TL;DR: ten US companies raised $656M between July 2025 and July 2026 to automate some part of hiring, and the money is not spread evenly. Mercor alone took $350M and is now reportedly negotiating at a $20B valuation on more than $2B of annualised revenue. Below it sits a much younger cohort (Juicebox, Paraform, Contrario, Alex, HeyMilo) where the real signal is revenue velocity, not round size.
Recruiting has become one of the busiest corners of applied AI, and July 2026 is a good moment to take stock. The category has produced a decacorn, a handful of companies compounding revenue faster than most enterprise SaaS ever has, and a long tail of seed-stage entrants who reached a few million in revenue before they bothered to announce themselves.
Key takeaways
- Ten US AI recruiting startups raised $656M between July 2025 and July 2026; Mercor took more than half of it.
- Revenue velocity, not round size, separates the field. micro1 grew 2.4x in four months on a $35M Series A.
- The category splits into five distinct products; buying the wrong one is the most common failure mode.
- Vendors are moving from licensed tools toward delivered hiring outcomes, and pricing is following.
- Automated screening compliance (Local Law 144, HB 3773, the EU AI Act) is the employer’s liability, not the vendor’s.
The 10 fastest-growing AI recruiting startups in the USA
1. Mercor:
Mercor matches domain experts (physicians, lawyers, quantitative traders, chess masters, senior engineers) to the AI labs that need them to train and evaluate frontier models. It is recruiting in the strict sense: source, screen, contract, pay. What makes it unusual is the buyer. Mercor’s customers are not HR departments, they are research organisations with effectively unlimited demand for expert human judgement.
The numbers are the story. Mercor raised a $350M Series C at a $10B valuation in October 2025, a five-fold jump from its Series B eight months earlier. In July 2026 Bloomberg reported it was in talks at $20B, with founder Brendan Foody stating publicly that annualised revenue had passed $2B, double where it stood four months earlier. Its own homepage now shows more than 300,000 roles created and daily payouts above $3.7M.
The caveats are real and worth stating: Mercor disclosed a data breach in April 2026 and has faced lawsuits from contract workers over classification and pay. Neither has visibly slowed the business, but a company scaling a contingent workforce this quickly carries employment-compliance exposure that a software vendor does not.
2. micro1:
micro1 competes with Mercor for the same AI-lab budget, but its recruiting layer is the differentiator rather than a means to an end. Zara, its AI recruiter, runs the 20-to-30-minute technical interview that every expert on the platform has to pass, and micro1 sells that screening capability to outside employers too. Deel has said it cut recruiting costs by more than 80% in a month using it.
micro1 raised $35M at a $500M valuation in September 2025, taking total funding to roughly $41.6M. Sacra estimates the company hit about $300M in annualised revenue by April 2026, up from around $125M at the end of 2025, a 2.4x move in four months on a fraction of Mercor’s capital. On capital efficiency it is arguably the strongest business on this list.
3. Contrario:
Contrario came out of stealth in May 2026 having already crossed $6M in annualised revenue and paid out more than $1M to recruiters, all inside six months. Founded by Stanford dropouts Arya Marwaha and Aditya Sood, it pairs specialist human recruiters with AI agents that absorb the operational load of scheduling, follow-ups and pipeline hygiene, and it raised a $2.3M seed led by Nexus Venture Partners on the back of that traction. Management has said it expects to pass $25M annualised by the end of 2026.
Its explicit thesis is that the agents should make recruiters better rather than replace them, which is the opposite bet to the pure AI-interviewer companies further down this list. On current evidence both bets are working, which suggests the market is splitting rather than converging.
4. Juicebox:
Juicebox is the clearest example of the sourcing thesis. Its PeopleGPT engine takes a plain-English description of who you want and returns matched candidates from a very large index, with CRM and outreach agents layered on top. Teams at Ramp, Perplexity, Notion, Cursor and Linear use it, and customers report up to 90% less time spent identifying shortlists.
In March 2026 Juicebox announced an $80M Series B at an $850M valuation led by DST Global, with Sequoia, Coatue and Y Combinator participating, bringing total capital to $116M. ARR has tripled since the July 2025 Series A, the platform now serves around 5,000 customers, and it has powered more than 560,000 searches against over 3 million candidates. If sourcing is your bottleneck, this is the category leader, and our roundup of AI sourcing tools for recruiters compares it against the rest of the field.
5. Paraform:
Paraform connects companies to thousands of specialist recruiters, then puts custom AI agents underneath them to analyse hiring data, candidate preferences and role requirements. More than 1,000 companies have hired through it, including Palantir, Rippling, Decagon, Abridge and Scale, and the platform reports that clients typically meet the candidate they eventually hire within about twelve days.
Its $40M Series B in March 2026, led by Scale Venture Partners, took total funding to $65M and drew in operators from Palantir, Canva, Stripe, Shopify and Uber. As of July 2026 Paraform’s own homepage banner claims the business has crossed $100M in annualised revenue. That is a company-stated figure rather than an audited one, but it is a striking claim from a company that was seed-stage two years ago.
6. Ashby:
Ashby is the only company here selling the hiring system of record (ATS, CRM, scheduling and analytics in one product), and it is growing at a rate that most point solutions would be pleased with. Its $50M Series D, led by Alkeon with Lachy Groom co-leading, landed after a year in which ARR rose 135% and the customer base went from 1,300 organisations to more than 2,700. Interviews scheduled through the platform rose 170%.
What is notable is the discipline behind it: Ashby reported a burn multiple under 1x and had barely touched its Series C when the Series D closed. The AI work (notetaking, talent rediscovery, an MCP integration so the ATS can be driven from external AI tools) is being layered onto a profitable-shaped core rather than funding the growth itself. Xero, Shopify, Ramp, Notion, Linear, Vanta, Snowflake, Replit and Zapier run on it.
7. Alex:
Alex, formerly Apriora, runs live, conversational first-round interviews and hands recruiters a structured, scored write-up. The pitch is not that AI hires better than a recruiter; it is that a recruiter’s judgement is wasted on the first screen, and that every applicant deserves an interview rather than a resume filter. The company has now run more than a million AI-led screening interviews.
Backed by Y Combinator and 1984 Ventures, Alex raised a $17M Series A in October 2025, taking total funding to $20M. It is the smallest raise of the interviewing group relative to volume delivered, which is a good sign. If you are weighing this class of tool, our guide to the top AI tools for candidate screening sets out what to test before you buy.
8. Humanly:
Humanly is the only company on this list built primarily for hourly, frontline and high-volume hiring, where the constraint is not finding candidates but responding to them fast enough. Its conversational AI engages, qualifies, interviews and schedules across chat, phone and video, and it carries a 4.8/5 rating on G2 across more than 100 reviews.
In May 2026 the Seattle company raised a $25M Series B led by SEEK Investments, taking total funding to $55.6M, and used it to reposition around what CEO Prem Kumar calls “service-as-a-software”: delivering pre-vetted, ready-to-hire candidates on demand rather than selling recruiters another tool. That shift, from software licence to delivered outcome, is the most interesting strategic move in this cohort.
9. Findem:
Findem argues that the constraint on hiring AI is not the model, it is the data underneath it, so it has built what it calls the largest expert-labelled talent dataset, and runs sourcing and workforce-planning agents on top. It closed a $51M Series C led by SLW in October 2025, taking total funding to $105M, after tripling year on year and placing in the top 10% of the Inc. 5000.
It has also been the most acquisitive company here, buying Glider AI to fold skills assessment into the same dataset. Of everyone on this list, Findem is the one most clearly building for the enterprise buyer rather than the fast-moving startup.
10. HeyMilo:
HeyMilo is the smallest raise on the list and the one with the most enterprise proof per dollar. The New York company screens and interviews applicants by voice, video, SMS and form, operating around the clock in more than 20 languages, and has run more than a million candidate screens in production for Randstad, WilsonHCG and Neo Financial among others.
In June 2026 it announced it had reached $6M in total funding, led by Category Ventures with Canaan Partners, Alumni Ventures and ERA participating. Landing staffing groups of that size on $6M of capital is unusual, and it is the reason HeyMilo makes this list ahead of better-funded competitors.
Twelve months of funding, in order
Laid out chronologically, the cadence is the point. Four rounds landed in the second half of 2025; six landed in the first half of 2026, and three of those came in a single ten-week window in the spring.
| Company | HQ | What it sells | Latest round | Growth signal |
|---|---|---|---|---|
| Mercor | San Francisco, CA | Expert supply for AI labs | $350M Series C, Oct 2025 | >$2B annualised, +100% in 4 months |
| micro1 | Palo Alto, CA | AI recruiter + expert supply | $35M Series A, Sep 2025 | ~$300M annualised, up from $125M |
| Contrario | San Francisco, CA | Recruiter marketplace + agents | $2.3M seed, May 2026 | $0 to $6M annualised in 6 months |
| Juicebox | San Francisco, CA | Sourcing and candidate search | $80M Series B, Mar 2026 | ARR tripled; ~5,000 customers |
| Paraform | San Francisco, CA | Recruiter marketplace + agents | $40M Series B, Mar 2026 | 1,000+ companies; ~12 days to hire |
| Ashby | San Francisco, CA | ATS, CRM and analytics | $50M Series D, Jul 2025 | ARR +135%; 1,300 to 2,700 customers |
| Alex | San Francisco, CA | Live AI interviewing | $17M Series A, Oct 2025 | 1M+ AI-led screening interviews |
| Humanly | Seattle, WA | High-volume conversational AI | $25M Series B, May 2026 | $55.6M raised; 4.8/5 on G2 |
| Findem | Redwood City, CA | Talent data and sourcing agents | $51M Series C, Oct 2025 | 3x year on year; acquired Glider AI |
| HeyMilo | New York, NY | Multi-channel AI screening | $6M total, Jun 2026 | 1M+ screens; Randstad, WilsonHCG |
These ten are not competing with each other
The most common mistake buyers make with this category is treating “AI recruiting” as one purchase. It is at least five different products sold to different budget owners, and a company that buys the wrong one concludes that AI recruiting does not work when what actually happened is that it solved a problem it did not have.
What the funding pattern says about where this goes next
Three things stand out from twelve months of rounds.
The money has moved from tools to outcomes. Humanly repositioning as “service-as-a-software”, Contrario and Paraform paying human recruiters inside an AI product, Mercor and micro1 selling vetted labour rather than software: the pattern is vendors taking responsibility for the result instead of licensing a capability. Seat-based recruiting software is the part of this market with the least momentum.
Capital efficiency is unusually high. micro1 reached roughly $300M annualised on about $41.6M raised. Contrario reached $6M on $2.3M. HeyMilo landed Randstad on $6M. Recruiting is one of the few AI categories where the workload is expensive enough per unit that inference costs disappear inside the gross margin, so companies here do not need to raise mega-rounds to scale.
The AI labs are the single biggest customer. The two largest businesses on this list sell to model developers, not to HR. That demand is real but concentrated, and if frontier-lab spending on human expert data slows, the top of this ranking is far more exposed than the bottom of it.
If you are hiring rather than investing
None of these ten solves the problem most growing companies actually have, which is not “we cannot find candidates” but “we cannot employ the ones we find”. A sourcing engine will hand you a shortlist in Manila, Ho Chi Minh City or Kraków in an afternoon; it will not tell you how to put those employees on a compliant contract, run their payroll, or take on the statutory obligations that come with them.
That is the gap Second Talent sits in. We combine human-vetted candidates with employment infrastructure in the markets where the staff actually are, so the hire lands as a properly employed team member rather than a name in a pipeline. If your bottleneck is the tooling itself, start with our comparison of the top AI recruiting tools and our guide to AI recruitment and staffing for US startups. If the bottleneck is the recruiting expertise itself, you can hire an AI recruiting specialist or an AI talent sourcing specialist through us directly.
Frequently asked questions
Which AI recruiting startup is growing fastest right now?
By absolute scale, Mercor: annualised revenue passed $2B in July 2026, double its March level. By rate on a small base, Contrario, which went from zero to $6M annualised in under six months. By the balance of the two, micro1: roughly 2.4x in four months on about $41.6M of total funding.
Are any of these companies profitable?
None has claimed profitability outright. Ashby comes closest to disclosing the shape of one, reporting a burn multiple under 1x and saying most of its Series C was still unspent when the Series D closed. Marketplace models such as Mercor, Paraform and Contrario carry a very different cost structure from the software vendors and should not be compared on the same margin basis.
Why are Mercor and micro1 on a recruiting list at all?
Because sourcing, screening, contracting and paying specialists at scale is recruiting, whoever the buyer is. Both run AI screening as the core of the operation. micro1’s Zara interviews every expert on its platform. The distinction worth keeping in mind is that their customers are AI labs, so their growth tracks frontier-model training budgets rather than corporate hiring demand.
Will AI recruiters replace human recruiters?
The fastest-growing companies here are not betting on it. Contrario, Paraform and Humanly all pay human recruiters inside an AI-native product, and Contrario’s public position is explicitly that agents should amplify recruiters rather than replace them. What is being automated is the repetitive middle of the funnel (sourcing, first-round screening, scheduling and follow-up), not judgement, closing or candidate relationships.
What does this software cost?
It varies by model: per-seat for the ATS and sourcing products, per-interview or per-screen for the interviewing tools, and a percentage of first-year salary for the marketplaces. Our AI recruitment tools cost comparison works through the pricing structures side by side, and AI tools to hire developers covers the engineering-specific stack.
Funding, valuation and revenue figures in this article are as reported by the companies themselves or by TechCrunch, Bloomberg, GeekWire, Business Wire, PR Newswire and Sacra as of 28 July 2026. Private-company revenue figures are unaudited.
