The Federal Reserve held its target range at 3.50%–3.75% this afternoon. This is a fifth consecutive hold and Chair Kevin Warsh's second meeting at the helm. The vote split 9–3, with three regional presidents dissenting in favor of a 25bp hike: Hammack (Cleveland), Kashkari (Minneapolis), and Logan (Dallas). Warsh's own description of it was “I asked for a good family fight and I got one.”

The decision itself was not the story. What moved markets was Warsh's argument for why standing pat was not actually standing pat. He returned repeatedly to the idea that the bond market has already done the tightening on the Fed's behalf, now that he has pulled back from forward guidance. In his words, “nominal and real yields are materially higher across the Treasury curve,” and some of the intermeeting increases in market rates rank “among the most significant in the last two decades.” He put it more plainly later: “while at some level, we haven't done much in 42 days, the markets have done quite a bit.” Asked whether this amounted to a pause, he said financial market prices “would take the other side of that.”

The market took the other side of his whole premise. Equities were green going into the press conference and reversed hard once it started. The S&P 500 fell 1.85% in the 65 minutes between 2:55pm and the close, finishing the day down 1.52% at 7,316.15. The Dow lost 1,153 points, or 2.19%, its worst session since April 2025. The Nasdaq fell 1.74%.

The Treasury curve tells the more important story for our business. The long end sold off while the front end rallied. The 10-year yield jumped 9bp to 4.69%, and the 30-year rose 12bp to 5.21%, its highest level since July 2007. The 2-month yield fell 9bp as the hike that had been priced into this meeting came back out. That is not a market agreeing that higher long rates have solved the inflation problem. It reads to me as the opposite: the bond market is telling Warsh that letting term premium do the work is not a substitute for policy, and it is demanding more compensation for holding duration until the Fed acts.

The uncomfortable part for originators is that Warsh's logic runs directly through our rate sheets. He is treating a higher 10-year as evidence the framework is working. The 10-year is also what prices 30-year mortgages. If he continues to read a long-end selloff as tightening delivered rather than as a warning, the path of least resistance for mortgage rates is higher, not lower.

Warsh named the problem himself in his prepared remarks, listing the "economic shocks of recent years" as strained supply chains, military conflicts, energy-supply disruptions, higher tariff rates, and the surge in AI-related investment. Framing those as shocks rather than transitory noise implied he expects them to persist. That is a large part of why the bond market is unconvinced.

The underlying inflation data has cooled some from the spring's stagflation scare. June CPI decelerated to around 3.5% year-over-year, down from an April peak near 3.8%, while core CPI came in flat month-over-month. Core PCE still ran around 2.8% in June. That remains well above the Fed's 2% target, which is what the three dissenters were pointing at.

AI and the Mortgage Market: A Meaningful Opportunity for Disruption

We've spent a lot of time thinking about how AI can improve efficiencies for lenders. We're using Title Wave in conjunction with Encompass and Claude for assistance with work outside the LOS. The more I find use cases for AI, the more I started to think about the borrower, and how AI can impact them.

Nearly every mortgage borrower has a smartphone and access to mobile banking, yet the mortgage origination and refinancing process still requires substantial time, paperwork, expense, and coordination.

This creates a meaningful opportunity for AI to improve the borrower experience, increase lender efficiency, and change the behavior of the broader mortgage market.

Why Borrower Behavior Matters
Homeowners generally have the right to refinance when interest rates decline, but they do not always act in a purely rational or financially optimal way. Their decisions are influenced by closing costs, paperwork, processing time, interest rate awareness, and the difficulty of determining whether the potential savings justify the effort.

One way to evaluate this inefficiency is by examining borrowers who could reduce their mortgage rate by approximately 1%. Over the past 25 years, only about 29% of these borrowers have refinanced annually, although the response has varied considerably based on market conditions, consumer awareness, credit availability, and other factors.

AI can remove many of the barriers that have historically prevented borrowers from acting when a refinancing opportunity becomes available.

How AI Can Improve the Borrower Journey
1. Discovery
Today, borrowers often have limited awareness of changes in mortgage rates and potential refinancing opportunities. With AI, borrowers can receive automatic rate monitoring, personalized savings estimates, and timely notifications when refinancing or debt consolidation opportunities become available.

2. Application
Borrowers currently spend significant time locating financial records, completing forms, and providing documentation. AI can organize and retrieve financial information, prefill applications, identify missing documents, and reduce the burden of completing the application process.

3. Underwriting
The underwriting process often requires multiple borrower interactions, manual document review, and repeated requests for explanations. AI can help borrowers understand documentation requirements, prepare explanations, and respond more quickly to underwriting conditions.

4. Appraisal and Title
Borrowers remain dependent on third-party scheduling, property data, title reviews, and other processes that can add time and cost. AI-enabled data sharing, automated valuation tools, and streamlined title reviews can reduce these dependencies and shorten the closing timeline.

5. Servicing
Borrowers typically receive limited proactive assistance from their mortgage servicer. AI can provide automated notifications, personalized financial guidance, and alerts when opportunities arise to refinance, consolidate debt, remove mortgage insurance, or improve the borrower's overall financial position.

6. Refinancing
Many borrowers delay refinancing because they are unaware of the potential savings or assume the process will be too difficult. AI can continuously compare available offers, calculate the borrower's potential benefit, and begin the refinancing process when the economics become attractive.

Implications for Mortgage Originators
I expect AI to improve efficiency throughout the mortgage process. Potential benefits include:
   · Shorter application and closing timelines
   · Reduced documentation and compliance burdens
   · Greater use of automated income, asset, employment, and collateral verification
   · Lower fulfillment and closing costs
   · Fewer errors and underwriting defects
   · More proactive and personalized borrower outreach
   · Greater lender capacity without a corresponding increase in staffing

As these capabilities develop, expect the industry to generate more originations at a given interest rate level. Borrowers may also refinance more frequently as rates decline, rather than waiting for a much larger rate incentive.

Expect AI adoption to contribute to additional industry consolidation. Larger and more technologically capable lenders may gain market share, while smaller lenders that cannot make the necessary investments may find it increasingly difficult to compete.

Regulatory and Structural Limitations
The mortgage process won't become a one-click, same-day transaction in the near term. Existing disclosure requirements, waiting periods, title processes, appraisal requirements, and other regulatory protections continue to place practical limits on how quickly a mortgage can close.

More significant disruption would likely require changes to the Truth in Lending Act, closing requirements, title insurance, or other parts of the mortgage infrastructure. To that point, the CFPB just called for comments from the public on how to expand mortgage credit, with a focus on possible changing or reducing mortgage lending regulations.

The near-term impact may also be limited because relatively few existing borrowers currently have a strong economic incentive to refinance. However, the effect could become much more significant during the next meaningful decline in mortgage rates.

Implications for Mortgage-Backed Securities

The impact extends well beyond mortgage origination.

Mortgage investors are compensated in part for taking prepayment and convexity risk. That risk exists because borrowers do not exercise their refinancing option perfectly or consistently. If borrowers become more efficient, mortgages will refinance sooner and in greater volume following declines in interest rates.

Historical prepayment models are based on decades of borrower behavior across millions of loans. AI could change that behavior in ways that historical data does not fully capture. As a result, models trained on prior refinancing patterns may become less reliable.

The amount and speed of the change in borrower response will be important drivers of mortgage valuations in the coming years. Smaller rate declines could produce greater refinancing activity, shortening mortgage durations and increasing prepayment risk.

Shorter MBS durations could also create additional demand for longer-duration assets. This may be particularly important for banks and other investors that use mortgage assets as part of their broader asset-liability and duration-management strategies. If the duration of mortgage assets shortens while the duration of deposits and liabilities remains relatively unchanged, institutions may need to add duration elsewhere in their portfolios.

Summary
AI has the potential to reshape the mortgage industry for both lenders and borrowers. Borrowers will be better informed, use AI to reduce friction throughout the loan process, lowering costs, and increasing the efficiency of refinancing.

The exact timing and magnitude remain uncertain. However, the market may be underestimating how quickly borrower behavior, origination volumes, prepayment speeds, and mortgage valuations could change once AI tools are more fully integrated into the borrower and lender experience.

Federal Funds Rate

Next FOMC Meeting

Source: CME FedWatch, various (data as of July 28–29, 2026); probability estimates have been volatile and should be treated as directional, not precise

Fed Influencing Factors

Source: BLS, BEA, Federal Reserve (FRED), NAR, CNBC (data through July 29, 2026)

Freddie Mac PMMS & Mortgage Rates

Additional Housing Statistics & Originator Update

Upcoming Market-Moving Events

Some figures in the Fed/rates section reflect fast-moving, same-day reporting and should be reconciled against a terminal (Bloomberg/FactSet) before external distribution.