Frontier AI lab building System One Models — machine-native, typed decisions with calibrated confidence for software to act on, not chat.
TypeSafe AI is a San Francisco frontier AI lab, founded in 2024 by former OpenAI researcher Diogo Almeida with Erik Gafni and Sasha Sheng, that emerged from stealth on 2026-09-15 with about $40 million in seed funding led by DCVC — a round Forbes reports valued the company at $200 million, citing a person familiar with the deal. Rather than generating text, its System One Models return typed decisions with calibrated probabilities and confidence scores that software can act on directly, so a business can automate outputs above a confidence threshold and route the rest to people. The first public model, Jev, is in waitlisted early access; the company claims sub-100-millisecond responses, published pricing of $42 per billion input tokens, and up to two orders of magnitude lower cost and latency than frontier LLMs on decision-shaped workloads, and it argues that RLHF-trained chat models are structurally overconfident and therefore need humans in the loop. Almeida, who worked on the instruction-following research behind ChatGPT, says most intelligence should eventually run quietly inside software; the company was part of the 2024 AWS Generative AI Accelerator before launch.
AI📍 San Francisco, CAFounded 2024
Current Valuation
$7.5B
as of 2026-10-09
UNICORN
QL Price
$7.5B
0% vs last round
QuantLogix's current estimate for TypeSafe AI, published 2026-10-11 — in line with the last disclosed round of $7.5B (2026-10-09).
Confidence
low
Basis
round anchored
What backs this price
Last disclosed round
$7.5B
2026-10-09
weight 1
How this price has held up
No round has priced for TypeSafe AI since QuantLogix began publishing a price for it, so there is nothing to grade yet. The claim above is on the record from 2026-10-11 and will be graded against the next priced round. See the full accuracy record →
Which publicly-traded funds already hold it, what $10,000 buys, and whether any route is open to you.
Cite this price as QuantLogix QL Price, TypeSafe AI, 2026-10-11 (146ded840c80313c).
The id pins this exact claim, so a citation stays checkable after the price moves.
Model estimate for research — not a quote, target price, or investment advice.
IPO Outlookprivate
Is TypeSafe AI going public?
No. TypeSafe AI is private and has not filed to go public. There is no S-1 on file.
What is TypeSafe AI's valuation in 2026?
TypeSafe AI is valued at $7.5B as of 10/09/2026, set by its Series A ($870M, Oct 2026). That is the last confirmed mark, not a live price — private shares do not trade continuously.
Has TypeSafe AI filed an S-1?
No public S-1 is on file for TypeSafe AI.
Who are TypeSafe AI's investors?
TypeSafe AI's investors include DCVC, Andreessen Horowitz, Sequoia Capital.
How much has TypeSafe AI raised?
TypeSafe AI has raised $910M in total across 2 disclosed rounds, most recently a Series A of $870M in Oct 2026.
Is TypeSafe AI publicly traded — can you buy TypeSafe AI stock?
No. TypeSafe AI is a private company — its shares are not listed on any exchange and there is no ticker. The routes to exposure are pre-IPO secondary marketplaces and funds (accredited investors), employee tender offers when the company runs one, and public companies with read-through exposure. QuantLogix maps every route it can verify for TypeSafe AI on its Pre-IPO Access Desk.
Answers are derived from the roster record shown on this page — filing status, last round and confirmed marks. Where the record is silent, the answer says so rather than guessing.
Company Profile
CEO
Diogo Almeida
Founded
2024
HQ
San Francisco, CA
Total Raised
$910M
Growth
Out of stealth 2026-09-15 with a $40M DCVC seed (reported $200M post); 24 days later announced an $870M a16z-led Series A at $7.5B (company post + Bloomberg). Company says Jev passed one million users within days of launch and is used by about a third of Fortune 500 companies (unnamed, unaudited). Pitch is typed, calibrated decisions for software at roughly two orders of magnitude lower latency and cost than frontier LLMs on decision-shaped tasks.
Last Round
Series A — $870M (Oct 2026) · Lead: Andreessen Horowitz
IPO Status
private
Fundraising Status
Series A · $870M · this month
privateTotal raised $910M2 disclosed rounds
Series A Impact
Round Size
$870M
Primary financing announced
Post-Money Valuation
$7.5B
Company-announced mark
Valuation Step-Up
+3650.0%
vs Seed at $200M
Round / Post-Money
11.6%
New capital ÷ post-money
Disclosed Funding
$910M
Cumulative disclosed rounds
Step-up compares disclosed post-money valuations. Round / post-money is a financing-size ratio, not an ownership or dilution estimate.
Growth & Valuation
Valuation across funding rounds
Capital raised per round
Charts built from publicly disclosed funding rounds and revenue estimates. Approximate, not real-time, not investment advice.
Funding History
Round
Amount
Date
Lead Investor
Post-Money
Series A
$870M
Oct 2026
Andreessen Horowitz
$7.5B
Seed
$40M
Sep 2026
DCVC
$200M
Vertical Intelligence — AI / ML
Open models
0
Downloads (30d)
0
across all models
Followers
89
Attributed to Hugging Face organization "TypeSafe AI" as of 2026-10-09. Open-weights footprint only — closed-model usage is not publicly measurable. Verify on Hugging Face ↗
Talent Signal
Hiring trend
● First snapshot
trend accrues with monthly sweeps
Open roles
6
via ashby
Product-build share
50%
engineering, research, product, design
Trajectory
steady
product-led hiring — 50% of open roles build product; the rest are go-to-market and G&A.
Actively hiring with no tracked executive churn.
Open roles by month
Open roles by function
■ product-building ■ go-to-market / G&A
Member of Technical Staff, Backend/PlatformDeveloper AdvocateMember of Technical Staff, Infrastructure (Kubernetes Specialist)Member of Technical Staff, Model CapabilitiesMember of StaffFounding Marketer
Counted from TypeSafe AI's live job board as of 2026-09-17; role functions classified from posting titles. Open-role counts are an activity signal, not headcount. Refreshed monthly.
Founders & Key People
Diogo AlmeidaErik GafniSasha Sheng
Investors
DCVC · Andreessen Horowitz · Sequoia Capital
Products
Jev — first public System One Model, in waitlisted early access since 2026-09-15: takes unstructured state plus a predefined schema of questions and returns typed decisions with calibrated probabilities and confidence scores, all outputs generated in parallel rather than token by token
System One Models — the model class, trained with the company's Reinforcement Learning for Calibrated Decisions (RLCD) instead of RLHF; outputs are schema-constrained so type errors are impossible and no free text is generated
Published pricing of $0.042 per million input tokens ($42 per billion) with output tokens free; company-measured end-to-end latency of 70–500 ms and, on its own workflow evals, up to 193.6x faster and 444.6x cheaper than frontier LLMs on System One tasks
Developer surface: console and playground at console.typesafe.ai, docs at docs.typesafe.ai, public workflow-evals site (evals.typesafe.ai) and an open-source System One adapter that constrains ordinary LLMs to the same decision API
Competitors
OpenAI · Anthropic · Google DeepMind · DeepSeek · Mistral AI
AIFoundation ModelsAI Research LabAutomationEnterprise AIDeveloper ToolsRecently FundedUnicorn
Private-company numbers are not real-time. Reflects publicly disclosed valuations from press releases, news reports, and tender offers as of 2026-10-09. Refreshed quarterly.