# Is Surge AI Legit? Surge AI Review and vs Scale AI

> Is Surge AI legit? Yes: TechCrunch put its 2024 revenue at $1.2B, and it took no venture funding. How it pays experts, if it buys company data, vs Scale AI.

Canonical: https://origindatapartners.com/guides/surge-ai-review
By: Origin Data Partners
Status: Published
Published: 2026-11-10
Updated: 2026-11-10
Sources checked: 2026-11-10

## Quick answer

Yes, Surge AI is a legitimate AI training-data company: TechCrunch reported its 2024 revenue at $1.2 billion, it says it has never raised venture funding and names OpenAI, Anthropic, Meta and Google as partners. It pays experts as contractors. For companies, MIT Technology Review reports it buys data archives for RL gyms, but it publishes no seller program.

## Key points

- Surge AI is one of the largest AI data companies: TechCrunch cites reported 2024 revenue of $1.2 billion, and Surge says it has been profitable without venture funding.
- Experts work as contractors. Current listings range from $100 to $150+ an hour for software engineers to $200 to $400 an hour for some writers.
- Surge buys company archives to build RL gyms, per MIT Technology Review, but has no public intake. The archive market it describes runs mostly through wind-down services.

## Is Surge AI legit? The short answer

Yes. Surge AI is a large, profitable company that supplies human expertise, datasets and RL environments to the biggest AI labs. "Is Surge AI legit" searches come from experts deciding whether to take contract work, and from owners who read that Surge buys company archives. The answer differs for each, so they are covered separately.

For experts, Surge pays well by industry standards. For companies, Surge does buy data, but it publishes no way to sell to it directly, no price and no terms.


## What Surge AI is: size, funding and customers

Surge describes itself as "Human Intelligence for AGI" ([Surge AI](https://www.surgehq.ai/)). Its listings say "We've run a profitable business from day one without raising venture funding" and that it works "in partnership with companies like OpenAI, Anthropic, Meta, and Google" ([Surge AI](https://www.surgehq.ai/workforce)). TechCrunch, comparing data suppliers, cited "Surge's reported $1.2 billion in 2024" in revenue ([TechCrunch](https://techcrunch.com/2025/12/04/micro1-a-scale-ai-competitor-touts-crossing-100m-arr/)).

It sells three things. Expert work for labs. Ready-made datasets and RL environments, including "Realistic engineering tasks built from real company codebases", where some labs can "pay only when it moves the metrics that matter" ([Surge AI](https://www.surgehq.ai/ots)). And enterprise programs that "include supervised fine-tuning, reinforcement learning, preference optimization, and custom RL environments" ([Surge AI](https://www.surgehq.ai/enterprise)).

Source: [Surge AI: expert workforce listing](https://www.surgehq.ai/workforce)

## Surge AI review for contractors: pay and what to check

Surge hires experts as contractors. Its software-engineer listing pays "$100–$150+/hour" to build coding RL environments, noting that "Frontier labs train on the agentic coding RL environments you create" ([Surge AI](https://www.surgehq.ai/swe)). A writing listing pays "$200–400 / hour, contractor" ([Surge AI](https://www.surgehq.ai/workforce)). Rates vary by role and are set per listing.

Contractor work is set per listing: confirm the hourly rate, any unpaid onboarding or assessment time, and how time limits on tasks affect pay before you start. The listings describe contract roles, not employment, so plan for your own taxes, and keep a record of hours and approved tasks in case a payment needs to be queried later.

Source: [Surge AI: software engineer listing](https://www.surgehq.ai/swe)

## Does Surge AI buy company data?

Yes, according to MIT Technology Review: "Surge AI, for example, buys data to build "reinforcement-learning gyms" where clients' AI agents practice tasks". Nick Heiner, who leads Surge's environment work, "says he seeks data archives reflecting a range of situations, from disagreements to tricky judgment calls", and that "one company's Slack history, if it captures conversations with variety and depth, can be worth more than another's Slack, email, and calendars combined" ([MIT Technology Review](https://www.technologyreview.com/2026/10/13/1145755/ai-labs-buying-data-from-dead-startups-to-train-models/)).

What Surge does not have is a seller program. We found no form, price, payment terms or eligibility rules on its site. The archive market the magazine describes runs mostly through wind-down services such as SimpleClosure, which sell closed startups' code and workspace data to labs and to vendors that build simulated workspaces. It also notes that "Founders have little to no visibility into who is buying the data and what they plan to do with it."

If you are approached, ask the intermediary whether Surge is the buyer, the end user, or both, and whether the data could pass to a lab. Surge sells datasets and environments to labs, so ask whether a lab will receive your records or only environments built from them, and get the answer into the license.

On price, Heiner told the magazine that buyers quote very different numbers because valuations rest on "research theses" about how many useful tasks an archive yields. For a seller, that means there is no list price to compare against, and the second and third quotes matter.

If your company is closing, see [defunct company data](https://origindatapartners.com/guides/sell-defunct-company-data), [selling Slack and email archives](https://origindatapartners.com/guides/sell-slack-email-archives-ai-training) and our [SimpleClosure review](https://origindatapartners.com/guides/simpleclosure-review). If it is operating, the programs in [company data buyers compared](https://origindatapartners.com/guides/company-data-buyers-compared) publish terms Surge does not.

Source: [MIT Technology Review: AI labs are buying data from dead startups (Oct 2026)](https://www.technologyreview.com/2026/10/13/1145755/ai-labs-buying-data-from-dead-startups-to-train-models/)

## What Surge AI's products tell a seller

Surge's catalogue shows where company data ends up. Its coding datasets include "Realistic engineering tasks built from real company codebases", and its enterprise-agent datasets cover "office documents, enterprise workflows, and live external services" in long-horizon tool-use environments ([Surge AI](https://www.surgehq.ai/ots)). Those are built from the same kinds of records an operating company keeps: repositories, documents, tickets and the messages around them.

Its pricing to labs is also telling. Some labs in its Trusted Program can train first and pay only if the data improves their model. A supplier that sells on results needs material that reliably produces hard, realistic tasks, which is why its environment lead prizes archives with disagreement and judgment over sheer volume.

For a seller, that points to what to emphasise in any conversation: how many years the records cover, how many people and decisions they capture, and whether outcomes are recorded (a ticket closed, a deal won, a bug fixed). Volume alone does not set the price.

Source: [Surge AI: off-the-shelf datasets](https://www.surgehq.ai/ots)

## Surge AI vs Scale AI

Surge and Scale AI compete for the same lab budgets, and the main difference since 2025 is ownership. In June 2025 Scale announced "a significant new investment from Meta Platforms, Inc. (Nasdaq: META) that values Scale at over $29 billion"; "Scale's founder, Alexandr Wang, is joining Meta", and Meta holds a minority of the equity ([Scale AI](https://scale.com/blog/scale-ai-announces-next-phase-of-company-evolution)). TechCrunch later reported that OpenAI and Google DeepMind "reportedly cut ties with Scale AI following Meta's $14 billion investment in the vendor" ([TechCrunch](https://techcrunch.com/2025/12/04/micro1-a-scale-ai-competitor-touts-crossing-100m-arr/)). Surge remains independent and self-funded.

| Question | Surge AI | Scale AI |
| --- | --- | --- |
| Outside funding | None, by its account; profitable from day one | Meta investment valuing it at over $29 billion (June 2025) |
| Ownership | Independent | Meta holds a minority stake; founder joined Meta |
| Lab relationships | Names OpenAI, Anthropic, Meta and Google as partners | OpenAI and Google DeepMind reportedly cut ties after the Meta deal |
| Size signal | $1.2 billion 2024 revenue, as reported by TechCrunch | Not compared here |
| Buying company data | Buys archives for RL gyms; no public seller program | See [Mercor competitors](https://origindatapartners.com/guides/mercor-competitors) for its data programs |

Source: [Scale AI: Meta investment and founder joins Meta (June 2025)](https://scale.com/blog/scale-ai-announces-next-phase-of-company-evolution)

## What to check before selling data to Surge AI or a reseller

Because Surge has no public intake, an owner usually meets it through an intermediary or an inbound approach. That changes the questions. With a published program you can read the terms first; with an unnamed or indirect buyer you have to ask for them, and you may not learn who ends up with the data unless the contract says so. These checks cover both routes.

- Ask who the end buyer is, by name, and whether the license lets them resell or share the data.
- Get the price, the payment trigger and the payment date in writing, not only a range.
- Ask which records are excluded (HR files, health information, customer secrets) and who de-identifies the rest. See [de-identifying company data before licensing](https://origindatapartners.com/guides/de-identify-company-data-before-licensing).
- Get at least two quotes. Buyers price on research theses, so numbers vary widely.
- Run the full list in [how to vet an AI data buyer](https://origindatapartners.com/guides/how-to-vet-an-ai-data-buyer) and compare buyers in our [directory of AI data companies](https://origindatapartners.com/guides/ai-data-companies).


## Next step

If you hold archives that a buyer like Surge would want, find out what they are worth to more than one buyer before you sign through anyone.

Not sure whether a buyer or an offer is sound? Origin checks who the buyer is, what the license allows and how the price compares with published deals before anything is signed. [Get a free estimate and a second opinion](https://origindatapartners.com/check-fit).

Origin Data Partners works with companies that own data and helps get it bought. We are independent of the services on this page, and a buyer may pay us when a deal closes ([how Origin is paid](https://origindatapartners.com/referral-disclosure)). [See what your data is worth](https://origindatapartners.com/check-fit).


## Questions and answers

### Is Surge AI legit or a scam?

Surge AI is legitimate, not a scam: TechCrunch reported its 2024 revenue at $1.2 billion, and Surge says it has been profitable without venture funding and names OpenAI, Anthropic, Meta and Google as partners.

### How much does Surge AI pay?

It depends on the role. Current listings show $100 to $150+ an hour for software engineers building coding environments and $200 to $400 an hour for some expert writers, all as contractors.

### What is the difference between Surge AI and Scale AI?

Surge AI is independent and self-funded, while Scale AI took a Meta investment in June 2025 that valued it at over $29 billion, and its founder joined Meta. OpenAI and Google DeepMind reportedly cut ties with Scale afterwards.

### Does Surge AI buy company data?

Yes, according to MIT Technology Review, which reported in October 2026 that Surge buys data archives to build reinforcement-learning gyms. Surge publishes no seller program, price or terms.

### Has Surge AI raised venture capital?

Surge says no: its listings state it has run a profitable business from day one without raising venture funding.

### How can a company sell data to Surge AI?

There is no public form. The archive sales MIT Technology Review describes run mostly through wind-down services, so ask any intermediary to name the end buyer and put the price and payment date in writing.

### What company data does Surge AI value?

Archives with variety and depth. Surge's environment lead told MIT Technology Review he looks for disagreements and tricky judgment calls, and that one rich Slack history can outweigh another company's Slack, email and calendars combined.

## Sources

- [Surge AI: homepage](https://www.surgehq.ai/)
- [Surge AI: expert workforce listing](https://www.surgehq.ai/workforce)
- [Surge AI: software engineer listing](https://www.surgehq.ai/swe)
- [Surge AI: off-the-shelf datasets](https://www.surgehq.ai/ots)
- [Surge AI: enterprise](https://www.surgehq.ai/enterprise)
- [MIT Technology Review: AI labs are buying data from dead startups (Oct 2026)](https://www.technologyreview.com/2026/10/13/1145755/ai-labs-buying-data-from-dead-startups-to-train-models/)
- [TechCrunch: micro1 touts crossing $100M ARR, with rivals' revenue (Dec 2025)](https://techcrunch.com/2025/12/04/micro1-a-scale-ai-competitor-touts-crossing-100m-arr/)
- [Scale AI: Meta investment and founder joins Meta (June 2025)](https://scale.com/blog/scale-ai-announces-next-phase-of-company-evolution)

## Related reading

- [Mercor Competitors: Who Else Pays Companies for Data (2026)](https://origindatapartners.com/guides/mercor-competitors)
- [micro1 vs Mercor vs Handshake vs Turing: company data buyers compared (2026)](https://origindatapartners.com/guides/company-data-buyers-compared)
- [Defunct company data: who buys it and on what terms](https://origindatapartners.com/guides/sell-defunct-company-data)
- [RL environment companies and deals tracker (2026): who builds them and who pays for company data](https://origindatapartners.com/guides/rl-environment-companies)
- [Scale AI Competitors: Companies Like Scale AI (2026)](https://origindatapartners.com/guides/scale-ai-competitors)