Sponsored by
Welcome back Chronicler!
This week Anthropic filed to go public, and the structure of that filing matters as much as the revenue behind it. Meanwhile the money kept flowing to the biggest bets, with a one-year-old lab landing a $1B Series A and a Dallas firm closing a $10B fund.
Plus: a $1B first-round check for a one-year-old lab, and Beady AI's daily portfolio screening for funds.
As always, thank you for subscribing and engaging week-in week out. Still a work in progress, but I appreciate all 2,375 of you.
With Love,
Kev
THIS WEEK IN VC & TECH
What matters this week
LEAD STORY
Anthropic Officially Files for Its Public Debut - DEEP DIVE by Venture Chronicle
Anthropic's first-half 2026 revenue was $16.3B, about 14x a year earlier, and it booked an $80.7B net loss that is almost entirely non-cash. Figures below come from press summaries of the filing, not the filing itself (see Sources).
Let’s start with a graph, Anthrophic’s compute bill as a percentage of revenue has been on a downward slope, strong indication of top line uptake, but for how long and is this sustainable?
Numbers to know
Metric | Figure | Why it matters |
|---|---|---|
H1 2026 revenue | $16.3B (vs $1.16B in H1 2025) | 14x year on year, and 3.5x all of FY2025 ($4.6B) |
Q2 2026 revenue | $11.6B | Implies Q1 near $4.7B and a run rate near $46B (my math) |
H1 net loss | $80.7B | $72.4B is non-cash marks on convertible notes, derivatives and SAFEs |
H1 operating loss | $693M | About 4% of revenue; Q2 alone reportedly showed $974M of operating income |
H1 adjusted net income | $2.7B | Non-GAAP; adjusted operating income was $1.6B, a 10% margin |
H1 compute and infrastructure cost | $8.9B | About 55% of revenue (my math); reportedly 48% in Q2, down from 251% in Q1 2025 |
Cash and short-term investments | $131.7B at June 30 | $112.4B cash plus $19.3B short-term; $141.1B raised in total |
Customers spending $100K+ | 11,900+ | Up from about 950 a year earlier |
Net revenue retention | About 900% | Existing customers spent roughly 9x what they did a year before |
Concentration | Top five customers about 25% of revenue | Amazon, Google and Microsoft channels reportedly 48% of H1 revenue |
Control | 50.1% of votes, held by a 7-person Founder LLC | Via one Class F share; public Class A gets one vote per share |
Listing | Nasdaq, ticker ANT | Goldman Sachs and Morgan Stanley lead 22 banks; price, size and valuation are blank |
What the numbers mean
The headline loss is an accounting artifact, and a flattering one. Of the $80.7B net loss, $72.4B is non-cash fair-value marks on convertible notes, derivatives and SAFEs. Those liabilities grow when Anthropic's own valuation rises, so the loss is partly a scoreboard of the markup. Strip it out and the operating loss is $693M on $16.3B of revenue.
Operating leverage is the real story. Revenue grew about 14x year on year while compute and infrastructure cost grew about 4x (300%, per one summary). That gap is why Q2 reportedly flipped to operating profit. The open question is whether it holds once the next generation of models needs far more compute per dollar of revenue.
Revenue is consumption, not seats. Net revenue retention near 900% and a jump from about 950 to 11,900 customers above $100K means existing customers keep expanding usage. That compounds fast on the way up. It is also revenue that can be reallocated quickly, which public investors will price as volatility.
There is no gross margin to anchor on. One summary says compute is booked as an operating expense rather than cost of revenue, so the filing presents no gross margin. Expect analysts to build their own, and expect the debate over what counts as cost of goods to be the first valuation fight.
Control versus ownership
Public investors would buy economics with no path to control: one share held by seven founders carries 50.1% of the vote, and another elects most of the board. Per the summaries, there are five share classes:
Class | Holder | Votes and economics |
|---|---|---|
A | Public | One vote per share, full economics |
N | Mostly strategic partners; 5% of fully diluted shares reportedly goes to the Anthropic Public Benefit Fund | Non-voting except where law requires, full economics |
T | Long-Term Benefit Trust (one share) | Elects a board majority, no economics |
F | Founder LLC of seven co-founders (one share) | 50.1% of general voting power, no economics |
E | Employees (planned post-IPO) | None outstanding at listing |
The unusual part is not the dual-class structure. Meta and Alphabet founders hold control alongside large economic stakes. Here, control sits on shares that carry no economics at all, so it cannot be diluted by selling stock. Class A holders could own nearly all the equity and still hold at most 49.9% of the vote, with the board chosen by a trust.
The filing also flags the tension: public benefit duties may conflict with stockholder interests. For anyone writing a term sheet, this is the clearest template yet for decoupling governance from capital.
What it means for VCs and founders
ANT becomes the comp for every late-stage AI mark. The last private round, Series H in May, was reported at $965B. Press reports (not the filing) put the IPO target at $1.5T to $2T, a 1.6x to 2.1x step-up in about six months. Expect every late-stage AI portfolio mark to be re-underwritten against where ANT trades.
Governance becomes a priced term. If public markets accept control decoupled from capital, other labs will ask for the same structure. Investors should decide now what a dollar of economics is worth without a vote, rather than negotiate it deal by deal.
Circularity is now a disclosed risk. Cloud providers are investors, suppliers and customers at once. The filing lists those conflicts of interest as a risk. Anyone underwriting AI infrastructure startups should ask how much of their demand traces back to the same few balance sheets.
The IPO will mint operators and angels. Stock comp includes a $2.9B catch-up at IPO and $23.7B more over roughly 2.1 years. Newly liquid employees fund and found the next wave of startups, and they know exactly what frontier-lab economics look like. That is a sourcing edge for seed funds, and a competitive threat to the median one.
Risks and what to watch
The biggest risk is the gap between cash and commitments. Anthropic holds $131.7B of cash and short-term investments, while named compute contracts with Google, Amazon, Microsoft, Broadcom, SpaceX, Nscale and Lambda sum to roughly $600B or more (one author's math; the filing leaves the aggregate blank), running as far out as 2045. The company also states it may not achieve or sustain profitability, and cites catastrophic risks from its own technology.
Amended filing: price range, share count, raise size (press reports say about $100B) and valuation are all blank today.
Compute total: whether an aggregate commitment figure appears in an amendment.
Q3 numbers: the filing omits them, and a second profitable quarter would validate the operating leverage story.
Note conversion: how the convertible notes and SAFEs settle at the offering, and whether more non-cash charges follow.
Customer concentration: whether the top five customers and the cloud channels stay near 25% and 48% of revenue.
The first great AI IPO is selling shares that come with almost no say.
QUICK HITS
Disruptive's $10bn Fund Disruptive is dropping its deal-by-deal SPV model for a single $10bn fund that makes about 10 bets of roughly $1bn each in late-stage AI and defence. When a non-hub firm can write $1bn checks and decide for itself, SPV-dependent syndicate investors lose their route into the best late-stage deals.
Applied Materials Doubles Down on Besi Applied is doubling its stake in the Dutch hybrid-bonding specialist to about 20%, pending US antitrust review, with no board seat sought. It is a strategic claim on advanced AI-chip packaging that stops short of the 30% Dutch bid threshold.
Automattic's Layoffs The WordPress parent is cutting staff after a failed attempt to oust its CEO led to a full board turnover and several executive exits. Governance fights at a major open-source steward are a platform risk for the many startups built on WordPress.
China's AI+ Manufacturing Push On October 19 China's Ministry of Commerce and eight other departments issued guidelines to expand its "AI+ manufacturing" campaign and boost supply of smart devices. State demand for AI-enabled hardware is a tailwind for Asian hardware startups.
PRESENTED BY ATLASSIAN
State of Product reveals what AI still hasn’t solved
80% of product professionals say AI helps them ship faster, but customers aren’t seeing value any sooner. Atlassian’s State of Product 2027 explores the gains AI is delivering and the challenges that remain, from decision-making that hasn’t kept pace to gut instinct overriding customer evidence.
MONEY IN MOTION
The deals that shaped the week
Flapping Airplanes landed a $1B Series A at a reported $6B valuation, a first-round check size that used to be reserved for late-stage companies. It signals that investors are now pricing elite research teams, not revenue, as the scarce asset.
Flapping Airplanes | Frontier AI research lab | Series A, $1B at $6B valuation | Index Ventures, Lux Capital | Oct 21, 2026 Link
Bedrock Robotics | Autonomous construction equipment | Series C, $600M at $4.1B valuation | CapitalG, Emergence Capital | Oct 26, 2026 Link
Otter.ai | AI meeting tools | Growth round, $320M | General Catalyst (Growth fund) | Oct 20, 2026 Link
Warp Sciences | AI for drug chemistry | Series A, $215M | Lightspeed Venture Partners | Oct 22, 2026 Link
Whop | Online selling platform | Series D, $200M at $5.1B valuation | Sequoia, Shopify Ventures | Oct 21, 2026 Link
Pattern: money concentrated in AI at every layer, from research labs (Flapping Airplanes) to applied science (Warp Sciences) and physical AI (Bedrock Robotics). Mega-sized early rounds next to growth checks for revenue-proven companies like Otter.ai and Whop show capital chasing both the frontier and proven revenue.
AI TOOLKIT FOR VCs
Our curated stack for the modern investor
Leave Granola and get up to 12 months free of Wispr Flow Notetaker + Dictation
If you have paid time left on an individual Granola plan, we'll match it with a Wispr Flow subscription that includes Notetaker and dictation, and add bonus time, up to 12 months total. Sign in or create a Wispr account and submit proof of your plan to check eligibility.
THIS WEEK'S STANDOUT
Beady AI launched a risk screening and daily portfolio monitoring service built for VC and investment funds. It screens founders, companies and funds against sanctions, PEP, court records and adverse media across 100,000+ sources, maps ownership, then re-screens portfolio companies daily with alerts. It matters now because LP and regulatory pressure on KYC and diligence is rising, and it automates a slow manual step (the vendor claims up to 95 percent fewer false positives, unverified).
Altvia Investor Engagement Manager - Tracks each LP across every fund, flags weakening or rising relationships and drafts cited next steps for fundraising and IR, with human approval.
S&P Global iLEVEL Agent - Adds a native AI agent, AI data ingestion and natural-language search to iLEVEL, plus an MCP server connecting private markets portfolio data to Claude and Copilot.
FinBursa 360 - An agentic front office for PE, VC and private credit with agents across CRM, deal management, data rooms, fundraising and research, requiring human approval before outreach.
Presentations.AI - A beta Presentation Refresh regenerates investor-update decks from Snowflake, Salesforce and HubSpot data, useful for portfolio reporting.
FROM X
This week's best take
PRESENTED BY KALSHI
The goat-herding company that hedged a wage law
For US businesses with big payrolls. Western Grazers in California hedged a state wage-law change on Kalshi. If a new rule would raise your costs, you can hold a position that it passes, so the payout offsets the hit. Federally regulated, from a few dollars.
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:
Apply early and tailor your application to each specific job posting.
Research the Venture Capital firm thoroughly, including their portfolio companies and investment thesis.
Customize your CV to highlight the strengths most relevant to the role.
Go the extra mile: Identify a startup that aligns with the VC's portfolio and investment strategy. Then, craft a two-page memo explaining why you believe it would be an excellent fit for their investment
🔥Must Apply | 🌱Impact Investing | 🧑💻Remote Role | 🥳Easy Apply
🇺🇸 Naver Ventures - Investment Analyst
🇺🇸 Future Back Ventures - Managing Partner
🇺🇸 Amex Ventures - Portfolio Manager 🔥🔥
🇺🇸 Capital One Ventures - Manager 🔥
🇺🇸 Ericsson Ventures - Vice President
🇺🇸 Samsung Next - AI & Robotics Investor 🔥🔥🔥
🇺🇸 Lilly Ventures - Sr. Director 🔥
🇬🇧 Lockheed Martin - Venture Capital Associate
🇫🇷 Bpi France - Fund of Funds Investor 🔥🔥🔥
🇩🇪 First Momentum Ventures - Visiting Analyst
🇩🇪 MarktLink Capital - Investment Associate
🇳🇱 Plug & Play Tech Center - Senior Associate 🔥🔥
🇯🇵 Plug & Play Tech Center - Principal 🔥🔥🔥
How was the newsletter today?
That’s it for today, hope you enjoyed this as much as I did curating it, see you next week!
-Kev