← All firm reviews
Refinery · Private research

Menlo Ventures: strategic review

23 September 2026 · Deep public-evidence review · Laurie review pending

Menlo Ventures: the AI market-data firm with a relationship engine behind it

Menlo Ventures appears to be trying to own a specific form of AI authority: the investor that can measure where adoption and spending are moving, then connect founders to the companies and model provider shaping that movement. The current homepage still promises founder partnership, “When we invest, we’re invested”, while its most prominent proof is now AI research, the Anthropic relationship and $3bn of new capital described as going “ALL IN on AI” [identity-home] [history-50].

That places Menlo beside Bessemer in recurring market intelligence, a16z and Lightspeed in large-firm AI visibility, and Redpoint in research distributed through named investors. Menlo’s strongest ownership is narrower: repeated survey-led measurement of enterprise, healthcare and consumer AI adoption. Bessemer owns longitudinal cloud benchmarks and public decision artefacts more strongly; a16z owns media infrastructure and distribution as a founder service; First Round and Index make operating help more publicly reusable [peer-bessemer] [peer-a16z] [peer-index].

The system is effective for AI recognition, useful proof and third-party transmission. It is partly effective at making Menlo’s founder partnership distinctive. The strongest public evidence of “ALL IN” is found in Anthology and Inception, yet those programmes sit behind the market-data franchise in the current public impression.

The reports have become Menlo’s clearest public product

The 18-month website census contains 68 dated items: 37 tagged funding, 14 perspective, 5 report, 7 firm, 3 how-to and 10 portfolio-only or exit items, with category overlaps [publication-census]. Transaction writing supplies volume. The 5 reports supply the stronger association. They form 3 developing sequences: enterprise AI in 2023, 2024 and 2025; consumer AI in 2025 and 2026; and healthcare AI in 2025 [reports-index].

The work has costly ingredients. The 2025 enterprise report combines 495 US decision-makers with a bottom-up market model. The healthcare report uses more than 700 executives and over 20 additional stakeholder conversations. The 2026 consumer report surveys 5,067 US adults and discloses weighting, changed questions and the limits of year-on-year comparison [report-enterprise-2025] [report-healthcare-2025] [report-consumer-2026]. This moves Menlo beyond opinion into a benchmark people can cite, challenge and reuse.

Transmission is observable. TechCrunch used the mid-year LLM work to report Anthropic’s enterprise share. NPR’s Marketplace interviewed Amy Wu Martin about consumer adoption. Cisco cited the consumer report in a 2026 infrastructure study. Independent LinkedIn users reproduced and criticised the enterprise findings, including a critique that contrasted reported application spend with announced infrastructure commitments [reception-techcrunch] [reception-npr] [reception-cisco] [reception-critical]. Laurie’s January 2026 market map had already identified the mechanism: Menlo combines blog and research, then distributes individual findings through partners [laurie-market-map].

This is effective category-building. The limitation is the interested vantage point. Menlo invests across the markets it measures, and the reports frequently end with portfolio examples and an invitation to founders. Methodology and limitations increase trust; they do not make the work independent. The 2025 enterprise report also restates prior-year totals after excluding inference, showing that the benchmark is still being refined [report-enterprise-2025]. Evidence that independent practitioners rely on the series over several annual cycles, including when its findings complicate Menlo’s portfolio thesis, would strengthen the ownership claim.

Anthropic turns the AI claim into access, not just language

The 2026 anniversary account presents Anthropic as Menlo’s 2023 flag-planting moment, followed by the firm’s largest investment, repeated follow-ons and the 2024 Anthology Fund [history-50]. Anthology makes that relationship available to earlier founders. Its rolling application, 2-week response expectation, $100,000-and-up starting investments, model credits, quarterly deep dives, biannual demo days and access to Anthropic leaders create a concrete route from attention to relationship [programme-anthology]. By October 2025, Menlo reported 45-plus companies and a founder event featuring Anthropic product and research leaders; by June 2026 it reported more than 60 companies and 3 exits [programme-anthology-update] [history-50]. Those outcomes are firm-reported.

This is Menlo’s most distinctive programme because the proof is scarce access, a named partner and repeated portfolio formation. It also creates concentration risk. Anthropic strengthens Menlo’s AI credibility while making part of the position depend on one company relationship. Bessemer’s cloud authority rests on several compounding objects and public-market partnerships. Menlo’s AI authority currently combines its own surveys with privileged proximity to one frontier lab [peer-bessemer]. A founder survey showing that Anthology access materially changes product, distribution or fundraising would turn a credible mechanism into demonstrated effectiveness.

Inception is the better proof of the founder promise, but the research system rarely points to it

Inception offers about 10 founders a year up to 3 months of 10 to 20 hours a week with Menlo, initial SAFE funding, customer and wedge work, recruiting and GTM support, followed by possible pre-seed investment [programme-inception]. This is unusually concrete evidence for “When we invest, we’re invested”. It reveals who Menlo meets first, how early it starts and what hands-on help means.

The observable business loop is therefore substantial. A report or partner post creates an encounter. Anthology or Inception supplies an application. A selected founder receives capital, technical access or company-building support. Investments and programme updates become later evidence. The assets and routes are observed; content-attributed applications, investments and company outcomes are unknown.

The weakness is public integration. Report conclusions usually lead to “we’d love to meet you”, while the homepage promotes Anthology more visibly than Inception [identity-home] [report-consumer-2026]. The reader has to discover that Menlo also operates a highly selective company-building sprint. a16z connects media, programmes and portfolio service more explicitly. Index turns recurring operating problems into tools. Menlo’s genuine difference may be under-explained: it can move from a large adoption dataset to a small, labour-intensive founder intervention [peer-a16z] [peer-index].

The partner network distributes one research house more than 6 separate voices

The selected people create a federated expert system. In the 18-month WordPress taxonomy, Derek Xiao appeared on 12 items, Deedy Das on 11, Tim Tully on 10, Amy Wu Martin on 8, Greg Yap on 5, Shawn Carolan and Johnny Hu on 4 each [partner-tag-census]. Derek operates the sparse Directionally Correct newsletter, with 4 visible essays from January 2025 to January 2026, cross-linking Menlo research. Deedy’s technical builder identity travels through interviews and a substantial personal X presence, although X itself was blocked. Amy and Shawn carry consumer AI. Greg, Johnny and Derek carry healthcare [programme-directionally-correct] [partner-deedy] [partner-guest-deedy] [partner-shawn-guest].

The firm account then atomises reports across LinkedIn. The accessible current sample showed repeated charts and findings, partner reposts, portfolio reuse and external interviews around the 2026 consumer report [linkedin-current]. That confirms Laurie’s distribution observation. It also reveals the tradeoff: partners are effective carriers of one institutional research object, while fewer people sustain an independently memorable argument. Derek is the clearest exception.

No official owned podcast, YouTube channel, Instagram programme or TikTok programme was located after site, platform and name-format searches. X was inaccessible and LinkedIn exposed a dynamic sample, so those channels remain unknown beyond what was visible [platform-searches]. The report pages use short Vimeo banner loops and polished downloadable PDFs; direct video playback was blocked, and the inspected 2026 embed was an 8-second atmospheric banner rather than an editorial programme [creative-report] [audiovisual-limit]. Menlo’s creative signature is data visualisation, compact claims and partner-led atomisation. Redpoint owns platform-native creative experimentation more strongly; Lightspeed and Sequoia own recurring video and audio formats more strongly [peer-redpoint] [peer-lightspeed] [peer-sequoia].

Effectiveness and overturn conditions

Recognition: effective, especially around AI and Anthropic. Association: effective for survey-led AI adoption intelligence; shared for broad AI leadership and founder partnership. Proof: strong for research production, applications and programme design; partial for support outcomes. Transmission: effective through press citation, practical reuse and partner distribution; full social denominators are unavailable. Action: strong through Anthology and Inception applications, direct partner contact, jobs and portfolio routes. Commercial conversion is unknown.

The practical market implication is that Menlo already owns something narrower and more credible than “ALL IN on AI”: a recurring view of where AI demand moves, joined to selective routes for founders building into that demand. The reading would change if target founders primarily remember Anthropic rather than Menlo; if independent audits challenge the survey models; if private attribution shows little qualified inbound; if Inception outcomes reveal a distinctive repeatable company-building method; or if a peer’s adoption dataset becomes the category’s default reference.