RuncheyResearch

AI Absorption Ledger / MS

Morgan Stanley

MS · Q2 2026 · reported 2026-07-15 · revenue $21.35bn

Assessment

Q2 2026 says less about AI inside the firm than Q1. The call has fewer AI passages than in Q1, the CEO’s operational, equities, advisor and cyber remarks are not repeated, and the only internal statement is the CFO’s: higher technology-driven spend relates to investments to support infrastructure, AI-enabled efficiencies and business growth. The 10-Q names AI for the first time in coverage, and only as support for active capital markets beside investor sentiment.

The only sized channel is the operations saving, , the ledger’s own estimate on the Q1 multipliers. Build investment moves from a reference-class estimate to described and unsized, because the quarter’s statement names AI beside infrastructure and growth (a step from the rule applied to this quarter’s wording, not a change at the firm); its ceiling is the information processing and communications line, which rose to . The model bill, advisor co-piloting, the equities client agent and the cyber toll are not mentioned.

The AI capital cycle is where management spoke at length: the CEO put the 2026 data center spending forecast at about , up from , and declined to put a share on the firm’s role in raising it. Institutional Securities net revenues rose ; nothing separates the AI part, so the channel stays unsized, and any of it would be the buildout’s money passing through a bank. Compensation rose on advisor payouts and incentive pay, and worldwide employees fell over the quarter with no cause attributed to AI.

Sized channels against the income statement, Q2 2026

1 of 7 channels sized

Each blue mark is one channel's dollars for the quarter; a bar is the range of an estimate. Grey marks are the company's reported lines. The distance between them is the point: how large the AI channel is next to the line it sits in.

New money and old money, Q2 2026

1 new4 expanded2 relabelled

Each channel is tagged once for whether its money existed before language models, from the company's annual report and call at the start of the period. A bar splits one flow's sized dollars by that tag. The incremental total is the part that would not be there without the models: a new channel counts in full, an expanded one only for what AI added, a relabelled one at zero.

Cost displaced by AI$2.6mn to $106mn sized

expanded $2.6mn to $106mn

Incremental total $2.6mn to $106mnpoint $17mn

The sized total counts every channel the company credits to AI, including relabelled money that existed before language models and the ledger's own estimates for it. The incremental total counts relabelled channels at zero. A flow is split by layer where its dollars sit at more than one: end use, compute sold to builders, and hardware. The same dollar can be a buyer's spend, a cloud's revenue and a chipmaker's revenue, so the layers are never added together.

Paid for AI2 channels · 2 not sized

vendor bill

Model access and AI tool bill

Not sized

No source this quarter mentions a model provider, a model or an AI tool bill.

inscrutabledisclosure: not mentioned· motive: exploratory· before LLMs: new

The call, the release and the 10-Q are silent on model access this quarter. In Q1 the channel was described and unsized, since access to Claude Mythos Preview was stated and no charge was; nothing this quarter adds a price or a payment.

Evidence: 0 quotes, 1 from before coverage

What the firm pays outside model providers for model access and AI tools. The only provider named on the calls is Anthropic, named by an analyst; the CEO confirms the firm is permissioned on and working with Claude Mythos Preview. Anthropic is private and not on the ledger. No price, charge, volume or term is stated, and no source shows that any money has started. The bill would sit inside information processing and communications or professional services, and neither is split.

Why this motive

Carried from Q1: a frontier model under test with no measure.

Before LLMs: new

The anchor names no model or AI tool bill; information processing and communications expense was for FY2024, with rising technology spend among the reasons non-compensation expenses grew (claim ms-anchor-c4). Paying for access to a large language model could not exist without LLMs, so the channel is new whatever the anchor says.

“Non-compensation expenses of $17,723 million in 2024 increased 3% from the prior year, primarily driven by higher execution-related expenses and increased technology spend, partially offset by lower legal expenses and lower FDIC special assessment cost.”
Filing, mdna, 10-K periodic report, 2025-02-21

By quarter

  • Q1 2026described · described, no size · exploratory
  • Q2 2026not mentioned · inscrutable · exploratory

engineering

Internal investment in AI and agentic infrastructure

Not sized

The CFO names AI-enabled efficiencies together with infrastructure and business growth as what the higher technology-driven spend supports (claim c2), and nothing separates AI’s part (the methodology rule for AI named beside another cause). The ceiling is the rise in information processing and communications, , and the line itself, .

described, no sizedisclosure: described· motive: exploratory· before LLMs: expanded

This quarter the build investment has a statement and no measure: the CFO said higher technology-driven spend relates to investments to support infrastructure, AI-enabled efficiencies and business growth. Information processing and communications rose to , and the 10-Q puts increased technology spend among the drivers of non-compensation expenses without naming AI. Because AI is named beside infrastructure and business growth, the ledger quotes the line as the ceiling and gives no ballpark this quarter. The step from Q1, whose reading rested on a statement naming AI alone, comes from the joint-cause rule applied to this quarter’s wording, not from any change at the firm; whether the rule should follow the quarter’s evidence or the channel awaits a ruling.

Evidence: 2 quotes, 4 figures, 2 confounds, 2 from before coverage

Internal money spent building AI into the firm: the agentic infrastructure the CFO says the firm is investing in, and the technology-driven spend the CFO relates in part to AI-enabled efficiencies. The lines it sits in are compensation (technologists) and information processing and communications. Separate from the outside model bill.

Why this motive

The CFO relates higher technology-driven spend in part to AI-enabled efficiencies, beside infrastructure and business growth (claim c2), with no measure: investing ahead of a measured result, the exploratory tell, carried from Q1.

Before LLMs: expanded

At the anchor the firm was already investing in technology and giving technological tools to its advisors (claims ms-anchor-c3 and ms-anchor-c4); information processing and communications expense was and compensation for FY2024. No AI share of either line was given then or since, so no quarter of the activity before AI can be traced. The size is the whole of an activity that existed before.

“We've discussed a lot about investing in Parametric and technology. We've been giving technological tools to our advisors and investing in the business.”
CFO, qa, earnings call, 2024-04-16
“Non-compensation expenses of $17,723 million in 2024 increased 3% from the prior year, primarily driven by higher execution-related expenses and increased technology spend, partially offset by lower legal expenses and lower FDIC special assessment cost.”
Filing, mdna, 10-K periodic report, 2025-02-21

Figures

  • Information processing and communications expense · 2026-CQ2
  • Information processing and communications expense, prior year · 2025-CQ2
  • Information processing and communications expense growth, year over year · 2026-CQ2
  • Information processing and communications expense against the prior year · 2026-CQ2

What else could explain it

  • other: Infrastructure and ongoing business growth are named beside AI-enabled efficiencies as reasons for the same higher technology-driven spend, with no split.
  • line composition: Information processing and communications carries the whole technology bill, and compensation the technologists, neither split by purpose.

Quotes

“Higher technology-driven spend relates to investments to support our infrastructure, AI-enabled efficiencies, and ongoing business growth.”
c2 · CFO, prepared remarks, earnings call, 2026-07-15
“Non-compensation expenses of $5,715 million in the current quarter and $10,644 million in the current year period increased 19% and 14%, respectively, compared with the prior year periods, primarily due to higher execution-related expenses and increased technology spend.”
c10 · Filing, mdna, 10-Q periodic report, 2026-08-04

By quarter

  • Q1 2026described · our inference · exploratory · $14mn to $81mn
  • Q2 2026described · described, no size · exploratory

Cost displaced by AI2 channels · $2.6mn to $106mn sized · $2.6mn to $106mn incremental · 1 not sized

back office · expensive to verify

Operations and surveillance work displaced by AI

0.01% to 0.5% of the quarter’s revenue

Incremental total: counts in full.

our inferencedisclosure: described· motive: exploratory· before LLMs: expanded

The CFO said higher technology-driven spend relates in part to AI-enabled efficiencies. The sentence names efficiencies as a purpose of spend, not a realized saving, and gives no area, count or saving, so the share of work AI handles is carried over from Q1; the CEO did not return to the operational uses he described in Q1. Compensation rose on advisor payouts and incentive pay; worldwide employees fell over the quarter and rose over the year. The size shown is the ledger’s own estimate, , on the Q1 multipliers; every multiplier is a judgment.

Evidence: 2 quotes, 6 figures, 3 confounds, 1 from before coverage

Payroll in operational functions and surveillance that AI displaces or avoids: the CEO says efficiency around classic operational flow and surveilling is under way inside core infrastructure, and the CFO in Q2 names AI-enabled efficiencies as a purpose of technology spend. The CEO calls earlier replacements of call-center or operational work pure efficiency exercises and not a new phenomenon, so they are context, not credited to AI. The line it would show in is compensation and benefits. Money movement and surveillance errors are costly, so verification is expensive.

Why this motive

The CFO names AI-enabled efficiencies with no measure (claim c2), and no displaced line visibly shrinks: compensation rose on advisor payouts and incentive pay (claim c11). Exploratory, carried from Q1.

Before LLMs: expanded

At the anchor employees, consultants and internal systems already processed a high volume of transactions (claim ms-anchor-c6), inside compensation of for FY2024 and a workforce of . LLM tools put into that work change its cost per unit, not the existence of the line. The size is the change AI made, not the whole line.

“We rely on the ability of our employees, our consultants, our internal systems and systems at technology centers maintained by unaffiliated third parties to operate our different businesses and process a high volume of transactions.”
Filing, risk factors, 10-K periodic report, 2025-02-21

Figures

  • Compensation and benefits · 2026-CQ2
  • Compensation and benefits growth, year over year · 2026-CQ2
  • Worldwide employees at quarter end · as-of 2026-06-30
  • Worldwide employees, change over the year · 2026-CQ2
  • Worldwide employees at the end of the prior quarter · as-of 2026-03-31
  • Worldwide employees, change over the quarter · 2026-CQ2

What else could explain it

  • line composition: Compensation is dominated by advisor payouts and incentive pay, which rose on higher revenues.
  • transformation program: The March workforce management action, which the 10-Q ties to operational efficiency and performance and not to AI, falls in the year to date; worldwide employees fell over the quarter, and nothing in the fall is attributed to AI.
  • operating leverage: The CFO credits operating leverage through the first half to top-line growth and disciplined execution.

Quotes

“Higher technology-driven spend relates to investments to support our infrastructure, AI-enabled efficiencies, and ongoing business growth.”
c2 · CFO, prepared remarks, earnings call, 2026-07-15
“Compensation and benefits expenses of $8,187 million in the current quarter and $16,729 million in the current year period increased 14% compared with the prior year periods, primarily due to an increase in the formulaic payout to Wealth Management advisors and higher discretionary incentive compensation within Institutional Securities, both based on higher revenues.”
c11 · Filing, mdna, 10-Q periodic report, 2026-08-04

By quarter

  • Q1 2026described · our inference · exploratory · $2.7mn to $108mn
  • Q2 2026described · our inference · exploratory · $2.6mn to $106mn

customer support · expensive to verify

Client questions answered by an AI agent on the electronic trading platform

Not sized

No source this quarter mentions the client agent or AI in the Equity business.

inscrutabledisclosure: not mentioned· motive: exploratory· before LLMs: expanded

The CFO credited the Equity result to active markets, technology trends and the firm’s investments without naming AI. Equity net revenues were .

Evidence: 0 quotes, 1 figure, 2 from before coverage

Coverage and support time saved in the Equity business because technical client questions are answered directly by a client agent inside the electronic trading platform, as the CEO describes. The line it would show in is Institutional Securities compensation and benefits. Errors on institutional trading questions can be costly, so verification is expensive.

Why this motive

Carried from Q1: an implemented tool with no measured result.

Before LLMs: expanded

At the anchor the firm already ran clients through electronic and automated trading platforms and processed a high volume of transactions with its own staff and systems (claims ms-anchor-c5 and ms-anchor-c6); Institutional Securities compensation and benefits was for FY2024. An agent answering client questions changes the cost of that service, not its existence. The size is the change AI made, not the whole line.

“The trend toward direct access to automated, electronic markets, and the move to more automated trading platforms has resulted in the use of increasingly complex technology that relies on the continued effectiveness of the programming code and integrity of the data to process the trades.”
Filing, risk factors, 10-K periodic report, 2025-02-21
“We rely on the ability of our employees, our consultants, our internal systems and systems at technology centers maintained by unaffiliated third parties to operate our different businesses and process a high volume of transactions.”
Filing, risk factors, 10-K periodic report, 2025-02-21

Figures

  • Equity sales and trading net revenues · 2026-CQ2

By quarter

  • Q1 2026described · our inference · exploratory · $326k to $24mn
  • Q2 2026not mentioned · inscrutable · exploratory

Revenue arriving through AI2 channels · 2 not sized

sales force · expensive to verify

Financial advisor effectiveness from AI co-piloting

Not sized

No source this quarter mentions AI tools for financial advisors.

inscrutabledisclosure: not mentioned· motive: exploratory· before LLMs: expanded

The call discussed advisor matching, referral models and workplace flows without naming AI, and the 10-Q and the release are silent on advisor AI. Wealth Management net revenues were .

Evidence: 0 quotes, 1 figure, 1 from before coverage

Revenue gained because AI co-piloting gives financial advisors the history of each client relationship and helps drive client action: the corridors of super agents the CEO says Wealth Management leadership is spending time on. The clients are wealth households. An unpriced tool on existing advice revenue. The roster note that the firm has deployed AI assistants for advisors with a model provider is not in any covered source and is not used.

Why this motive

Carried from Q1: co-piloting described as work in progress.

Before LLMs: expanded

At the anchor the firm was already giving technological tools to its advisors (claim ms-anchor-c3), and Wealth Management earned of net revenues in FY2024. The advice work existed; LLM co-piloting changes how much revenue an advisor produces, so the channel is expanded and sized as an increment. No lift has been measured. The size is the change AI made, not the whole line.

“We've discussed a lot about investing in Parametric and technology. We've been giving technological tools to our advisors and investing in the business.”
CFO, qa, earnings call, 2024-04-16

Figures

  • Wealth Management net revenues · 2026-CQ2

By quarter

  • Q1 2026described · described, no size · exploratory
  • Q2 2026not mentioned · inscrutable · exploratory

product revenue

Investment banking and markets activity from AI themes and the AI buildout

Not sized

The 10-Q names the adoption of AI together with improved investor sentiment (claim c3), and the CEO declines to attach a percentage to the firm’s role in raising the buildout’s capital (claim c7); nothing separates AI’s part (the methodology rule for AI named beside another cause). The ceiling is Institutional Securities net revenues, .

described, no sizedisclosure: described· motive: exploratory· before LLMs: relabelled

This quarter the filing itself names AI: the 10-Q says active capital markets were supported by the adoption of AI and improved investor sentiment. Asked for numbers on the AI capital spending cycle, the CEO said the 2026 data center forecast had risen to about from , economy-wide figures from the firm’s research, and that the firm’s share of raising that capital is going to be meaningful, without a percentage. Institutional Securities net revenues rose to and Investment Banking . The clients’ money is mixed: hyperscaler cash flow, corporate cash flow and investor capital. Any part that is the AI buildout is its money passing through an intermediary.

Evidence: 7 quotes, 9 figures, 2 confounds, 1 from before coverage

Institutional Securities revenue from client activity the firm ties to AI: the CFO names AI themes beside geopolitical uncertainty and market dispersion as drivers of client engagement, the Q2 10-Q names the adoption of AI beside investor sentiment as supporting active capital markets, and the CEO describes the AI capital spending cycle as a capital-raising opportunity whose share he declines to put a percentage on. Advising, underwriting and trading for clients predates LLMs; any of it that is the buildout is the buildout’s money moving through an intermediary, not absorption by the clients.

Why this motive

The 10-Q names the adoption of AI beside improved investor sentiment as support for active capital markets (claim c3), the CEO names AI beside regulatory normalization as urgency for M&A and declines to put a share on the buildout’s capital raising (claims c8 and c7): AI named as part of a cause with no measure, exploratory, carried from Q1.

Before LLMs: relabelled

At the anchor the firm already earned investment banking revenue as advisor, underwriter and distributor of capital (claim ms-anchor-c7): in FY2024, within Institutional Securities net revenues of . Management gives no measure of the AI part, so the tie-break reads it as existing activity described with AI named.

“Investment banking revenues are derived from client engagements in which we act as an advisor, underwriter or distributor of capital.”
Filing, mdna, 10-K periodic report, 2025-02-21

Figures

  • Institutional Securities net revenues · 2026-CQ2
  • Institutional Securities net revenues, prior year · 2025-CQ2
  • Institutional Securities net revenues growth, year over year · 2026-CQ2
  • Total Investment Banking revenues · 2026-CQ2
  • Total Investment Banking revenues, prior year · 2025-CQ2
  • Total Investment Banking revenues growth, year over year · 2026-CQ2
  • Equity sales and trading net revenues · 2026-CQ2
  • Data center capital spending in 2026, current expectation per the CEO (economy-wide) · FY2026
  • Data center capital spending in 2026, forecast of November 2025 per the CEO (economy-wide) · FY2026

What else could explain it

  • other: Improved investor sentiment, regulatory normalization, IPO markets reopening and sponsor monetization are named beside AI as causes of the same activity, with no split.
  • mix shift: Institutional Securities revenue moves with volumes, volatility and issuance across all clients; the 10-Q credits the Equity increase to client activity, particularly in Asia.

Quotes

“The economic environment exhibited strength in the second quarter of 2026, characterized by active capital markets supported by the adoption of AI and improved investor sentiment.”
c3 · Filing, mdna, 10-Q periodic report, 2026-08-04
“The forecast for 2026 on data center CapEx that was taken late last year, around November of 2025, was that $575 billion would be spent this year, and it's coming in at about $850 billion.”
c4 · CEO, qa, earnings call, 2026-07-15
“You're basically looking at us being around 10%-15% of the way through the investment cycle.”
c5 · CEO, qa, earnings call, 2026-07-15
“I think the answer is that some of it can just be done naturally through cash flow generation of the leading hyperscalers. Some of it is going to have to be with fresh capital, whether it be debt or equity, and a bunch of it's going to have to be done very creatively to access the right investor base.”
c6 · CEO, qa, earnings call, 2026-07-15
“On that one, I'd say I wouldn't attach a percentage to it, but I would say it's going to be meaningful.”
c7 · CEO, qa, earnings call, 2026-07-15
“The urgency, in our view, is twofold, and it gets to the two themes of the letter. The first is AI.”
c8 · CEO, qa, earnings call, 2026-07-15
“Active markets and technology trends serve as tailwinds, and our multi-year investments in our global franchise allowed us to prosecute greater levels of client engagement.”
c9 · CFO, prepared remarks, earnings call, 2026-07-15

By quarter

  • Q1 2026described · described, no size · exploratory
  • Q2 2026described · described, no size · exploratory

Cost imposed, or revenue lost, by others’ AI1 channel · 1 not sized

other

Cyber defense against AI-enabled attacks

Not sized

No source this quarter mentions AI-enabled attacks or the cost of defending against them.

inscrutabledisclosure: not mentioned· motive: imposed· before LLMs: relabelled

Cyber risk from AI did not come up on the call, and the release and the 10-Q add nothing on it. In Q1 the channel was described and unsized, since AI was named beside a broad rise in cyber risk; nothing this quarter separates its part.

Evidence: 0 quotes, 2 from before coverage

Security spending the firm carries because attackers use AI: the CEO says ecosystem risk is likely increasing because of the quality of the newest model and the firm needs to take its defense to another level. The cost sits in information processing and communications and compensation and is not split.

Why this motive

Carried from Q1: a toll the firm did not choose.

Before LLMs: relabelled

The FY2024 annual report already said attackers may become more effective by using artificial intelligence and called cybersecurity a continuing investment (claims ms-anchor-c1 and ms-anchor-c2); information processing and communications expense was for FY2024. No line is shown to move because of AI, so the tie-break reads it as the existing cyber budget described with AI named.

“Any of these parties may also attempt to fraudulently induce employees, customers, clients, vendors or other third parties or users of our systems to disclose sensitive information in order to gain access to our networks, systems or data or those of our employees or clients, and such parties may see their effectiveness enhanced by the use of artificial intelligence.”
Filing, risk factors, 10-K periodic report, 2025-02-21
“We continue to make investments with a view toward maintaining and enhancing our cybersecurity, resilience and information security posture, including investments in technology and associated technology risk management activities.”
Filing, risk factors, 10-K periodic report, 2025-02-21

By quarter

  • Q1 2026described · described, no size · imposed
  • Q2 2026not mentioned · inscrutable · imposed

Reported lines, year-over-year growth

Revenue +27.1%

Q2 2026. Growing slower than revenue: total non-interest expenses (+16.1%), compensation and benefits (+13.9%), information processing and communications (+10.5%). A displaced cost shows up as a line that stays under the dashed revenue line. These are the audited lines, as first reported; nothing here is attributed to AI by the filing.

Total non-interest expensesCompensation and benefitsInformation processing and communicationsRevenue
0%10%20%30%Q1 2025Q2 2025Q3 2025Q4 2025Q1 2026Q2 2026RevenueTotal non-interest expensesCompensation and benefitsInformation processing and communications
Reported values and filings
LineQ1 2025Q2 2025Q3 2025Q4 2025Q1 2026Q2 2026
Total revenues, net of interest expense$17.74bn$16.79bn$18.22bn$17.89bn$20.58bn$21.35bn
Total non-interest expenses$12.06bn$11.97bn$12.20bn$12.11bn$13.47bn$13.90bn
Compensation and benefits$7.52bn$7.19bn$7.44bn$7.06bn$8.54bn$8.19bn
Information processing and communications$1.05bn$1.09bn$1.12bn$1.16bn$1.15bn$1.20bn