Analysis
Pulse covered a16z's $1.1 billion Machine Age Fund when it closed on Aug. 28 -- the firm's first vehicle dedicated entirely to the physical layer of AI: semiconductors, memory, networking equipment, storage, data centers, robotics and connected hardware. What's worth adding now that the initial headline has cooled is the context that actually explains why this fund exists: hardware now represents more than 20% of a16z's overall deal flow, up from a negligible share just a few years ago, per TechCrunch's original reporting. That is not a diversification experiment sitting alongside the firm's core software practice -- it's a formal acknowledgment that a meaningful and growing share of what a16z is already funding no longer fits the software-first playbook the firm built its reputation on.
The fund's stated goal, in partner language, is to "open the throttle and accelerate the physical buildout of AI." In practice that means checks into the same category Pulse has tracked all month racking up outsized rounds:
- Castelion — $1 billion Series C at a $13 billion valuation
- Base Power — $1 billion Series D
- Etched — chip round that doubled its valuation to $21 billion
Those are the exact deal shapes a dedicated hardware fund is built to write more of, faster, without competing internally against a16z's software funds for allocation.
The genuinely new information here, beyond the fund's existence, is the deal-flow statistic: more than one in five deals now crossing a firm built on software-era thinking are physical-AI or hardware-adjacent. That is the number that should reset how founders pitch a16z going forward -- a hardware-heavy pitch is no longer a mismatch for the firm's model, it's increasingly the median deal in the door.
What to watch next: whether the Machine Age Fund's first checks land in already-hot categories like Castelion's hypersonics or Base Power's grid batteries, where a16z is competing against Carlyle, JPMorgan and other growth-stage capital already active in defense and energy hardware, or whether it goes earlier-stage into component-level bets -- chips, memory, networking silicon -- where a16z's software-era relationships with hyperscalers could actually be a differentiated edge over traditional industrial investors.
