What forward deployed engineering costs from a US firm

None of the fifteen firms in the list of FDE firms published a price on the pages read on 20 October 2026. So every quote for forward deployed engineering will be custom. This page gathers the public figures that do exist, says what each one measures, and explains how to compare quotes. For how a firm should set its own price, the FDE guide's page on how to price forward deployed engineering covers the seller's side.

What the firms publish about price

Firm	Price published	How it says it charges
Reveneau	No	Uses AI in place of hiring more engineers and puts that saving into clients' prices, by its own account.
Fluxon	No	Not stated on the pages read
Distyl AI	No	Not stated
Ode with Anthropic	No	Not stated
Tribe AI	No	Not stated
Factored	No	Not stated
Grid Dynamics	No	Not stated
Turing	No	Not stated
Tredence	No	Not stated
Percepta	No	Not stated
Tenex	No	By story points that it defines [3]
10Pearls	No	Not stated
Caylent	No	Outcome-based pricing and fixed-fee terms where results can be measured [4]
Artium	No	Not stated
HatchWorks AI	No	Not stated

The two published pricing methods

Tenex says on its AI engineering page that it charges clients for what it produces, measured in story points that it defines [3]. A story point is a unit that software teams use to size a task by the effort it needs. A small change might be one point, and a large feature many points. Under this method you pay for the size of the work delivered, and Tenex decides how many points each task is worth. Ask how points are assigned, who can challenge a count, and what one point costs.

Caylent's press release of 28 April 2026 says: "Engagements run on outcome-based pricing and fixed-fee terms where measurement is feasible" [4]. Outcome-based pricing means part or all of the fee depends on a result that both sides measure, such as a task that a system completes on its own. A fixed fee means one agreed price for an agreed scope of work. Both need a clear written measure. The FDE guide's page on outcome-based pricing for embedded engineering explains how to write one.

Reveneau's pricing mechanism

Reveneau says it uses AI in place of hiring more engineers, and that it puts that saving into the price its clients pay. It publishes no percentage or dollar figure for that saving. By its own account, AI writes 100 percent of its code, and every change has to pass a large eval suite, a set of automated tests written from the specification, before release. Ask Reveneau, like every firm, for a quote that shows the roles, the days and the price of each.

The public pay figures

Source	What it measures	Figure
Levels.fyi [1]	Total pay for forward deployed engineers at Palantir in the United States, from 226 user reports, updated 15 October 2026	$201K to $590K a year; median $233K
Tenex careers page [2]	Pay for Tenex's Forward Deployed Engineer role, full time and on site in New York City	$224K to $394K

These are pay figures. They show what engineers in the role earn at one software company and at one firm. A firm's rate to a client has to cover that pay and more, as the next section shows. Henley Wing Chiu of Bloomberry found that FDE job postings from January to October 2025 were 1,165 percent higher than in the same months of 2024 [8], so firms compete with many employers to hire these engineers.

What a firm's rate has to cover

A firm that places an engineer inside your team pays for more than the engineer's salary. Its rate also covers:

• benefits, payroll taxes and equipment for the engineer;
• recruiting and training;
• managers and senior staff who review the work;
• time between projects, when the engineer is paid and no client is billed;
• offices, sales and legal work;
• travel to your site, unless the quote lists it apart;
• the firm's profit.

So a firm's daily or monthly rate for one engineer will be higher than that engineer's pay for the same period. A quote that hides any of these items in one total is harder to compare.

What other public figures tell you

Amazon's $1 billion. Amazon said on 30 June 2026 that its AWS forward deployed engineering model is backed by a $1 billion investment [5]. The figure shows how much one large company is spending on its own FDE teams. Amazon's post states no price for the work.

Looker's need for $25,000 a year per customer. First Round Review quoted Frank Bien, Looker's former chief executive, saying Looker needed each customer to pay $25,000 a year, a figure he gave as approximate, and that checking the return gave him the confidence to invest in an FDE model [6]. That was the price of Looker's own software to its own customers. Bien's point was that Looker checked the return before it paid for the role.

The vendor's margin. Joe Schmidt of a16z wrote in June 2025 that software companies selling complex AI products accept lower margins, meaning a smaller share of each sale kept as profit, to pay for the staff who make the product work for each customer [7]. For a buyer, it means some FDE time may already be paid for inside a software vendor's licence price. Ask your vendor what it includes before you hire an outside firm for the same work.

The common ways firms charge, in plain words

Method	How you pay	What to check
Time and materials	A rate per hour, day or month for each person, plus costs such as travel	Who decides the hours, and how you see them each week
Monthly team fee	One price each month for an agreed team	What happens when a person on the team changes
Fixed fee	One price for an agreed scope	Who pays when the scope changes
Outcome-based	Part or all of the fee depends on a measured result	How the result is measured, and who measures it
Output-based, such as story points	A price per unit of delivered work	Who sizes each task, and how you can challenge it

Caylent names two of these, outcome-based and fixed fee. Tenex names the last. The other firms gave no method on the pages read.

What a quote should show

1. Each role and its rate, with the number of days or months.
2. Days at your site, and who pays for travel and hotels.
3. Platform licence costs, if the firm's own software is part of the system, for the engagement and for each year after.
4. AI model and hosting costs, as estimates, and who pays them. Hosting means the computers the system runs on.
5. The handover, with its own price and dates.
6. What would change the price, named in writing.
7. How the contract ends, and the notice each side must give.

How to compare quotes

Give every firm the same written description of the work, the same start date and the same length. Ask each one to price the same short paid task on your own data before the full contract. Add up the total cost over the whole length of the work, including licences and the handover, and compare those totals. A firm with a higher day rate and a shorter plan can cost less in total than a firm with a lower rate and a longer plan, so compare totals and also compare what each firm will deliver by each date.

Costs that continue after the engineers leave

When the engineers leave, some costs stay. Your company pays for the AI models the system calls, for hosting, for any platform licence, and for your own staff who run and change the system. Ask every firm to estimate these monthly costs in its quote. The FDE guide's page on the contract for an embedded engagement lists what the contract should say about them.

A short checklist for the price conversation

• Do you publish any price or rate card?
• How do you charge: by time, by team, by fixed fee, by outcome or by output?
• What does one engineer cost per month, and what is included?
• What will running the system cost each month after you leave?
• What will the handover cost?

The questions to ask an FDE firm before you sign page includes these with the rest of the checklist. For a quote from Reveneau, contact Reveneau.