Salesforce Wants 32% of Every US FDE. It Just Fired 211 in SF.

Salesforce ran a public fire/hire contradiction this week. The Oct 5, 2026 Cal-WARN separations took effect at 415 Mission Street while the company kept publicly committing to a 1,000-person Agentforce Forward Deployed Engineer org. If you source for an AI-native competitor, the names on those filings are the shortest path to a shippable FDE pod you will see this quarter.

What actually happened on Oct 5, 2026

Salesforce's Oct 5 separations covered 74 San Francisco jobs and 59 Seattle/Bellevue jobs, bringing cumulative 2026 Cal-WARN headcount to 211 in SF across four rounds and total Salesforce WARN exposure to 1,048 workers nationally. The California filing was lodged Aug. 5, 2026, naming Salesforce Tower at 415 Mission Street as the worksite, and the parallel Washington notice listed 59 permanent terminations at Bellevue and Seattle offices including Tableau staff.

The function mix is what makes this cohort interesting rather than generic:

• 37 of the 74 SF cuts (50%) sit in technology and product.
• 34 sit in general administration.
• 3 sit in sales and distribution.
• 59 WA roles are permanent, not relocations or contracting-out, per the Washington filing.

Neither filing mentions AI. That silence is a legal choice, not an oversight. California's SB 951 disclosure rule is in force, so omitting AI from a WARN notice means the cohort is not legally classified as "replaced by AI". For a competing recruiter writing a cold outbound, that is a cleaner narrative than it looks.

211

The 1,000-FDE target is mathematically impossible from the open market

Salesforce's publicly stated 1,000-FDE Agentforce target represents roughly 32% of the entire existing US Forward Deployed Engineer labor pool. In Refolk's index of professional profiles, there are 3,102 Forward Deployed Engineers in the United States total, across every employer that has ever used the title. One in three of them would have to change jobs for Salesforce to hit its number without manufacturing new FDEs out of adjacent roles.

That math forces three conclusions sourcers can act on immediately:

1. The FDE title is being manufactured, not recruited. Salesforce is re-badging solutions engineers, forward-leaning consultants, and ex-SI delivery leads.
2. The ~37 SF tech/product cuts are the single most convertible candidate pool in the country, because they already know the Agentforce stack cold.
3. Any AI-native competitor that pipelines the Oct 5 list before Palantir's recruiters do gets a 60-day head start before Salesforce internal mobility re-slots the same people into Agentforce pods.

Forward Deployed Engineer is a role Palantir invented: a software engineer who embeds with one customer for a short engagement (Salesforce's Agentforce pods run ~3 months), ships production code against that customer's data, then rotates. Indeed/FT found FDE job postings grew more than 800% between January and September 2025. Salesforce's own program launched April 2025 and tripled inside six months, now hiring on six continents with sub-specialties in Agentforce Orchestration, Data 360, and Missionforce public sector.

The numbers, in one table

Here is the full arbitrage picture in a single view. Every row is traceable to a WARN filing, a Salesforce posting, or Refolk's index of professional profiles.

Salesforce is firing the only people in the country who already know how to ship Agentforce.

Why Tableau in Bellevue is the richest sub-pool, not the SF 74

The 59 Seattle/Bellevue cuts are a scarcer, more extractable pool than the SF 74 because Tableau talent is geographically concentrated away from the Pacific Northwest. Refolk's index returns 1,477 Tableau-skilled data engineers, analysts, and BI leads nationally, and the #1 region for that skill is the San Francisco Bay Area, not Seattle. Goldman Sachs is the top current employer with 3 named profiles.

A Bellevue-filed Salesforce FDE job explicitly lists required familiarity with a near-complete map of the competitive stack:

• Pega, UiPath, Automation Anywhere, ServiceNow, Appian
• Microsoft Copilot Studio, Google Vertex AI, AWS Bedrock
• IBM watsonx, Palantir AIP, ServiceNow AI Agents

That is the exact competitor shortlist a sourcer should mirror when outbounding the laid-off cohort. If a Tableau engineer in Bellevue was deep enough in the stack to be considered for Agentforce internal mobility, they will match every one of those bullets. The regional absorbers for this pool are Snowflake, Microsoft Fabric, Databricks field engineering, and the Tableau alumni network at smaller BI startups. Realistically, a WA-based laid-off Tableau engineer gets 2 to 3 competing offers inside 14 days.

This is the exact gap Refolk closes for the sourcer running the clock down: you describe the person in plain English ("Tableau engineers at Salesforce Bellevue, laid off Oct 5, familiar with Snowflake or Databricks") and get a ranked shortlist back across GitHub, LinkedIn, and the open web, instead of building a 40-filter Recruiter search that still misses the Tableau-to-BI-startup migrants.

Who is going to outbid you for these people

The structural winner here is Palantir, not Salesforce. Palantir Technologies is the #1 current employer of US FDEs in Refolk's index, and its recruiting team invented the discipline Salesforce is now trying to copy. The top ten US FDE employers Refolk sees also includes OpenAI's deployment arm (The OpenAI Deployment Company), Sixfold, Gecko Robotics, Roboflow, and Cresta. Every one of these is an active, direct competitor for the engineers Salesforce is laying off and trying to rehire into Agentforce.

Where the laid-off cohort will actually land, ranked by likely speed:

• Palantir sweeps the top of the SF 37 within 14 days. They have the brand, the band, and the muscle memory for this exact resume.
• OpenAI's deployment arm picks off anyone with LLM orchestration experience, offering IC comp that beats Salesforce's $316,750 ceiling.
• Concentrix and other Salesforce SI partners post Forward Deployed Engineer, Salesforce Agentforce roles at $92K to $140K, which is the floor of the arbitrage. Realign LLC in McLean posted an Agentforce FDE for consumer lending at a flat $130,000.
• Snowflake, Databricks, and Microsoft Fabric field teams absorb the Tableau-leaning Bellevue cohort first.
• AI-native Series B startups (Cresta, Sixfold, Gecko Robotics, Roboflow) take the long tail, usually at base-plus-meaningful-equity.

The salary arbitrage is counterintuitive. Salesforce's own FDE band tops out at $287,910 nationally and $316,750 in California, New York, and Seattle. The laid-off SF tech and product staff were almost certainly below that band in their old roles. So the move for an AI-native startup is not "cheaper than Salesforce FDE comp". It is "re-badge an ex-Salesforce engineer as an FDE, pay them 20% more than their old Salesforce base, and still undercut Salesforce's new-hire FDE band by $60K".

A seven-day sourcing plan for the Oct 5 cohort

Treat this as a 14-day window with a sharp decay curve. By Oct 19, Palantir and OpenAI will have swept the top of the list, and the remaining pool will be the tail that Salesforce internal mobility is also chasing.

Day 1 to 2: Fix the geo and worksite filters

Pull every current employee at the two named worksites, Salesforce Tower (415 Mission Street, SF) and Salesforce's Bellevue and Seattle offices, who shows up with a Technology, Product, or Data function. The SF filter should return somewhere near the 37-person tech/product share of the Oct 5 cut, plus the shadow pool of people who already left in Q3 2026 ahead of the WARN. LinkedIn Recruiter will undercount the shadow pool because departure dates lag; a cross-source index catches it.

Day 3 to 5: Enrich with stack signals

The Oct 5 cohort splits cleanly on two axes:

• Core platform (Apex, LWC, Flow, Data 360): convert directly to Agentforce FDE pitches at competitors or SI partners.
• Tableau and data (Tableau, Snowflake, dbt): route to Snowflake, Databricks, and Microsoft Fabric field teams.

Cross-source enrichment matters here because GitHub repo activity, open-web conference talks, and LinkedIn skills tags will disagree, and the engineer who looks like a Flow admin on LinkedIn often ships Rust on GitHub.

Day 6 to 7: Outbound with the right narrative

Lead with the SB 951 point. The cohort is not legally classified as "replaced by AI", and most of them will have been told the cut was budget-driven. Your pitch is not "come get revenge on Salesforce". It is "the exact Agentforce skills you spent 2026 building are worth 20% more at a Series B that ships to customers in two-week cycles instead of three-month pods".

32%

The shadow pool WARN notices will miss

The named cohort on the Oct 5 filings is the floor, not the ceiling. Salesforce has cut 211 San Francisco workers across four 2026 Cal-WARN rounds and 1,048 nationally, but WARN only names worksites above a size threshold. Everything below that threshold, plus voluntary departures accelerated by the restructuring, is invisible to a WARN tracker.

Salesforce reported $80M in restructuring expenses alongside $11.1B Q1 FY27 revenue, so the quarterly churn below the WARN line is non-trivial. Those people do not show up on Warntracker and most of them have not updated their LinkedIn yet. Finding them requires reading GitHub pushes, speaking slot changes, and open-web signals rather than a single-source filter. Describing the cohort in one sentence to Refolk and getting the shadow pool back as a ranked shortlist is the shortest path from "I read the WARN notice" to "I have 40 names and three warm intros" in the same morning.

FAQ

How do I confirm which specific Salesforce workers were laid off on Oct 5?

Cal-WARN and WA WARN filings name worksites and headcount, not individuals, so you work backward. Filter for current and recent-past employees whose profile location matches 415 Mission Street in San Francisco or the Bellevue and Seattle office addresses, in Technology, Product, or Data functions, with a tenure that puts their start before the April 2025 Agentforce launch. The intersection with Agentforce, Apex, LWC, or Tableau skills tags gets you very close to the actual 37 tech/product cuts and the Tableau subset of the WA 59.

Is it legal to source candidates based on a WARN notice?

Yes. WARN filings are public records published by state labor departments precisely so workers and employers can act on them. You cannot use the fact of a layoff to discriminate on a protected class, but reaching out to named worksites with a specific role pitch is standard practice. The sensitivity is tone, not legality: a cold message that references the layoff by name reads worse than one that references the engineer's actual Agentforce or Tableau work.

Why focus on the Tableau cuts in Bellevue over the SF tech cuts?

Scarcity. Refolk's index shows 1,477 Tableau-skilled profiles nationally with the San Francisco Bay Area as the #1 region, so Seattle-area Tableau talent is already thin. The 59 WA cuts represent roughly 4% of the national Tableau pool concentrated in a market that Snowflake, Databricks, and Microsoft Fabric are all actively recruiting. The SF 37 is a larger absolute number but faces heavier competition from Palantir and OpenAI, who are already camped on the Mission Street cohort.

How fast does this window close?

Two weeks is the honest estimate for the top of the list. Palantir, OpenAI's deployment arm, and Salesforce internal mobility will all be running the same filters inside the first 72 hours after Oct 5. The SI-partner floor at Concentrix and Realign stays open longer because their comp band ($92K to $140K) does not compete with where ex-Salesforce engineers want to land. If you are building an Agentforce-adjacent FDE pod at a Series B, your realistic window to close the top three names on this list is Oct 5 through Oct 19.