Executive Summary
Venture capital has never concentrated on one idea the way it has concentrated on artificial intelligence. In 2025, by the OECD's count, AI firms took 61 percent of all global venture capital, $258.7 billion of $427.1 billion, a share that has more than doubled since 2022. By CB Insights' count the numbers differ but the story does not: AI drew $226 billion, close to half of a record $469 billion venture year, and the six largest rounds on the planet all went to AI companies. Corporate venture capital has followed the same gravity, with AI now described as a pillar of corporate innovation strategy even as CVC teams turn to secondary markets for liquidity (Silicon Valley Bank).
This report reads that market from a seat the big funding datasets do not cover: the working investor, in the room, talking about the portfolio actually in hand. Open Future Forum convened an investor room in Silicon Valley in July 2026 with more than 290 distinct applicants — venture investors, corporate and strategic investors, growth and seed funds, family offices and angels among them — and asked two plain questions about the companies they back. Who increasingly owns the AI buying decision across the portfolio, and where is AI making the biggest difference.
The answers point in one direction with one important hesitation. On value, conviction is high. The AI Investor Conviction Index, the flagship metric of this report, is the share of investors who see AI already producing a measurable difference somewhere across their portfolio. In Edition 1 it reads 86 percent (base 237), with better products the most-named area at 51 percent, ahead of cost cutting and customer impact. An investor-only cut of the same question reads 83 percent (base 126), so the signal is not an artifact of the operators in the room.
On control, the picture is less settled. The room names the CEO as the rising owner of the AI buying decision at 51 percent, well ahead of the CIO or CTO at 24 percent and finance at 12 percent (base 245). AI is being bought high, as a bet on the whole company rather than a tool for one function. Yet 22 percent still answer that it is too early to say who owns the decision at all. Conviction that AI matters has arrived before conviction about who runs it.
These first-party reads are early and directional, drawn from an application-stage instrument, with the base printed on every figure. They are set here against the external market record, including the sharp counter-evidence that most enterprise AI pilots still show no measured profit impact (MIT). The gap between 86 percent conviction and a thin record of returns is the tension this report exists to track, edition over edition.
Investors are convinced AI is already creating value; they are much less sure who inside their companies is in charge of capturing it. Open Future Forum investor room, 2026. Base 237 and 245, directional, bases on every figure.
The Answers We Have Now
This section is the first-party data in hand. It is drawn from the application instrument for an Open Future Forum investor room convened in Silicon Valley in July 2026, with more than 290 distinct applicants, reported as a rounded floor. The instrument asks screening and opinion questions inside the application flow. It is not an attendance list and it is not a survey of a closed panel. Every figure below carries its base, and multi-select questions are reported as any-mention, so they can sum past 100.
Finding 1: The portfolio is deployed, and investors can see it
The first thing the room agrees on is that AI is already in the portfolio doing work. Asked where AI is making the biggest difference across most of their portfolio, only 16 percent say nothing is measurable yet (base 237). The flagship AI Investor Conviction Index reads it from the other side: 86 percent name at least one concrete place AI already pays off, and only 14 percent name no concrete area at all. Better products leads at 51 percent, ahead of cutting costs at 35 percent and helping customers at 34 percent. This is the supply side of the venture thesis stated by the people funding it: the value is not hypothetical, it is showing up in products first.
Finding 2: The CEO is the buyer investors see rising
Asked who increasingly owns the AI buying decision across their portfolio, the room names the CEO most often, at 51 percent, roughly double the CIO or CTO at 24 percent, with individual teams at 18 percent and finance last at 12 percent (base 245, any-mention). Read from the investor seat, this is a statement about how AI purchases behave. They are company-level bets that reach the top of the house, not line items settled inside IT or finance. For an investor selling into these companies, or backing the founders who do, it says the decision maker who matters is the one who owns the whole P&L.
Finding 3: Where the value lands, in the portfolio's own words
The value question resolves toward the product line rather than the cost line. Better products at 51 percent outpaces cutting costs at 35 percent and helping customers at 34 percent, with nothing-measurable-yet at 16 percent (base 237, any-mention). That ordering matters for how investors underwrite AI. A cost-cutting story is a margin story with a ceiling; a better-product story is a growth story with a market. The room is telling us AI is being valued, so far, more as a top-line lever than as a headcount substitute.
Finding 4: Conviction is running ahead of clarity
The most useful number in the room is a hesitation. Even with 86 percent naming measurable value, 22 percent still say it is too early to say who owns the AI buying decision across their portfolio (base 245). Put those two readings together and the shape of 2026 appears: investors are sure AI is working before they are sure who, inside the company, is accountable for making it work. Value has outrun governance. That gap is exactly where an early investor either finds an edge or takes a loss, and it is the gap this series will track as the operating model settles.
How to read these numbers
- Floors, not totals. Applicant counts are distinct emails to the investor room, reported as a rounded floor. Invitation outreach is excluded from every count.
- Application, not attendance. These are answers given in the application flow, not a poll of a seated panel, and not a claim about who attended.
- A mixed room, read honestly. The investor room drew venture, corporate and strategic investors, growth and seed funds, family offices and angels, alongside founders and operators. The portfolio questions are answered most meaningfully by those who hold portfolios; an investor-identified cut of the conviction question reads 83 percent (base 126) against the full-room 86 percent (base 237), so the headline is not carried by non-investors.
- Directional bases. Every figure prints its base. Reads move edition over edition as responses grow. The flagship conviction question in its sharpest, investment-decision form is entering the field at upcoming investor events; the portfolio-impact read is the labeled seed.
No figure without its base; never present the applicant floor as a count of any one title; never publish an invitation, approval, or acceptance figure in any form; never call the portfolio-impact seed the final Index without the base and the directional label.
Early Signal from the Room
Three supporting reads, drawn from adjacent Open Future Forum rooms, put the investor answers in context. Each carries its base on its face and each is directional.
The portfolio is already running AI, not evaluating it
In the Open Future Forum finance rooms, 71 percent of respondents say their team is already running an AI tool, against 22 percent still evaluating and 8 percent not started (base 185, single-select). This is the operator-level counterpart to investor conviction: the companies in the portfolio are past the pilot question. When investors say AI already makes a measurable difference, this is the ground truth underneath it.
Where AI actually enters the company it is sold to
Founders in the Open Future Forum rooms, asked who owns the buying decision inside the company they sell to, name a business-unit leader most (39 percent), then the CIO or CTO (36 percent) and finance (28 percent). Not one of the 148 founders named security or the CISO (base 148, any-mention). Set against the investor read that the CEO increasingly owns the decision, the supply side and the demand side agree on the direction: AI buying is drifting up and toward the business, away from the traditional technology gatekeeper.
Pricing that meters value
How AI founders charge is the supply-side tell on the value question. Usage-based pricing leads at 43 percent, with outcome-based at 24 percent, flat subscription at 24 percent and per-seat at 20 percent (base 148, any-mention). Usage and outcome pricing together are the models that only work when the software delivers measurable value, because the customer pays in proportion to what it produces. Founders are pricing for the better-product world investors say they are seeing.
The flagship conviction question, in its sharpest form, asks investors directly how far AI has changed their investment pace, their diligence and their conviction, on a fixed scale. That question is entering the registration flow at upcoming Open Future Forum investor events. Until it clears the role-tagged response floor, the portfolio-impact read reported here is the labeled seed for the AI Investor Conviction Index, and this edition says so plainly.
By the Numbers
External figures first, each attributed and used for context only, then the first-party reads with their bases.
- AI took 61 percent of global venture capital in 2025, $258.7 billion of $427.1 billion, more than double its 30 percent share in 2022 (OECD, February 2026).
- By a second methodology, AI drew $226 billion of a record $469 billion venture year, close to half of all funding, in a year total venture funding rose 47 percent (CB Insights, State of Venture 2025).
- The six largest venture rounds of 2025 all went to AI companies and together took $111 billion, about 49 percent of AI funding, led by OpenAI at $41 billion and Anthropic at $32.5 billion (CB Insights).
- Mega-rounds of $100 million or more accounted for 738 deals and $307 billion, roughly 65 percent of all venture dollars, even as total deal count fell 17 percent (CB Insights). The market is writing fewer, larger checks.
- United States startups took $328 billion, about 70 percent of global venture funding, concentrating the AI trade geographically as well as by sector (CB Insights).
- Corporate venture capital is leaning in and adapting: 22 percent of CVCs used secondary markets for liquidity in 2025, up from 15 percent a year earlier, and two in three financial CVCs now operate off the balance sheet (Silicon Valley Bank, State of CVC 2025).
- The counterweight: roughly 95 percent of enterprise generative-AI pilots have shown no measurable profit-and-loss impact (MIT), even as 47 percent of AI deals reach production, nearly twice the rate for traditional software (Menlo Ventures).
From our own rooms, with the base shown:
- The AI Investor Conviction Index reads 86 percent (base 237).
- The CEO leads the AI buying decision at 51 percent (base 245).
- Portfolio teams already running an AI tool, 71 percent (base 185).
- Usage or outcome pricing among AI founders, a majority of those charging (base 148).
The Thesis, in One Line
Investors have priced AI as the dominant venture story and can already see it working in their portfolios; the open question of 2026 is not whether AI creates value but who inside the company is accountable for capturing it, and the money is concentrating faster than that answer is arriving.
The Investor Conviction Curve
This report tracks investor posture on a simple four-stage curve, its own construct, that later editions will measure movement along.
Watching
The investor sees AI as a theme but has not changed pace, price or diligence for it. In 2026 this stage is nearly empty at the top of the market and still populated at the edges, visible in the 22 percent who say it is too early to say who even owns the decision.
Backing
The investor is actively deploying into AI and can point to portfolio companies where it works. This is where most of the room sits: 86 percent name measurable value, and better products lead the way they describe it.
Concentrating
Conviction turns into position size. Fewer, larger checks chase the winners, and the six-largest-rounds-are-all-AI pattern is this stage expressed at market scale. The risk of the stage is crowding: capital arriving faster than the operating proof beneath it.
Repricing
The stage the market has not reached at scale. It begins when the measured-return record either catches up to conviction or fails to, and portfolios are marked to that reality. The MIT pilot finding is the early warning that this stage exists. The Conviction Index is built to catch the turn.
The marker that distinguishes each transition is accountability. Watching becomes Backing when the portfolio can name where AI works. Backing becomes Concentrating when conviction sets position size. Concentrating becomes Repricing when someone has to own the result and the result is measured.
What AI Looks Like Across the Investor Table
One market, seen from the different seats that make it. Each seat reads the same conviction differently.
The generalist VC
Sees AI as unavoidable and is repricing the whole book around it. The 61 percent share of venture dollars is this seat's reality: to be a generalist in 2026 is to be an AI investor whether or not that was the plan. The risk it manages is paying entry prices set by the six largest rounds.
The corporate and strategic investor
The CVC seat invests for strategic sightlines as much as return, and AI is now the pillar of that strategy (Silicon Valley Bank). It is also the seat most exposed to the who-owns-it question, because a corporate parent has to route AI decisions through its own CEO, CIO and business units, the same tangle the portfolio read exposes at 51, 24 and 12 percent.
The growth investor
Underwrites companies far enough along to show the measurable value the room describes, and is therefore closest to the repricing risk. When 47 percent of AI deals reach production but 95 percent of pilots show no profit impact, the growth seat is where that contradiction gets settled at a valuation.
The seed investor
Buys conviction before proof by definition, and in 2026 that means paying up for the AI story earliest, when a company is little more than a founder and a thesis. This seat's edge is the accountability gap itself: backing the founders who will answer the who-owns-it question inside their customers before the rest of the market prices it.
The family office and the angel
The most numerous group in the room and the least standardized. They carry conviction without a mandate, which lets them move early and also leaves them most exposed to concentration at the top. For this seat the Conviction Index is a peer benchmark: a way to see whether their read matches the room's.
The LP behind them all
Does not appear in the room but sets its weather. Every seat above is deploying capital an LP committed on an AI thesis, and the repricing stage, when it comes, is felt here first. The Index gives the LP a forward read on whether portfolio conviction is still rising or starting to cool.
Read together, the seats describe one shift seen from many chairs: capital has agreed on AI faster than the companies it funds have agreed on who runs it.
AI Across the C-Suite, Read from the Portfolio
The investor room's answer to who owns the AI buying decision is, at heart, a read on the C-suite of every company they back. Broken out by seat, base 245 for the ownership figures and 237 for value, any-mention, all directional.
The CEO, the rising owner
Named by 51 percent as the seat that increasingly owns the AI buying decision, roughly double any other role. To investors this says AI is being decided as a whole-company bet, at the seat that carries the whole result. It is the clearest single signal in the data, and it reframes AI selling: the economic buyer is moving up to the person who owns the P&L.
The CIO or CTO, the technology gate that is no longer sole
Named by 24 percent in the portfolio read, and the technology seat is confirmed from the other side: founders name the CIO or CTO as their internal buyer 36 percent of the time (base 148). Still central to how AI actually ships, but the decision has climbed above it. The CIO and CTO now share the AI call with the CEO rather than holding it alone.
The CFO and finance, the budget and not the buy
Named as the deciding owner by just 12 percent, the smallest share of any named seat. Yet the finance rooms show 71 percent of teams already running an AI tool (base 185). The read is precise: finance holds the money question and the return clock, but not the buy decision. AI spend is being approved above finance and measured after the fact.
The CISO, governance without the gavel
The sharpest gap in the data. Not one of 148 founders named security or the CISO as the buyer of what they sell (base 148). The security seat carries the cost of governing AI without owning the decision to bring it in. For investors this is a risk marker: the seat accountable for AI safety is structurally downstream of the seat that buys it.
The CMO and growth leader, where the value shows
No single seat owns the value read, but its shape favors the growth office: better products at 51 percent and helping customers at 34 percent are the top-line areas the CMO and growth leader own. When investors say AI is landing as a better-product story rather than a cost story, they are describing value that accrues first to the growth seat.
Individual teams, the shadow buyer
Bottom-up adoption is named by 18 percent as a rising owner of the decision. Real, and a live channel for how AI enters a company, but secondary to the executive seats. For an investor it is the early-usage signal that often precedes a company-level purchase, not the decision itself.
The board and the CAIO, the newest and least-settled seats
Neither shows up as a buyer in the data, and that absence is the point. With 22 percent of the room still unable to say who owns the AI decision, board-level accountability and the new chief-AI-officer seat are where the who-owns-it question is being resolved next. This is the governance layer the conviction is currently running ahead of.
Read down the C-suite, one pattern holds: AI is bought high and governed low. The seat that signs is not the seat that secures it, and the seat that funds it is not the seat that decides.
The Same View, by Sector
Five sectors investors are backing hardest, read on where conviction is firmest and where the measured-return question bites hardest. External benchmarks are representative of 2026 cross-sector venture data and are used for context.
AI infrastructure and compute
The clearest conviction and the largest checks. This is where the six mega-rounds landed and where the better-product story is really an enabling-layer story. The risk is capital intensity: the returns have to justify buildouts priced in the tens of billions.
Enterprise software and agents
Where the portfolio-impact read is most direct, because agentic software is what most of the room's companies deploy and sell. The 71 percent already-running figure lives here. The repricing risk is the pilot-to-production gap that the MIT and Menlo numbers bracket.
Fintech
The sector where cost cutting and better products meet, and where usage and outcome pricing are most natural because value is measurable in dollars moved. Investor conviction is high; regulatory and model-risk questions set the ceiling.
Healthcare and life science
The sector where AI and corporate venture capital are jointly leading the recovery in deal activity, and where the measured-return timeline is longest. Conviction here is patient capital, underwriting outcomes that take years to prove.
Consumer
The most divided seat. Better products can mean a genuine step change or a thin wrapper, and this is where the nothing-measurable-yet 16 percent concentrates. Investors are most selective here and most alert to the repricing stage.
The sector does not change the conviction question. It changes how fast, and how measurably, the answer arrives.
What Investors Say About the AI Trade
Two cases are being argued at once, and the honest read holds both.
The reckoning case
Most enterprise AI pilots still show no measurable profit impact (MIT). Concentration is extreme: fewer, larger rounds, 65 percent of all venture dollars in mega-deals, and the six biggest checks all in AI. Critics call it a conviction bubble, capital arriving on a story the operating record has not yet earned, with the repricing stage of the curve still ahead.
The counter-case
The value is not hypothetical to the people closest to it. In the room, 86 percent of investors name measurable portfolio value and 71 percent of operators are already running the tools. Production rates beat the software benchmark, with 47 percent of AI deals reaching production, nearly twice the rate for traditional software (Menlo Ventures). Pricing is moving to usage and outcomes, which only holds when value is real. This is what a market looks like when a general-purpose technology actually lands.
The reconciliation
Both are true because they describe different stages of the same curve. Conviction and measured return are not yet in sync, and the concentration at the top is the market betting they soon will be. The investors who do best from here are the ones who can tell, portfolio by portfolio, which companies have closed the accountability gap and which are still living on conviction alone. That is the distinction the Conviction Index is built to measure, and it is why this report exists.
The Evidence Behind the Theses
Five theses, each with support, counter-evidence and a verdict.
Thesis 1: AI is now the default venture asset class
Support: AI took 61 percent of global venture capital in 2025, more than double its 2022 share (OECD), and close to half by a stricter count (CB Insights). Counter-evidence: none material on the share itself. Verdict: well supported.
Thesis 2: Investor conviction that AI creates value is high and early
Support: the AI Investor Conviction Index reads 86 percent, with an 83 percent investor-only cut (base 126). Counter-evidence: the base is directional and application-stage, and the sharpest conviction question is not yet fielded. Verdict: supported first-party, stated as directional.
Thesis 3: AI is being bought high, as a company-level decision
Support: the CEO is the most-named rising owner of the AI buying decision at 51 percent (base 245), and founders confirm the drift toward business-unit and executive buyers. Counter-evidence: any-mention framing lets several owners be named at once. Verdict: supported and directional.
Thesis 4: Conviction is running ahead of governance
Support: 22 percent still say it is too early to say who owns the AI decision, even as 86 percent name value (base 245 and 237). Counter-evidence: the two figures come from different questions and are read together, not as one distribution. Verdict: supported as a directional tension, the core finding.
Thesis 5: The measured-return record has not caught up to the money
Support: roughly 95 percent of enterprise AI pilots show no measured profit impact (MIT); capital is concentrating in mega-rounds (CB Insights). Counter-evidence: 47 percent of AI deals reach production, nearly twice the rate for traditional software (Menlo), and portfolio operators report real value. Verdict: well supported and the honest counterweight to the conviction reads above.
About the Index
The AI Investor Conviction Index is the share of investors who, looking across their portfolio, see AI already producing a measurable difference somewhere, as opposed to nothing measurable yet. It is reported as a single percentage with the full distribution beneath it, overall and, as bases grow, by investor type. In Edition 1 it reads 86 percent (base 237).
The Index is built to move. Its value is not the level in any one edition but the direction across editions, read against the external market record. A conviction reading that keeps rising while measured returns stay flat is the signal of the concentrating stage of the curve; a conviction reading that falls is the first sign of repricing. No other published number reads investor conviction this way, at the portfolio level, in the same community, over time.
Definitions
- AI Investor Conviction Index. The share of investors who name at least one place AI already makes a measurable difference across their portfolio. The flagship metric of this report.
- Investor. For this report, a participant whose role or firm is in venture capital, corporate or strategic venture, growth or seed investing, a family office, or angel investing.
- Portfolio-impact read. The application-stage question on where AI makes the biggest difference across a portfolio, used as the labeled seed for the Index.
- Any-mention. A multi-select figure counting every respondent who named an option; such figures can sum past 100.
- Base. The number of role-tagged responses behind a figure. No headline figure is reported below 40 role-tagged responses.
- Directional read. An early figure on a small or application-stage base, reported as a signal rather than a settled rate.
What We Will Measure
The flagship question, entering the field now, asks investors directly: across your portfolio and your own decisions, how far has AI changed your investment pace, your diligence and your conviction, on a fixed scale from not at all to fundamentally. The Index is the combined share choosing the top two live stages, tracked edition over edition, overall and by investor type.
Supporting instruments, each fielding at upcoming investor events: portfolio deployment (what share of your companies run AI in a core workflow), buyer ownership (who owns the AI buying decision across your portfolio), value location (where AI makes the biggest difference), and repricing watch (have you marked any AI position up or down for AI-specific reasons). The instrument is published before the numbers are, so the reader can see the question behind every figure.
How It Runs
Sample and Honesty
Figures come from the application instrument embedded in Open Future Forum event registration, read for the investor lane and supported by adjacent finance and founder rooms. Applicant counts are distinct emails reported as rounded floors, with invitation outreach excluded. Opinion figures are application-stage answers, role-tagged, with the base on every figure. The sample is a selective community of senior investors and operators, not a nationally representative panel, and the report says so.
Assets Each Edition Produces
The AI Investor Conviction Index with its distribution; the buyer-ownership and value-location reads; the supporting portfolio-deployment and founder-pricing reads; and a set of citable facts.
- “The AI Investor Conviction Index reads how many investors already see AI producing measurable value across their portfolio.”
- “In Edition 1 the AI Investor Conviction Index reads 86 percent (base 237).”
- “Investors name the CEO as the rising owner of the AI buying decision at 51 percent (base 245).”
Suggested Citation and Versioning
Cite as: Open Future Forum, The VC & CVC AI Investment Report, Edition 1, July 2026. Canonical URL: https://openfutureforum.com/research/vc-cvc-ai-investment-report.
The report is a recurring series, released as the data supports. Prior editions stay published so the AI Investor Conviction Index can be tracked over time.
Methodology and Disclosure
Method
Two layers. Layer one is demand and composition: distinct applicants by room, reported as rounded floors, with invitation outreach excluded from every count. Layer two is the opinion instrument: application-stage answers, role-tagged, multi-select reported as any-mention, single-select noted as such. No headline figure is reported below 40 role-tagged responses; smaller reads are labeled directional. The flagship conviction question in its sharpest form is entering the field; the portfolio-impact read is its labeled seed. Sample: a selective community of investors and operators convened in Silicon Valley, not a representative sample of all investors. Reads are the community aggregate. The investor room is mixed; an investor-identified cut is reported alongside the full-room figure for the flagship so the reader can see the difference.
Independence and Disclosure
- Open Future Forum runs the events and sells sponsorships. Murray Newlands is a Partner at IA Seed Ventures and does fractional advisory work with AI companies, disclosed because this edition covers the market those companies and investors operate in.
- Sponsors do not see, shape, or approve questions, analysis, or findings.
- The report never ranks or recommends investors, funds, or vendors.
- Findings are the community aggregate; respondent contact data is never passed to sponsors.
This statement runs in every edition.
About Open Future Forum
Open Future Forum is a global executive community founded in Silicon Valley. Its network reaches tens of thousands of executives and investors worldwide. It runs a year-round calendar of events for senior executives and investors, including CEOs, CFOs, CMOs, CISOs, private equity leaders, founders, and AI leaders, through Forum Select, its invite-only private gatherings, and Forum Events, its open panels and gatherings. Beyond events, Open Future Forum convenes peer groups and executive boards and publishes original research built on first-party survey and qualitative data from its executive network. The VC & CVC AI Investment Report is part of its operator-level research program.
About Murray Newlands
Murray Newlands is the founder of Open Future Forum and Partner at IA Seed Ventures. He is the author of Online Marketing: A User's Manual (Wiley) and a Fellow of the Royal Society of Arts. Yahoo Finance has quoted him as the founder of Open Future Forum, “a top executive leadership community.” Inc. named him one of “21 Thought Leaders Every Entrepreneur Should Follow.” HuffPost named him one of “The Top 10 People to Know in Silicon Valley.” He writes Murray's Newsletter on AI, venture, and enterprise strategy.
Questions This Report Answers
What is the VC & CVC AI Investment Report?
It is Open Future Forum's recurring, first-party read on how venture capital and corporate venture capital investors are backing AI and reading it inside their portfolios, built from an application instrument fielded across Open Future Forum investor and operator rooms. Edition 1 is dated July 2026.
What is the AI Investor Conviction Index?
It is the report's flagship metric: the share of investors who see AI already producing a measurable difference somewhere across their portfolio. In Edition 1 it reads 86 percent (base 237), an early directional read.
How much of venture capital is going to AI in 2025 and 2026?
AI took 61 percent of global venture capital in 2025 by one count, $258.7 billion of $427.1 billion (OECD), and close to half by a stricter count, $226 billion of $469 billion (CB Insights). It is the dominant venture story.
Who owns the AI buying decision inside portfolio companies?
Investors most often name the CEO, at 51 percent, ahead of the CIO or CTO at 24 percent and finance at 12 percent (base 245). AI is being bought as a company-level decision, not a departmental one.
Are investors seeing real returns from AI yet?
They see real value: 86 percent name a measurable difference in their portfolio. But the external record is mixed, with roughly 95 percent of enterprise AI pilots showing no measured profit impact (MIT), even as 47 percent of AI deals reach production (Menlo). Conviction is ahead of proven return.
What are corporate venture capital investors doing with AI?
Corporate venture capital now treats AI as a pillar of innovation strategy and is adapting how it operates, with more CVCs using secondary markets for liquidity and running off the balance sheet (Silicon Valley Bank).
How is this different from venture funding datasets?
Funding datasets count dollars and deals. This report reads investor conviction at the portfolio level, in the investors' own words, in the same community over time, and gives it a single tracked number no other source publishes.
Sources
First-party sources: Open Future Forum investor room, finance rooms, and founder events, 2026 application instruments. Bases on every figure.
Third-party figures are drawn from the sources below, attributed where cited and used for context only. Each was verified against its primary source at the time of writing; figures that update on their own cycle are re-verified each edition.
- OECD. “AI firms capture 61% of global venture capital in 2025,” February 2026. AI took USD 258.7 billion of USD 427.1 billion in 2025, up from a 30 percent share in 2022. oecd.org
- CB Insights. State of Venture 2025. Global venture funding USD 469 billion, AI USD 226 billion; the six largest rounds all AI, totaling USD 111 billion; OpenAI USD 41 billion, Anthropic USD 32.5 billion; US USD 328 billion. cbinsights.com
- Silicon Valley Bank. State of CVC 2025. 22 percent of CVCs used secondary markets for liquidity, up from 15 percent; two in three financial CVCs operate off the balance sheet. svb.com
- Menlo Ventures. 2025: The State of Generative AI in the Enterprise. 47 percent of AI deals reach production, nearly twice the conversion rate of traditional SaaS; enterprise generative-AI spend USD 37 billion in 2025, a 3.2x year-over-year increase. menlovc.com
- MIT. The GenAI Divide: State of AI in Business 2025, MIT Project NANDA. Roughly 95 percent of enterprise generative-AI pilots have shown no measurable profit-and-loss impact. nanda.media.mit.edu
External figures are used for context only. They are not affiliated with this report and do not endorse it.
This report is for information only. It is not legal, financial, investment, tax, or accounting advice, and it is not a recommendation to buy, sell, or hold any security or to make any investment. Figures drawn from the Open Future Forum instrument are early, directional reads on selective samples, with bases shown; they are not projections. Readers should consult their own advisors. The report is provided without warranty of any kind, and Open Future Forum accepts no liability for decisions made in reliance on it.
© 2026 Open Future Forum. The VC & CVC AI Investment Report and the AI Investor Conviction Index are named works of Open Future Forum. Quotation with attribution is welcome. Third-party marks belong to their owners.
Investing in AI? Compare Notes with Investors Doing the Same
Private, off-the-record gatherings for venture and corporate investors reading AI across their portfolios. No vendors pitching you. No agenda.