The headline from the week of 18–24 July 2026 wrote itself: Atoms, the physical-AI robotics company founded by Travis Kalanick, raised $1.7 billion in a growth round led by Andreessen Horowitz.

Crunchbase's list of the week's ten biggest rounds is worth reading past the first line, though, because three of the ten were security companies.

The ten

#CompanyRaisedStageValuationWhat it does
1Atoms$1.7BGrowthPhysical AI robotics
2Meshy AI$400MSeries B$1.5B3D AI generation
3Sila$300MGrowthBattery technology
4Etched$300MSeries C$10B pre-moneyAI inference chips
5Augustus$180MSeries B$1BFintech
6Cathedral$160MGrowth$14B reportedDefence / military cyber
7Crystalys$130MSeries BBiotech
8Candid Health$120MSeries DHealthcare revenue software
9Glow$100MAI endpoint security
10Neo Security$100MGrowthEnterprise software control

Cathedral, Glow and Neo Security account for $360 million across the three.

Why that matters more than the sector total

We looked recently at cybersecurity funding holding at $10.6bn for the half while Q2 fell about 30%. The conclusion there was that the total was being carried by a small number of very large rounds while deal counts fell — a barbell.

This week is that pattern in a single frame. Security did not appear as a sector having a good week. It appeared as three individual companies, in a top ten otherwise dominated by AI and hardware, taking nine figures each.

Note the investor names attached: Cathedral led by Sequoia and Andreessen Horowitz, Glow by Sequoia, Cyberstarts, Greenoaks and Redpoint, Neo Security by Bessemer and Andreessen Horowitz. These are the same firms writing the AI cheques. It is not a separate pool of security capital; it is the same capital, occasionally pointed at security.

Cathedral's number is the one to sit with

$160 million at a reported $14 billion valuation is an unusual shape. The round is small relative to the valuation — roughly 1% — which typically indicates a company that does not need the money and is raising on terms it likes, rather than one funding a burn plan.

The category is defence and military cyber. That places it in a segment with government customers, long procurement cycles and, at the moment, considerable political tailwind.

For founders, the useful read is not "defence tech is hot." It is that valuation and round size are decoupling, and a large valuation attached to a modest raise is a different signal than the same valuation attached to a large one.

What Etched says about the AI stack

$300 million at a $10 billion pre-money, from Sequoia, for AI inference chips.

That is the second-highest valuation in the table, for a company selling silicon rather than software. It fits a pattern visible in the same week's list — Sila on batteries, Atoms on robotics — where the largest cheques went to companies with physical products and heavy capital requirements.

Crunchbase's own framing quotes the thesis behind the Atoms round: the coming industrial revolution in which large industrial sectors get completely digitised.

Whether that thesis is right is a separate question. What is observable is where the money went, and it went to atoms rather than bits with unusual consistency for one week.

What founders should take from this

  • A sector total tells you nothing about your round. Security "appeared" in this list three times, and all three were at a stage most companies reading this will never reach. The early-stage market is a different market.
  • The same funds are doing both. If a16z is writing a $1.7bn physical-AI cheque and a $100m security cheque in the same week, "we're a security company" is not a differentiator to that firm. What you are doing that is hard is.
  • Watch round size against valuation. Cathedral raising 1% of its valuation and a company raising 25% of its valuation are in completely different negotiating positions, and the headline number obscures which is which.
  • One week is one week. This is a snapshot, not a trend. It is useful for seeing shape — who is writing cheques, at what stages, in what categories — and close to useless for predicting next quarter.

The most reliable observation in the table is the least exciting one: at the top of the market, capital is available and concentrated. Neither half of that sentence describes what it is like to raise a Series A.