Blackstone bets one part in 48,000 of its assets on a startup founded in 2025

$27m set against more than $1.3 trillion of assets under management comes to about 0.002 percent, one part in roughly 48,000 ^[16] -- a scale at which the position could be written off entirely without moving Blackstone's own numbers. That is what makes it worth reading rather than dismissing. A cheque this size cannot buy returns at fund scale; it buys sightlines into a category early enough to matter when the growth vehicles are writing amounts that do move a balance sheet, and Globes calls this the firm's first early-stage cybersecurity investment ^[19], which is another way of saying someone decided deal sourcing is a line item worth funding.

Okta Ventures and Zscaler Ventures both came in with commercial partnerships attached ^[7], and the cap table reflects that: the product story is that Huskeys sits on top of the network and security tools an organisation already runs rather than replacing them ^[8]; Shlomo Kramer and Eran Reshef, two of the people credited with originating the web application firewall, are also in ^[12], and the firewall in question, per Calcalist, is a nearly 30-year-old Israeli technology that has changed relatively little since ^[10].

Then the arithmetic of the company. Total funding is $35m and this round is $27m of it ^[1], so everything raised before came to $8m ^[14] and 77 percent of the money Huskeys has ever taken arrived in one go ^[15], from a company founded in 2025 ^[3]. Neither announcement attaches a valuation, a revenue figure or a headcount to any of it ^[13]. What is disclosed instead is a customer roster (TikTok, Legoland, Ro, Cava and Hugging Face, per Globes ^[17]) and a category name of the company's own coinage, Network Edge Security Management ^[9].

Calcalist's version of that roster swaps Cava for Blackstone ^[18]. That detail changes what the money means: not simply a customer funding its supplier, but a customer buying optionality on infrastructure it has already decided to depend on. The demand case underneath rests on one estimate, that non-human traffic has already passed human traffic and reaches 70 percent of the internet by 2027 ^[20], leaving the human share at 30 ^[22]; Gafni's framing is that most identities online will soon be autonomous agents needing infrastructure that can identify and manage them ^[11].

This is probably wrong, but I read the $27m as a sourcing fee rather than an alternatives manager going into competition with seed funds. It could play out a few different ways. Blackstone repeats within a few quarters and puts a named team behind it, which makes it a programme. The Okta and Zscaler partnerships turn into real distribution and Huskeys gets bought before any growth round, which makes the strategic money the money that mattered. Or the agentic traffic estimate slips, and blocking exploitation at the edge within minutes ^[21] ends up a feature inside somebody else's platform. Two more early-stage cyber cheques would prove me wrong, and I would rather be shown that than be right about a one-off.