A mortgage closing looks like an ending. The signatures are finished. The keys change hands. For the financial system, however, the interesting administrative work is still ahead. A loan may join a pool. The pool may support a security. Investors need disclosures, balances, payment calculations, and tax information. Somewhere, every month, the numbers have to agree.
That somewhere includes U.S. Financial Technology. The company, formerly Common Securitization Solutions, runs the Common Securitization Platform for Fannie Mae and Freddie Mac. Its position is unusual: two mortgage-market competitors jointly own the operator of their shared securitization machinery. Understanding that arrangement explains more about this business than any list of software features.
- The customers: Fannie Mae and Freddie Mac, also its owners.
- The work: securities issuance, administration, disclosures, and tax reporting.
- The lesson: standardize shared operations, migrate in stages, and count integration costs.
One security, two issuers
Consider the problem the platform had to solve. Freddie Mac’s own explanation says neither enterprise’s previous systems could handle commingled resecuritizations. These structures can combine eligible securities from both enterprises. A common security therefore required a common operational foundation, capable of storing, processing, and transmitting mortgage data at scale.
Freddie Mac went first, using the platform in November 2016. Fannie Mae joined for the June 3, 2019 launch of the Uniform Mortgage-Backed Security, or UMBS. The two releases separated an initial operational migration from the larger market transition. The resulting securities could trade in a common To-Be-Announced market, with delivery occurring after the trade is agreed.
The distinction matters: sharing infrastructure did not merge the issuers. U.S. FinTech acts as an agent and administrator. It handles the mechanics around securities, while Fannie Mae and Freddie Mac retain their institutional roles. A borrower cannot download the platform and apply for a mortgage. Lenders and investors benefit through the market it supports.
The bill behind the build
Shared infrastructure sounds economical. Getting there can be expensive. In a March 2019 report, FHFA’s inspector general put the projected combined cost of platform development and enterprise integration at $2.13 billion through June 2019. That was a project estimate, encompassing work beyond the operator itself.
The report criticized FHFA’s transparency and oversight. Earlier public updates had omitted substantial integration costs; internal assessments had also identified delivery risks. Release 2 slipped six months from an earlier timetable. The weak point in this account was already visible before full launch: the public picture of cost and execution risk lagged the internal one.
This history changes the lesson. Buying common technology does not erase the expense of connecting existing businesses to it. Integration belongs in the first budget. It also gives the company’s eventual operational achievement some proportion: a functioning market utility arrived through years of institution-building, testing, and spending.
When the rules changed
Then came a different test. Pandemic forbearance changed how distressed loans were treated inside mortgage securities. The company’s account describes previously familiar processing assumptions giving way to new borrower-assistance requirements. Fannie Mae, Freddie Mac, and CSS had to adjust their processes together.
CSS says its teams introduced repeated platform changes over a three-month period, using multiple releases. Here, the product was partly an ability to translate policy into functioning operations. A relief announcement still needed compatible software, testing, and administration before the securities machinery could accommodate it.
Resilience also became an architectural project. A 2024 AWS case study describes a move from an initial single-Region design toward active-active processing across two AWS Regions. It reports $10.7 million in annual AWS run-rate cost optimization by the end of 2023. Those operating savings belong on a different ledger from the original build cost.
Fewer tools, faster answers
The less theatrical improvements may be easier to borrow. In a February 2024 analytics case study, data executive Rishi Ranjan described a familiar mess: multiple tools, repeated architecture patterns, and systems requiring skilled people to maintain them. A request for a comprehensive view of securities and loans exposed the limitations.
“Our capabilities depended too much on people”Rishi Ranjan / AWS guest case study, 2024
The team chose Amazon QuickSight, drawing on Snowflake data and using identity controls. Exploration began in late 2021; the first product arrived by the end of the next quarter. The reported result was a 90% reduction in time spent generating reports and supporting customer and leadership inquiries. This measures a particular workflow, rather than every task in the company.
What is worth copying
Another experiment offers a smaller, cleaner instruction. In 2023, CSS described building a potential user-driven securitization prototype with AWS Amplify in under ten weeks. The team chose tools whose underlying components could support further development. The deliverable was a prototype, a way to test an idea before committing to a production system.
That approach travels. Identify a specific bottleneck. Give users something testable. Measure the resulting work. The broader joint-venture model is harder to transplant: it needs owners willing to fund common operations and accept shared governance. A cloud platform alone cannot supply that agreement, or make securities economically interchangeable without aligned market practices.
U.S. FinTech’s business rests on that institutional arrangement. Its owners invested in its creation and ongoing operation; its expertise combines securities administration, technology, data, and risk. Separate legacy systems are the relevant historical alternative. The 2025 name change makes the technology identity clearer, but the company’s useful distinction remains its job: keeping shared mortgage-market machinery working after everyone else has left the closing table.