Overall GTM Job Posts Down 15%, AI-Natives Up 50%: There were 22,988 GTM job posts in Q1 2026, down 15% year-on-year. AI-native companies bucked the trend with hiring up nearly 50% YoY, though they still represent only 5% of all digital-native GTM posts and 2% of total GTM headcount.

SDR Headcount at AI-Native Companies Has More Than Doubled: Across the broader market, SDR/BDR posts fell 21% year-on-year. AI-native companies including Cursor, Decagon, LangChain, and OpenAI are hiring SDRs at 50% higher rates relative to their GTM mix versus all digital natives.

GTM Engineering Doubles, Customer Support Collapses: There are now 400+ GTM engineers at US digital native companies, with headcount doubled year-on-year and Claude mentioned in 12% of GTM engineering job posts in Q1 2026, up 3x from Q4 2025. Customer support saw the steepest decline of any GTM role, down 37%, reflecting traction from AI support products like Sierra ($15.8B) and Decagon ($4.5B).

Median Lead Check Requirements Have Nearly Doubled: The median seed round reached $4.5M in 2026, up from $2.5M in 2019. Series A median is now $16M, up from $9M. The median check to lead a seed round is now $2.3M vs. $1.2M in 2019; Series A lead checks are $8M vs. $4.5M. Median lead checks are 25-50% larger than just three years ago.

Three Linked Fund Decisions Are Now Structurally Harder: Walker frames fund size as a function of three interrelated choices: whether to lead rounds (check sizes much larger), whether to follow-on (reserves must scale with Tier 1 markups), and whether to bridge companies (bridge capital competes with net-new bets). Each decision now costs materially more than the last fund cycle.

Incentives Point Toward Raising Bigger, Which Creates Misalignment Risk: The pressure across all three decisions points in the same direction: raise a larger fund next time. Walker notes LPs tend to welcome this, but a fund raised to a size that mismatches a GP's actual strategy or network density creates underperformance risk that compounds over years.

Thrive Appears in 85% of Fellowship Interviews With Zero Content: Nucleus Talent data shows Thrive comes up in more partner and fellowship conversations than almost any fund with an active content strategy, despite having no newsletter, no podcast, no blog. Benchmark has no marketing function at all. These are not failures to build a brand but deliberate choices: brand lives in portfolio quality, not editorial output.

The Anti-Brand Only Works With Three Conditions Already in Place: Owen argues the anti-brand requires network density, ecosystem entry point, and a concentration philosophy. Thrive's first fund was $40M; Greenoaks was founded in 2012. Neither had track records that justified going dark at launch. What they had were closed networks that predated the returns. The philosophy came first; the silence was chosen to enable it.

Conviction Represents the Explicit, Transferable Model: Sarah Guo built Conviction's brand around demonstrable network (strategic associations with Nvidia and OpenAI), a published worldview (LP letters, public positions), and partner visibility rather than institutional brand. Their website deliberately copies Berkshire Hathaway anti-design. All brand investment goes into partners, not the firm name.

54% of New Enterprise Logos Now Come Through Self-Serve: After Claude Opus 4.6 demand went vertical in December 2025, Anthropic launched an enterprise self-serve MVP in January 2026 and reached production in February. Within four months, 54% of new enterprise logos came through a fully self-serve funnel (real ACV, real terms, real invoicing) without an AE, using Clay and Claude for qualification and Intercom Fin to guide the buyer journey.

5 Claude Skills Replace the 6-Week Onboarding Ramp: Every new rep gets a territory and a plug-in via MCP connectors bundling: (1) Morning Briefing, a daily prioritised action list from Gmail, Gong, Slack, Salesforce, and calendar; (2) Call Prep, a full one-pager in under five minutes; (3) Customer Follow-Up, extracting action items and drafting responses against a 24-hour SLA; (4) Competitive Intel, dynamic deal-specific battle cards; and (5) Create an Asset, custom on-brand collateral or ROI calculators for any deal.

Claude as Connective Tissue, Not a Seventh Tool: Anthropic did not replace Salesforce, Gong, Ironclad, Clay, LeanData, or Slack. They threaded Claude through the seams between all six. Proposals are drafted, policy-validated, and uploaded to Ironclad via Claude; forecasting is run by Claude and inspected by managers.

Skills and Cowork Were the First Steps; /workflows Is the Final Form: Miessler argues every job is a series of steps to accomplish a goal. Skills and Cowork have been moving in this direction, already visibly impacting valuations in affected spaces. The /workflows feature converts regular, expected company work into pseudo-deterministic workflows that follow defined SOPs, executed by Claude Code rather than humans.

The Human Role Shifts to Three Functions: Once workflows handle execution, humans are left with determining what problems to solve (taste, experience, judgment), building new products from that insight, and optimising the workflows from above. The work itself becomes the workflows; human contribution becomes the meta-layer that defines and improves them.

Enterprise Is the Primary Unlock: Large organisations run on repeatable, documented processes currently executed by humans following SOPs. Converting those SOPs into Claude Code workflows is a direct line from current enterprise AI adoption to a materially different operating model at scale.

474 of 1,078 Unicorn Founders Are Foreign-Born, From 65 Countries: India leads with 90 unicorn founders, followed by Canada (42), the UK (31), China (27), and Germany (18), among founders from 65 countries across six continents. The study covered 1,110 unicorns defined as VC-backed US companies achieving a confirmed $1B+ valuation between 1997 and 2021.

Relocating to the US Multiplies Unicorn Probability Significantly: Of 1,110 unicorns, 88 (8%) were originally founded outside the US before relocating. The relocation effect is large across multiple origin countries, with startups from several nations 2.5x to 9x more likely to reach unicorn status after moving to the US than those that stayed home.

Per-Capita Productivity Varies Widely Across Immigrant Communities: Adjusting for first-generation immigrant population size reveals striking variation in unicorn output per 100,000 immigrants across origin countries, ranging from under 3 to over 43. The highest absolute contributor (India, 90 founders) has a much lower per-capita rate than smaller communities with tighter ecosystem networks and stronger US entry points.