Secondary liquidity unifies, Chinese model risk rises, and AI turns CRM into execution

Nasdaq Unifies Company and Fund Secondaries

On July 21, 2026, Nasdaq Private Market acquired Nasdaq Fund Secondaries and will merge company-share sales and multi-asset fund stakes into one secondary-liquidity platform. The combination joins Nasdaq Fund Secondaries’ end-to-end fund-secondaries execution, including QMS functionality, with Nasdaq Private Market’s SecondMarket platform for company secondaries, launched in 2024 with settlement and Tape D data. That matters because capital is concentrating in repeat buyers: Jefferies says evergreen funds now allocate about 41% of NAV to secondaries, and secondaries make up roughly 40% of the $113 billion raised by evergreen funds in 2025. After the move toward continuation vehicles and portfolio-level liquidity, this is the next step: a more unified market rail that can route both company shares and fund interests through a common process. For VC teams, that pushes secondary sales further toward a standardized workflow, with tighter valuation files, cleaner consent management, and earlier decisions on whether to sell shares or fund stakes.

Chinese Model Dependencies Are Now a Diligence Risk

Washington is weighing limits on Chinese AI models just as Beijing prepares export controls on AI model weights, turning a common technical shortcut into an immediate diligence issue. The exposure is already real: Cursor, valued at about $29.3 billion, built Composer 2 on Moonshot AI’s Kimi, and roughly 200 venture-backed startups signed a July 22 letter urging Washington not to restrict access to Chinese open-weight models from Moonshot AI and Alibaba. Industry estimates put about 80% of U.S. AI ventures on Chinese open-source models, leaving products built on Kimi, Qwen, Doubao, and DeepSeek exposed to substitution, interruption, or market-access constraints from either side’s policy move. This extends the compliance burden from infrastructure into application and procurement. DoD rules are pushing supplier vetting, bills of materials, and foreign ownership, control, or influence disclosures earlier in the cycle, with a May 2026 proposed rule requiring contractors and subcontractors on DoD contracts above $5 million to report FOCI status in NISS even for unclassified work. For investors and operators, the question is no longer which model performs best. It is whether the dependency is legal, replaceable, and contractable, and that means underwriting model provenance and procurement readiness before a term sheet, not after the first enterprise or government buyer asks.

Affinity and Navatar Push AI Deeper Into CRM Execution

Affinity’s launch of Ascend and Navatar’s Claude-integrated deal engine pushes the workflow layer one step further: the system of record is now starting to execute work. Affinity says Ascend automates sourcing, qualification, IC prep, deal and portfolio monitoring, and fundraising support, including warm-intro pathfinding, outreach drafting, record enrichment, meeting briefs, and follow-ups. Navatar takes a different route, keeping structured workflows and write-backs inside Salesforce while using Claude as an on-demand reasoning layer for sourcing, screening, research, summarization, and drafting.

That shift matters because the competitive fight is no longer AI-assisted research; it is AI-supervised operations. Affinity’s 2026 guide says 82% of firms already use AI for deal-sourcing research, so the next test is which platform can reliably update records, advance stages, and standardize follow-through across teams.

For practitioners, this extends the move away from manual venture operations toward exception handling, workflow design, and validating AI-generated actions. Analysts and associates gain leverage, but they also need tighter data hygiene, governance, and trust calibration in diligence, IC prep, and founder-facing workflows.