What happened
Begin Capital announced the start of what it calls its second “season”: Fund II, a $100 million fund aimed at founders building in deep tech. The announcement describes the firm as having begun in 2020 as a two-person micro-fund with $20 million. It says that first vehicle ultimately became a $72 million fund.
The firm presents Fund II as a more concentrated investment thesis. Its stated targets are companies addressing difficult technical problems in AI infrastructure, hardware and physical AI. That is a narrower description than the portfolio language on Begin Capital’s website for its first fund, which says Season 1 invested across industries and genres. The site describes Season 1 as 28 investments and says the firm led or participated in Seed and Series A rounds, writing checks from $500,000 to $7 million.
For the new vehicle, Begin Capital says it will continue to invest globally, with a focus on European technology talent. Its website states a pre-seed-to-Series-A mandate and a flexible ticket range of $500,000 to $8 million. The announcement also says the fund has already made nine investments: All3, SurrealDB, Areg, Crewline, xmemory, DesignVerse, Zencoder and TomorrowThings are named in the supplied account, alongside the statement that the investments were made with FirstMark, Initialized and Fly Ventures.
The operating approach described by the firm is to make relatively small, flexible first checks, then increase ownership in companies that emerge as winners. Begin Capital frames that approach as moving quickly and backing founders early. It also says its first fund is in the top 1% of funds of its vintage on its principal metrics and reports a 22% hit win ratio, but the supplied material does not define that ratio, identify the benchmark, or provide independently comparable performance data.
This is therefore a financing and mandate announcement, not an announcement of a new AI product, model, benchmark, or deployment. The concrete news is the $100 million Fund II and the sectors Begin Capital says it intends to finance. Its public site invites applications from teams and describes the firm’s investment range and stages; it does not set out a public application deadline, allocation schedule, or a dedicated program for the new fund.
What the archive adds
We have not covered this before; it opens a new thread.
Trend signal
No supplied pattern is a sufficiently direct fit. The announcement records an investor’s stated focus, not evidence that a particular operating or technical pattern has become durable.
Why now
The source gives no market explanation for the timing beyond the firm’s account of its first fund. Begin Capital says the first vehicle gave it confidence and that Fund II is intended for a focused set of deep-tech areas. It does not state a close date, limited-partner roster, fundraising process, or external trigger for the launch.
Connections to other events
The archive includes a prior Field Note on Eka, a company associated with physical AI and embodied-task completion: Eka is worth $1.45B thanks to a physical AI model that has experts raving. The connection is limited but real: Begin Capital explicitly includes physical AI in Fund II’s target areas. Eka is not named as a Begin Capital investment, and the supplied material does not establish any commercial relationship between them.
Why it matters
For founders working on AI infrastructure, hardware, or systems that operate in the physical world, Fund II expands the set of investors explicitly seeking these categories at early stages. The stated pre-seed-to-Series-A range matters more than the headline fund size for an individual company: it indicates that Begin Capital says it can participate from an early first check through later ownership-building.
The announcement also shows how the fund is choosing to describe deep tech. It is not treating AI as one broad software category; it separates infrastructure, hardware and physical AI, while saying it values both technical backends and proven industrial applications. Teams seeking capital can use that framing to test fit: identify the hard technical constraint, the application environment, and the evidence that makes the hypothesis measurable. That does not guarantee financing, but it is more actionable than a generic “AI fund” label.
What happens next
- Begin Capital will disclose additional Fund II investments in its stated target areas. (6 months, 65%). Watch for: portfolio additions or investment announcements naming AI infrastructure, hardware, or physical-AI companies.
- The firm will make follow-on investments after its initial checks in selected portfolio companies. (12 months, 55%). Watch for: later financing rounds that identify Begin Capital as an existing investor.
Limits and uncertainty
The supplied material is an announcement from Begin Capital and should be read as the firm’s own account of its vehicle, strategy and track record. It does not provide fund documents, limited-partner information, fees, capital-call status, deployment pace, or an audited performance report. The public site supports a second-season deep-tech mandate, the investment stages and the ticket range, but it does not independently detail the $100 million figure in the supplied homepage excerpt.
The named first investments are not accompanied here by dates, check sizes, ownership levels, company descriptions, or evidence that each one fits a particular target category. Nor does the source establish how much capital has been committed or deployed. The reported performance claims for Fund I cannot be evaluated from the supplied information because the measurement method and comparison set are absent.
Skill to build now
If raising an early deep-tech round, prepare a one-page evidence map before contacting investors: state the technical bottleneck, the specific application, the experiment or customer proof behind the claim, and the numbers that will change if it works. Begin Capital says it prefers hypotheses that are tested and translatable into numbers, and it names both technical depth and industry application as part of its Fund II thesis. This exercise will also reveal whether a company fits an infrastructure, hardware, or physical-AI investor category.
Reader takeaway
Begin Capital’s Fund II is a $100 million early-stage deep-tech vehicle whose stated focus is AI infrastructure, hardware and physical AI. The practical implication is not a new capability for builders today, but a clearer investor-fit signal for teams with technically difficult, measurable products in those areas.
Open question
How much of Fund II is already committed or deployed, and what evidence will Begin Capital use to distinguish infrastructure, hardware and physical-AI investments in practice?