What's Really Driving the Cost to Originate a Mortgage in 2026

The cost to originate a mortgage hit $11,800 in Freddie Mac's most recent analysis of retail lender expenses, and independent mortgage companies are running even higher, averaging $12,209 in 2025 according to MBA's Peer Group Roundtable data, with sales expense alone accounting for 60% of that total. Whichever figure you use, the driver is the same: personnel cost, not technology or compliance overhead, is still the single biggest lever lenders and brokerages can pull right now.

Where Does the $11,800 Figure Actually Come From?

Freddie Mac's analysis of retail lender expenses put the average cost to produce a mortgage at $11,800, with lenders reporting pre-tax net income of just $900 per loan on top of that, a thin margin that leaves little room for inefficiency anywhere in the process. MBA's own Peer Group Roundtable data tells a consistent story from a different angle: independent mortgage companies in the retail channel averaged $12,209 per loan in 2025, and sales expense, largely loan officer compensation and commission structures, made up 60% of that figure. Depositories ran even higher, averaging $16,320, with sales expense at 42% and corporate and production support allocations accounting for another 38%.

Different methodologies, similar conclusion: across every major data source, personnel-related expense dominates the cost structure of originating a loan.

What's Actually Driving This Cost?

Two forces show up consistently. The first is sales expense itself, commission and compensation tied to origination volume, which scales with production but doesn't shrink cleanly when volume slows. The second is staffing elasticity, or the lack of it. Lenders who staff for one rate cycle often find themselves scrambling to hire and train for the next one when volume shifts, and per-loan costs have followed that same boom-bust staffing pattern since 2008. A lender able to absorb a real jump in application volume without a matching jump in payroll isn't just more efficient in the moment. It's more stable, because it isn't forced to lay off people it spent months training every time rates move against it.

The Industry's Answer Right Now: A Wave of AI-Native Bets

Origination costs pushing past $11,000 to $12,000 a loan has triggered real investment activity. A handful of AI-native startups are each targeting a different layer of the cost problem:

Copperlane is building an AI agent, "Penny," designed to act as an autonomous loan officer assistant, automating document review and borrower communication. The company raised a $4.1 million seed round earlier in 2026.

Pylon has taken the most aggressive infrastructure bet, building a vertically integrated, API-first "mortgage rails" platform that automates processing, underwriting, closing, and delivery. The company claims a cost to originate 74% lower than the industry average and has partnered with Loan Factory, a large broker platform, to originate loans directly on its infrastructure. Pylon is backed by investors including Peter Thiel, Citi, and QED.

Ralo has gone furthest structurally, launching as what it calls the first AI-native mortgage broker, aiming to eliminate loan officers, processors, and underwriters from the process entirely. It closed a $2.9 million seed round and, as of mid-2026, is licensed in three states: California, Colorado, and Texas.

Bevri.ai has taken a narrower, deployment-first approach, an agentic AI point-of-sale platform now live with a closed group of loan officers at NEXA Lending, automating the 1003 application, income and asset verification, and Desktop Underwriter and Loan Product Advisor findings.

What These Bets Have in Common, and What They Don't Solve Yet

Each of these approaches represents real capability and real capital behind it. What they also share, at this stage, is limited scale: narrow state licensing footprints, closed pilots with a single launch partner, or a platform still proving itself beyond early beta. These are infrastructure bets with a multi-year horizon, not something a brokerage under margin pressure today can adopt this quarter and expect to feel the difference by next quarter.

That's the gap worth naming directly. If personnel cost is 60% of the problem, and the leading solutions on the table right now require a platform migration, a licensing expansion, or a structural rebuild of how originations flow, most brokerages need something that addresses the personnel-cost driver on a faster timeline.

The Available-Now Alternative: Outsourced Operations Staffing

Restructuring how back-office work gets staffed doesn't require waiting on a platform to mature. BrokerVA's five-role model, Setup, Disclosure, Closing, Funding, and Post-Closing, moves specific stages of the personnel-cost problem to directly employed specialists who scale with volume rather than sitting on a fixed payroll during slow months. This isn't positioned as a replacement for AI tooling. Paired with our Addy AI integration, it's a way to address the sales-expense and staffing-elasticity drivers named above immediately, while AI-native infrastructure continues to mature around the edges of the process.

See the full cost comparison → BrokerVA vs. In-House Staffing: A Full Cost Comparison

What Brokerages and Lenders Can Do About It Now

• Audit which back-office functions are currently costing you in idle capacity during slow months, not just in busy-season overtime
• Consider role-specific outsourcing for the stages driving the most rework or turnaround delay, rather than an all-or-nothing staffing decision
• Treat AI tooling and staffing flexibility as complementary levers, not a choice between one or the other
• Revisit whether your current staffing model can actually flex with volume, or whether it's locked into the same hire-fire cycle that's kept per-loan costs elevated since 2008

See how the broader retention and support case connects → Broker Support Services for Wholesale Lenders: A Retention Playbook

Frequently Asked Questions

What is the average cost to originate a mortgage in 2026? Freddie Mac's most recent analysis put the average cost to produce a mortgage at $11,800. MBA's Peer Group Roundtable data shows independent mortgage companies averaging somewhat higher, $12,209 in 2025, with depositories running higher still at $16,320.

Why is personnel cost the biggest driver of mortgage origination expense? Sales expense, primarily loan officer compensation, and corporate and production support staffing together account for the majority of per-loan cost across every major data source. Unlike compliance or technology costs, personnel expense scales with volume but doesn't shrink cleanly when volume slows, which keeps it the dominant line item.

Can AI alone solve rising mortgage origination costs? Not yet, for most lenders. Current AI-native platforms show real promise but tend to carry narrow licensing footprints, single-partner pilots, or multi-year infrastructure timelines. Pairing available-now staffing flexibility with AI tooling as it matures is a more immediate way to address the personnel-cost driver.

See how a flexible, role-specific staffing model can address your personnel-cost drivers now. Contact us and we'll walk you through our cost-savings calculator using your own numbers.