NVIDIA was hiring for a platform months before it bought one.
Our analysis of NVIDIA's job postings, as of August 1, found the company staffing up around AI agents. The roles cover a dedicated security team for AI agents, infrastructure for identifying and managing them, and software that runs agents directly on your device, which one listing calls "the foundation of the desktop AI operating system."
Those roles were open before the Hugging Face acquisition was announced. The deal gives them something to build on.
Since 2024, NVIDIA has acquired 15 AI startups, tying Accenture for the third most active behind OpenAI (22) and Salesforce (18).
The mix is lopsided:
9 deals focused on making AI infrastructure faster or cheaper to run, including Run:ai ($700M) and Enfabrica ($900M)
3 deals focused on data, including Gretel and Kumo ($400M)
1 developer tools deal and 1 industry-specific application
And then Hugging Face, which is the only one that’s a platform where developers actually go.
Run:ai's own data showed most GPU clusters ran below 25% utilization before its software. Gretel generated training data. Prior deals made something NVIDIA already owned run faster or cost less. Hugging Face is where developers show up before they ever touch NVIDIA hardware.
PayPal gave us the PayPal mafia. SpaceX may be building the industrial version.
We searched CB Insights People data for former SpaceX employees who have moved into senior roles elsewhere. They’re spreading across some of the hottest areas in industrial tech.
Doug Bernauer spent nearly 12 years at SpaceX before founding nuclear microreactor startup Radiant.
Maximilian Benassi worked on SpaceX’s Raptor engine before co-founding satellite manufacturer Apex.
Jordan Black worked in avionics R&D and manufacturing at SpaceX before starting aerospace wiring company Senra Systems.
And former SpaceX satellite VP Rajeev Badyal now leads technology for Amazon’s Project Kuiper.
Those are only a handful. For anyone scouting aerospace, defense, energy, or industrial tech, following the people leaving SpaceX may be one of the easiest ways to find what’s coming next.
See where other SpaceX alumni are going with our people search.
One company’s reported $400M growth plan.
We interview hundreds of startup executives every month. They tell us things that don’t always make it into press releases.
Galgo is one of the more interesting ones. It finances motorcycles for gig workers and informal earners across Latin America, people banks typically reject. Reports put its target at roughly $500M in revenue within five years, up from about $100M today. But how does Galgo plan to get there?
“As Galgo grows, every additional loan and dealer generates more data, improves underwriting, reduces costs, and strengthens our offer. That compounding advantage makes the platform smarter, more competitive, and increasingly difficult to replicate.”
Laszlo Godor,
VP of Capital Markets, Galgo
The company plans to launch a gig-economy lending product with Uber, enter Brazil, and use AI to approve more customers without taking on more credit risk.
But growth isn't guaranteed. Even Godor names the risk directly:“The main uncertainty is execution during rapid growth. Expanding into Brazil and entering new customer segments require disciplined underwriting, strong local execution, and careful control of credit risk. Competition is also increasing as more digital credit and rental companies recognize the opportunity.”
That gives us something concrete to evaluate: where the growth is supposed to come from, what has to go right, and what would derail it. And it gives dealmakers something better than another funding announcement.
CB Insights' emerging technology insights platform provides all the analysis and data from this newsletter. Our data is the easiest way to discover and respond to emerging tech.
