What is AI in CRE lending? AI in CRE lending is the use of machine learning, agentic software, and predictive analytics to underwrite, price, monitor, and stress test commercial real estate loans faster and with fewer errors than manual credit work. That definition stopped being theoretical in July 2026, when the five largest US banks reported second quarter earnings, quietly cut the reserves they hold against commercial real estate loan losses, and named AI and data centers as their single biggest growth story. For the full picture of how technology is reshaping debt, see our guide to AI CRE finance and capital markets.
Key Takeaways
- In Q2 2026, JPMorgan Chase, Goldman Sachs, Bank of America, Wells Fargo, and Citigroup cut or held CRE loan loss provisions as the long feared default wave failed to materialize.
- Goldman Sachs cut its provision for credit losses to $102 million from $315 million in Q1 2026, a 68% reduction, while its commercial real estate book grew 21% year over year to about $40 billion.
- JPMorgan CFO Jeremy Barnum called data center underwriting a bellwether for lender appetite, citing rising risk tolerance and power supply questions.
- Lower reserves signal more lender liquidity, which can widen credit availability and tighten spreads for well underwritten CRE deals in late 2026.
- Banks are pairing looser reserves with AI underwriting and continuous loan surveillance, changing how quickly deals get approved and how closely portfolios are watched.
What the Q2 2026 Bank Earnings Actually Showed
The largest US banks are far less worried about commercial real estate than they were a year ago, and their Q2 2026 numbers prove it. Across earnings calls, leaders said the dreaded pandemic era wave of CRE defaults never fully materialized, and they backed that view by trimming the money set aside for future loan losses. A provision for credit losses is the expense a bank books to cover expected defaults, so cutting it is a direct statement of confidence in the loan book.
Goldman Sachs made the boldest move, cutting its provision to $102 million from $315 million in the first quarter, a 68% reduction and the largest percentage cut among its peers. CEO David Solomon highlighted fast CRE lending growth, with the bank's commercial real estate book up 21% year over year to roughly $40 billion and net earnings of $6.3 billion for the quarter. JPMorgan Chase trimmed its provision by more than 10% to $2.5 billion and posted a 41% profit jump to $21.2 billion. Bank of America cut its provision by more than 10% to $1.4 billion, grew net income 27% to $9.1 billion, and reduced criticized commercial loan exposure by $2.3 billion quarter over quarter. Wells Fargo nudged its allowance for credit losses to 1.4% of total loans, still below the level it carried in Q2 2025, and pointed to real improvement in its office loan portfolio.
Why Banks Are Betting on AI and Data Center Lending
AI and data centers were the dominant theme on every major bank call, and that is not a coincidence. As legacy office risk fades, lenders are reallocating CRE capital toward the properties that power artificial intelligence. JPMorgan CFO Jeremy Barnum described the data center underwriting space as a bellwether for lender appetite, flagging both the power supply questions that dog large campuses and a rising tolerance for risk among lenders chasing the sector.
The scale of the opportunity is hard to ignore. The AI in real estate market is projected to reach $1.3 trillion by 2030 at a 33.9% compound annual growth rate, and data center demand is spilling into adjacent industrial and warehouse space. Banks see AI driven tenants and infrastructure as the rare CRE growth story in an otherwise cautious market, a dynamic that Deloitte's 2026 Commercial Real Estate Outlook frames as fresh capital and lender activity reengaging even amid pockets of distress. It also mirrors the consolidation trend we covered in AI lender consolidation in CRE debt, where scale and data advantages concentrate lending power among the biggest players.
What Lower Loss Reserves Mean for CRE Borrowers and Investors
Lower reserves usually translate into more available credit. When a bank releases provisions, it frees capital that can be redeployed into new loans, which tends to widen credit availability and compress spreads for qualified borrowers. That matters because the industry still faces a multi trillion dollar wall of CRE loan maturities coming due through 2027, tracked in detail by the Mortgage Bankers Association. A more confident lender is more likely to refinance a maturing loan, extend a bridge, or fund a value add business plan.
The caveat is that this confidence is selective. Banks are underwriting AI adjacent and improving office assets, not blanket approving every deal. Debt service coverage ratio, the ratio of net operating income to annual debt service, still governs loan sizing, and a 1.25x DSCR requirement does not relax just because reserves fell. Investors should read the reserve cuts as a green light for well underwritten deals, not a return to loose 2021 credit. The market's confidence is also visible in the bond market, a theme we explored in the bond market's AI bet.
How AI Is Reshaping CRE Lending Decisions
Behind the reserve math sits a quieter change: banks are increasingly using AI to decide who gets credit and how loans are watched. AI underwriting tools ingest rent rolls, trailing twelve month operating statements, and market comparables in seconds, flag inconsistencies, and produce a first pass credit view faster than a human analyst. Continuous loan surveillance models now monitor portfolios in real time, catching a slipping DSCR or a softening submarket before it becomes a default.
For lenders, the payoff is efficiency and earlier risk detection, which is exactly what justifies lower reserves: if you can see trouble sooner, you need to hold less against surprises. CRE investors and debt funds can use the same tools defensively. Running your own portfolio through AI loan portfolio stress testing lets you model rate shocks and vacancy scenarios the way your lender does. For personalized guidance on building these workflows, CRE investors can reach out to Avi Hacker, J.D. at The AI Consulting Network.
Action Steps for CRE Investors in a Reopening Credit Market
The practical response to easing bank reserves is to get financing ready before the window narrows. Use the current confidence to lock terms on well underwritten deals rather than waiting for perfect pricing.
- Refinance maturing debt early: With provisions falling, engage lenders on 2026 and 2027 maturities now, while risk appetite is rising.
- Lead with clean, AI ready data: Lenders reward borrowers who deliver structured rent rolls and financials their AI underwriting can process without back and forth.
- Position AI adjacent assets: Industrial, data center adjacent, and power connected sites are where lender appetite is strongest right now.
- Stress test before the bank does: Model your own DSCR, cap rate, and refinance scenarios so a lender's AI driven review holds no surprises.
If you are ready to transform your underwriting and financing process with AI, The AI Consulting Network specializes in exactly this kind of implementation for CRE owners, sponsors, and debt funds.
Frequently Asked Questions
Q: Why did big banks cut CRE loan loss reserves in Q2 2026?
A: Banks cut reserves because the widely feared wave of commercial real estate defaults never fully materialized, office loan portfolios improved, and AI and data center lending emerged as a growth engine. Goldman Sachs cut its provision 68% to $102 million, and JPMorgan and Bank of America trimmed theirs by more than 10%.
Q: What do lower bank reserves mean for CRE borrowers?
A: Lower reserves generally mean more available credit and greater willingness to lend. For CRE borrowers, that can translate into easier refinancing, tighter spreads, and more appetite for well underwritten deals, though banks remain selective and DSCR requirements still apply.
Q: How is AI changing commercial real estate lending?
A: AI is automating underwriting, term sheet comparison, and continuous loan surveillance. It lets lenders produce a first pass credit view in seconds and monitor portfolios in real time, improving efficiency and earlier risk detection. The AI in real estate market is projected to reach $1.3 trillion by 2030.
Q: Are data centers really driving bank CRE strategy?
A: Yes. On Q2 2026 earnings calls, AI and data centers were the dominant theme. JPMorgan CFO Jeremy Barnum called data center underwriting a bellwether for lender appetite, and banks are reallocating CRE capital toward AI infrastructure and away from legacy office risk.