The number
149%. That's how much Palantir's US commercial revenue grew year-over-year in Q2 2026 — to $764M, inside total revenue of $1.935B (up 93%) and a record $2.13B of US commercial contract value booked in a single quarter. On the call, president Shyam Sankar named the reason out loud: Palantir's forward-deployed engineers and its ontology are the "determinative advantage" competitors can't copy. Read that against everything we've tracked this summer — OpenAI's $4B deployment spin-out, Google's in-house corps, Ode with Anthropic — and the significance lands: the company that invented the FDE just put a public P&L behind the thesis. Deployment-led revenue isn't a pitch anymore. It's compounding at 149% a year, audited.
data: Palantir Q2 2026 results (Business Wire, Aug 2, 2026); earnings-call remarks per company transcript.
What changed this week
The Plank census is still frozen — observed: 2026-07-04, count 1,206, now nine-plus weeks static — so the signal again is org-chart shape, not headcount. Two moves stood out from this week's frontier-lab and Glassdoor spot-check:
OpenAI is verticalizing FDE. The careers board now splits the role by industry: alongside the generalist Enterprise track we already list ($350K–$550K TC), OpenAI is hiring a Forward Deployed Engineer, Life Sciences and a Forward Deployed Engineer, Semiconductor — both SF — plus dedicated FDE reqs in Tokyo, Singapore, Seattle, and NYC, and a standalone Recruiter, Forward Deployed Engineering. When a function grows its own vertical specialists, four new geos, and its own hiring pipeline in one quarter, it has left "our first FDE" and entered "build the org."
FDE became a permanent federal-and-enterprise org chart. Three separate tells this week: Databricks stood up a clearance-gated US Federal AI-FDE track, distinct from its commercial FDE org; AWS posted a Forward Deployed Engineering Manager inside a named "AWS Forward Deployed Engineering" org spanning IC → Senior → Manager → Sr-Manager, with a US-Government vertical behind TS/SCI-with-poly; and T-Mobile — a carrier, a buyer rather than a vendor — is building an in-house FDE ladder to embed engineers in its own business units. The biggest cloud added a management layer; a telco crossed the vendor-to-buyer line. Both say the same thing: this is now a standing function, not an experiment.
New roles
Databricks — Sr. AI Engineer, FDE (U.S. Federal Sector) Washington, DC / MD / VA (hybrid, ~20–25% travel) · $200K–$300K (employer-provided) Databricks already runs a broad commercial FDE bench; now it's carving out a clearance-gated federal track (US citizenship + Secret eligibility). The federal FDE land-grab — AWS, Scale, Cognition, now Databricks — is where the least-transparent, highest-margin deployment work is quietly concentrating. View role →
Amazon Web Services — Forward Deployed Engineering Manager, FDE US Government Arlington, VA · comp not disclosed (US citizenship + active TS/SCI with Polygraph) The management title is the whole story. AWS has a named Forward Deployed Engineering org deep enough to post IC, Senior, Manager, and Sr-Manager reqs, plus a US-Gov vertical gated behind TS/SCI-with-poly. When the largest cloud builds a management layer for FDEs, the function is permanent. View role →
Kaizen Labs — Forward Deployed Software Engineer, Cleared (TS/SCI) Washington, D.C. / New York (hybrid) · $220K–$280K base + equity A rare cleared FDE seat that isn't Palantir. You build on Foundry, AIP, and Maven Smart System as the integration muscle, inside classification constraints rather than around them. The TS/SCI + citizenship gate is exactly why base opens at $220K–$280K — the clearance is scarcer than the React/Node stack. NEA/a16z/Accel-backed govtech, $35M raised. View role →
Doppel — Forward Deployed Engineer San Francisco / Remote (US) · undisclosed (Glassdoor pegs the lead track ~$100K–$200K) Proof the model has jumped past AI labs and data platforms into cybersecurity. Doppel (a16z/Bessemer, Series C, "social-engineering defense," scanning 150B+ entities a day) wants FDEs fluent in OAuth/SAML/SCIM and SIEM/SOAR wiring — not just LLM glue. The highest-leverage FDE seats increasingly reward deep integration plumbing. View role →
Casca — Forward Deployed Engineer San Francisco (on-site) · $100K–$200K One of the rare FDE reqs explicitly open to strong new grads — the band spans the whole early-career-to-senior spectrum. Casca (YC, "AGI for banking") embeds you with banks to deploy an AI-native loan-origination system against 30-year-old legacy cores. Regulated fintech, agentic product, ground floor. View role →
BCG X — Forward Deployed AI Engineer, UAE Dubai, UAE · comp not disclosed The title has fully crossed into Big-3 strategy consulting. BCG X ran a global Forward Deployed AI Engineer wave across ~8 geos this summer; by mid-August every one had closed except this Gulf seat. That the Dubai req is the one still open is its own signal — BCG is planting FDE capacity where sovereign-AI budgets are. View role →
One insight: what a clearance is actually worth on an FDE offer
Three of this week's six new roles are clearance-gated, so it's worth pricing the trade. The federal FDE tier runs a completely different comp philosophy from the commercial one, and the numbers show it. Databricks' federal track posts $200K–$300K employer-provided; Kaizen Labs' cleared seat opens at $220K–$280K base plus equity; Cape's federal FDE lists $175K–$235K; AWS's US-Gov role won't even disclose behind TS/SCI-with-poly.
Compare that to the commercial frontier tier, where FDE total comp runs roughly $385K at mid-level to north of $1.2M at principal — but 60–70% of that is private-company equity you can't sell until a secondary tender or an IPO that may be years out (per Perspective AI's 1,200-FDE comp report, May 2026). The federal trade is the inverse: you give up the equity lottery ticket and take cash-heavy, low-volatility base. A cleared $220K–$280K base is money that doesn't depend on a 409A re-rate.
The clearance itself is the moat. A TS/SCI is slower to obtain and scarcer than any tech stack, so it commands a premium on base and near-total job security — but it comes with geographic lock (the DC metro), citizenship gates, and a hard ceiling. You won't clear $1M inside a SCIF. Read federal FDE as the bond allocation in an asset class otherwise built entirely of equity: lower upside, far lower variance, and — right now — the fastest-growing corner of the market.
data: role bands from fde.directory listings (Databricks, Kaizen Labs, Cape, AWS); commercial-tier and equity-share figures from Perspective AI, "2026 Forward Deployed Engineering Compensation Report" (May 21, 2026).
One ask
Reply with your FDE comp — base, equity, and whether a clearance is in the mix — for the salary project. Cleared or commercial, every band you send makes the next issue's analysis sharper. Anonymous by default; I only ever publish ranges.
We’re building the FDE salary dataset — hit reply with your comp (band + tier is enough, kept anonymous).