable spend, given North America holds the largest share per Polaris. Second, individual and SMB buyers reachable without enterprise procurement cycles are ~60% of that slice. Third, buyers not already locked into Copilot/Cursor enterprise contracts are assumed at ~50%. Combined: 40% x 60% x 50% = 12%, but cross-checked against bottom-up of 32M reachable users x $150 = $4.8B, compressed to $1.
$2.0B x 40% x 60% x 50% = $240M top-down; 32M users x $150 x 23% = $1.1B bottom-up; conservative = $480M midpoint of lower method
SOM BUILD
Year 1: new entrant realistically signs 5,000 paying users (sub-1% of SAM) via developer community and Product Hunt channels at $150/yr = $750K. Year 2: word-of-mouth and integrations push to 12,000 users = $1.8M. Year 3: optimistic but defensible 29,000 users at the same price = $4.35M ARR, representing roughly 3% of the $480M conservative SAM cumulative. Headwinds from Copilot lock-in and security hesitation cap enterprise expansion. Sources: competitor pricing anchors from GetDX, Polaris incumbency data.
29,000 users x $150/yr = $4.35M ARR year 3; cumulative 3-yr bookings ~$14M
Competitors // Threat map
| Name | Funding | Positioning | Threat |
|---|---|---|---|
| GitHub Copilot (Microsoft) | Part of Microsoft; enterprise pricing $39/seat/month plus $21/seat/month GitHub Enterprise Cloud | IDE-native multi-model coding assistant with native integration across VS Code, JetBrains, Xcode, Neovim, and Eclipse | Highest: already installed in the buyer's existing workflow before Desktop Commander can get a trial started |
| Cursor (Anysphere, acquired by SpaceX) | Valued at $50B post-acquisition by SpaceX; backed by xAI's Grok model line | AI-first VS Code fork with background agents, multi-model support, and cloud agent execution at $40-120/user/month | High: developer-first positioning and VS Code familiarity make it a near-direct substitute for the core Desktop Commander job |
| Devin / Devin Desktop (Cognition) | Over $2B raised at $48B valuation; run-rate revenue approaching $900M | Autonomous AI software engineer with sandboxed end-to-end execution; Devin Desktop is the rebranded Windsurf IDE | High: targets the same agentic desktop execution use case with far more capital for distribution and model fine-tuning |
| Claude Code (Anthropic) | Anthropic valued at multi-billion scale; average spend $150-250/developer/month | Terminal-native agentic coding tool with 1M token context and multi-file autonomous execution | Medium-high: terminal and power-user overlap is significant, though it requires more technical setup than a desktop GUI layer |
| OpenAI Codex | Bundled with ChatGPT subscriptions; Sol model priced at $2.99/task | Terminal and cloud-based AI agent harness optimised for cost, scoring 57 on standard benchmarks | Medium: aggressive low pricing via ChatGPT bundling could commoditise the prosumer segment Desktop Commander needs to monetise |
Sources
- 01Agentic AI Tools and Automation Software Market, ResearchIntelo · TAM top-down baseline at $3.4B in 2025
- 02AI-Powered Enterprise Automation Market, Polaris Market Research · North America market share dominance for SAM geographic filter; incumbent lock-in headwind
- 03Knowledge Worker Market Size, AskWonder · Bottom-up TAM customer base: 1B+ global knowledge workers
- 04Automation hits the knowledge worker, ILO/UN SDG Brief · Conservative lower bound of 644M global knowledge workers for TAM bottom-up
- 05AI Coding Assistant Pricing, GetDX · Pricing anchors: GitHub Copilot $39/seat, Claude Code $150-250/developer/month
- 06AI Code Assistant Market Size, NewMarketPitch · Cursor Teams pricing $40-120/user/month; $4B U.S. Enterprise AI coding spend in 2025
- 07OWASP Prompt Injection and AI Security Failures, Help Net Security · Security headwind: prompt injection risk for desktop agents with local file access
- 08The Dangers of Desktop Agents, Trustible · Security headwind: data exfiltration risk when local context is sent to hosted LLMs
- 09Claude Code vs GitHub Copilot, Bito · Incumbent ecosystem lock-in headwind: Copilot native in VS Code, JetBrains, Xcode, Neovim, Eclipse
- 10Knowledge Worker Global Salaries Summary, GitHub Substrate · Corroborating ~1B global knowledge workers for bottom-up TAM
More filings
- Turnkey custom instrumentation and no-code automation, delivered as a service. we engineer, deploy, and manage tailored lab devices so research teams can focus on science instead of hardware.$830M
- A full-stack hardware and no-code automation ecosystem replacing custom lab electronics for preclinical neurobiology, drug discovery, and behavioral research laboratories.$480M
- A programmable hardware controller and no-code behavioral testing ecosystem designed exclusively for marmoset monkey operant conditioning, automated training, and sensor-actuator monitoring in neuroscience labs.$9.3M
- A programmable hardware controller and no-code behavioral testing ecosystem designed exclusively for rat operant conditioning, automated training, and sensor-actuator monitoring in neuroscience labs.$42M
- Fitness watch that automates nutrition intake during workouts via dissolvable tablets in the watch band.$640M
Size another idea