A new generation of mortgage platforms is deploying artificial intelligence to overhaul the commercial lending process for middle-market borrowers, promising faster deal cycles and better pricing for properties that often fall outside the purview of major institutions. These loans, typically ranging between $5M and $100M, fuel strip malls, small industrial projects and independent apartment investments across a fragmented and decentralized segment of the market.
The Mortgage Bankers Association projects total commercial and multifamily mortgage originations will jump 27% year-over-year to $805.5bn in 2026, up from $633.7bn in 2025. The middle market accounts for roughly 40% of that annual lending volume, according to Mitch Ginsberg, founder of mortgage marketplace CommLoan. Borrowers in this tier often lack the credit relationships that institutional investors command, leaving them dependent on independent brokers who lean on personal networks and cold-calling to secure terms.
Platforms including CommLoan, Lev and Janover aim to replace that manual process with algorithmic matching, connecting brokers and borrowers to a curated universe of lenders while preserving the human relationship at the centre of dealmaking. CommLoan, which launched in 2015, has onboarded more than 1,000 lenders into its network and facilitated approximately $650M in loans last year. The firm is on track to double that figure in 2026 and plans a reverse merger with Windtree Therapeutics Inc. later this year to go public.
"We recognize the relationship is critical. That's why we onboard the actual mortgage broker and don't replace them," Ginsberg said. "We're really creating enormous efficiencies for the mortgage broker with that matching process." Every commercial mortgage has a genetic code consisting of factors such as borrower credit history, property-type experience, and market conditions, he explained. Most lenders end up in deals that work but fall short of their exact parameters.
CommLoan uses artificial intelligence and a proprietary database to standardise loan applications and continuously update lender appetite, showing borrowers and brokers varied options including local banks, credit unions and private lenders. Recent transactions facilitated through the platform include a $5.4M refinance of a car wash in Marshall, Texas, and a $4.9M acquisition loan for a single-tenant office in Los Angeles. The company is also expanding into larger deals, having helped close a $195M loan for the Kali Hotel in Hollywood Park, a development in Inglewood, California, adjacent to SoFi Stadium.
Using CommLoan, developer Kali Chaudhary secured a 6.3% interest rate with Bank of America's high net worth group for that project. Previously, other brokers had quoted loans at 10.5% or more, meaning the CommLoan deal shaved off roughly $850K a year in interest charges. Chaudhary's group, KPC, is also behind a bid for the unfinished Oceanwide Tower in Downtown Los Angeles, which Ginsberg hopes CommLoan may help finance.
"Even a borrower like that, Dr. Kali Chaudhary, who's the sponsor, who certainly has substantial net worth and liquidity, could go to any lender that he wanted to," Ginsberg said. "Using our matching technology, we were able to match him to the lender that would best suit his particular situation at the time." CommLoan operates nationally and covers all property types, from car washes and self-storage to multifamily. Industrial and retail deals are particularly active, especially in Phoenix near the Taiwan Semiconductor Manufacturing Co. plant, Ginsberg noted.
Established brokerages are also adopting AI to defend market share. Eastern Union, a middle-market commercial mortgage brokerage that facilitated roughly $2bn in loans last year, has deployed in-house artificial intelligence to accelerate deal matching, according to CEO Abe Bergman. "Real estate business margins have thinned over the last few years," Bergman said. "This will allow the real estate industry to be able to improve those margins."
The technology race extends beyond incumbents. YieldStack, founded last year by recent college graduates Rommin Adl and Daniel Chesney, is pursuing the same opportunity: accelerate commercial lending technology, which founders view as a decade behind residential mortgage infrastructure, and combine algorithmic matching with advisory services for smaller sponsors. The fragmented nature of middle-market lending leaves room for multiple platforms to capture share as origination volumes rise and borrowers seek pricing advantages in a tightening credit environment.