Zuckerberg Just Admitted Meta's AI Agents Aren't Working

Welcome back Chronicler!

This week Mark Zuckerberg admitted out loud what a lot of us have been whispering: the AI agent hype has outrun the actual product. Meta cut thousands of jobs to fund the pivot, and its own CEO now says the results have not shown up on schedule.

Meanwhile the money did not stop, it just moved further down the stack: into energy, compute, and the unglamorous infrastructure that makes any of this possible. Two different stories, same lesson: read the fundamentals, not the demo.

Plus: how drag-along rights actually force a sale through, and a founder detection tool VCs are quietly using to win deals before they hit Crunchbase.

As always, thank you for subscribing and engaging week in week out. Still a work in progress, but I appreciate all 2,264 of you.

With Love,
Kev

THIS WEEK IN VC & TECH
What matters this week

LEAD STORY

Zuckerberg admits Meta's AI agent push hasn't delivered, months after mass layoffs

Meta cut 8,000 jobs and reassigned 7,000 more into AI groups earlier this year, betting that agentic AI would justify the reallocation, and its own CEO now says results have not materialized on the expected timeline. This is the clearest signal yet from a hyperscaler insider that enterprise AI agent ROI is running well behind the narrative VCs have been underwriting in their portfolio company pitches. Founders selling "AI agents replace headcount" should expect buyers and enterprise customers to demand harder proof points before signing, since even Meta, with unlimited compute and data, has not cracked it in six months. VCs should stress test agent startup revenue projections against a longer, not shorter, adoption curve, and treat 2026 enterprise agent bookings with real skepticism.

QUICK HITS

Sound Ventures  Ashton Kutcher is exiting the firm he co-founded 11 years ago to start a new early-stage vehicle with Morgan Beller, formerly of NFX and Andreessen Horowitz, targeting AI infrastructure, energy, and deep tech. The split signals a shift in AI capital: Sound bet big on category-leading labs, while Kutcher and Beller want the picks-and-shovels layer underneath.

Tapestry VC  The London and San Francisco firm closed its $80 million Fund III, nearly triple its predecessor, with backing from the British Business Bank, Molten Ventures, Railpen, and OpenAI CFO Sarah Friar, and opened a new London office. The close signals growing institutional appetite for Europe's repeat founder wave, with Tapestry now able to write checks up to $3 million, up from half a million at the firm's start.

Bending Spoons  The Milan-based serial acquirer of internet brands like AOL, Vimeo, and Eventbrite priced its IPO at $29 and closed its first day up nearly 40%, valuing the company at more than double its last private mark. The listing is a live test of investor appetite for a buy-and-transform legacy software model at a moment when many SaaS multiples are compressed.

India-Japan AI pact  The two countries adopted a joint AI statement and a 16-point roadmap covering GPU compute, data centers, and semiconductors, including a plan to bring 500 Indian AI professionals to Japan by 2030. It is another sign that mid-sized powers are building bilateral AI and compute alliances outside the US-China axis, widening the map of where infrastructure-focused VCs should be scouting.

PRESENTED BY OYSTER

Global hiring is changing fast. From AI's impact on HR to evolving compliance requirements and international expansion strategies, the rules are constantly shifting.

Oyster's events bring together HR leaders, founders, operators, and global employment experts to discuss what's working now—and what's coming next.

Whether you're hiring internationally today or planning for tomorrow, you'll walk away with practical insights you can actually use.

Explore Hiring Events

MONEY IN MOTION
The largest deals that you should know about

A crypto-native, privacy-first AI startup called Venice AI hit a $1 billion unicorn valuation on a modest $65 million round this week, proof that "uncensored" AI access now commands the same investor enthusiasm as frontier labs. Meanwhile an energy infrastructure company out-raised every AI model or app startup in the same window, showing the real bottleneck in the AI boom has shifted from compute to power.

Joulent  Energy infrastructure | Strategic investment, $1.75 billion | Lead: National Grid Ventures | July 1, 2026  Link

Together AI  AI infrastructure | Series C, $800 million | Lead: Aramco Ventures | July 1, 2026  Link

LeapXpert  Enterprise compliance software | Growth financing, $180 million | Lead: Riverwood Capital | June 30, 2026  Link

Venice AI  AI privacy platform | Series A, $65 million | Lead: Dragonfly | July 1, 2026  Link

Bespoke Labs  AI agentic training infrastructure | Series A, $40 million | Lead: Wing VC | July 6, 2026  Link

VENTURE BITE: One Concept in One Minute
DRAG-ALONG RIGHTS

A drag-along clause isn't there to protect the majority, it exists so a single holdout can't block an exit everyone else already voted for.

CONCEPT

A drag-along right lets shareholders holding a specified voting threshold, typically a majority or supermajority, force all other shareholders to sell their shares on the same terms if the company is acquired. Without it, a single small holder could block a sale by refusing to sign, even if the vast majority of the cap table wants the deal to close. The right converts a scattered shareholder base into someone who can actually execute a clean, 100% sale.

LEARNING OBJECTIVE

Show how a drag-along threshold determines whether a minority holder can block a sale, and calculate what a dragged minority holder receives when the vote passes.

WORKED EXAMPLE

A startup has 1,000,000 fully diluted shares. The shareholders agreement sets the drag-along threshold at 65% of voting shares. Investor A holds 400,000 shares (40%), Investor B holds 300,000 shares (30%), the Founder holds 200,000 shares (20%), and Angel C holds 100,000 shares (10%). An acquirer offers $10.00 per share, or $10,000,000 total.

Step 1: check whether the deal can be dragged. Investor A (40%) and Investor B (30%) vote yes, totaling 70% of voting shares, above the 65% threshold.

Step 2: because 70% exceeds the 65% threshold, the drag-along clause is triggered. The Founder and Angel C, who together hold the remaining 30% and did not consent, are contractually forced to sell their shares on the same terms.

Step 3: calculate each holder's payout at $10.00 per share. Investor A: 400,000 x $10.00 = $4,000,000. Investor B: 300,000 x $10.00 = $3,000,000. Founder: 200,000 x $10.00 = $2,000,000, forced to sell despite objecting. Angel C: 100,000 x $10.00 = $1,000,000, forced to sell despite objecting.

Total paid out: $10,000,000, matching the deal price exactly, confirming every share transfers at the same per-share price regardless of consent.

Quiz (Difficulty: Medium)

AI TOOLKIT FOR VCs
Our curated stack for the modern investor

Subscribe to The AI Report, the free 5-minute daily AI brief for 400,000+ business leaders, and you’ll get 37 Claude prompts free in your welcome email. They’re organised by the 8 situations every manager faces. You get both: the newsletter and the prompts.

Subscribe & Get the Prompts

THIS WEEK’S STANDOUT

🕵️‍♂️ Evertrace

Evertrace is a founder detection engine that surfaces pre-seed and stealth-stage founders before they appear in any startup database, monitoring trade registries, GitHub, patents, grants, domain registrations, and social signals in real time. More than 200 VC funds, including Atomico, Creandum, Cherry, and Heartcore, now use it, making it one of the fastest-adopted deal sourcing tools this year. For VCs it matters now because the edge in sourcing has shifted from finding good companies to finding them first, before mainstream databases catch up.

🚣🏻‍♀️ Rowspace  -  AI platform backed by Sequoia and Emergence that turns messy internal memos, models, and documents into structured intelligence for due diligence and portfolio monitoring.

Ⓜ️ Standard Metrics  -  Portfolio monitoring platform that lets investors ask natural language questions against portfolio company financial data collected across the fund.

💥 VC Boom  -  Scores pitch decks in under 90 seconds and matches founders to relevant investors, giving VCs a fast filter for inbound deal flow triage.

FROM X
Our favorite post from this week

PRESENTED BY LEVANTA

Most brands running external traffic aren't scaling — they're just spending. Levanta's free playbook breaks down 7 proven strategies: what works, what bleeds budget, and how top brands drive millions in off-Amazon revenue.

→ Grab the FREE playbook

THE BOARD
VC jobs, curated for you

Here's what you've been waiting for: exciting job opportunities! To maximize your chances, remember these key tips:

That’s it for today, hope you enjoyed this as much as I did curating it, see you next week!
-Kev