Begin Capital Portfolio: 28 Fund I Investments and Original Cheques

Begin Capital's first investment programme started in 2020 and ultimately made 28 investments, according to the London VC firm's own website. Its early holdings ranged from resale marketplaces and fertility products to data infrastructure, developer tools, fintech and AI software.

The most useful historical evidence is not a promotional list of unicorn prospects. It is a January 10, 2021 first-year report by co-founder Alex Menn that disclosed the actual cheques Begin Capital wrote into six companies, including the currency and the investor's role. Those figures make it possible to distinguish Begin's own investment from a startup's later whole-company financing round.

First year: six primary-source investment tickets

The rows above are Begin's disclosed individual cheque amounts, not the total capital raised by each startup. They are quoted in the original currency rather than being added into a misleading dollar total at an invented exchange rate.

The 2021 account also records an exit from Samokat during the first investment year. It does not disclose Begin's sale proceeds, ownership percentage or return multiple. An exit announcement is not enough to reconstruct an investor's realised profit.

How Begin selected its first investments

The initial six businesses span consumer products, marketplaces, enterprise software and logistics. Begin was the named lead investor for Zadaa, but a co-investor in the other five. The variability in ticket sizes supports Begin's description of using flexible first cheques, although these six records alone do not prove a consistent strategy across all 28 eventual holdings.

What else did Fund I invest in?

Begin's current official portfolio lists additional Season 1 businesses. The following names complete the initial 28-company history when combined with the original six; they are company associations, not 22 individually verified dated funding rounds.

Developer infrastructure, software and AI

Runware provides AI media-generation APIs and inference infrastructure. Codesphere develops cloud-native development and deployment tools. Streamkap works on real-time data streaming, Middleware on software observability, Cloudchipr on cloud operations, Simplyblock on storage infrastructure, and AppFactor on modernizing enterprise software. Actable AI provides data and analytics tooling, while IVM Markets applies technology to financial markets.

Enterprise operations and intelligence

Vamstar focuses on healthcare procurement. Convious supports visitor-attraction commerce. Uncrowd provides customer intelligence, Tomi.ai applies predictive analytics to marketing, Fundraise Up sells donation and fundraising infrastructure, and Thrust Carbon works on travel emissions accounting.

Finance, property, education and consumer commerce

LoanBase is property-financing technology; Kapaga addresses business finance; Native Teams provides cross-border work and payment infrastructure; Dwelly operates technology-enabled rental property management; Naduvi is a furniture and interior-products marketplace; Motley is a consumer-commerce company; and EBAC Online offers online education.

These descriptions come from portfolio company summaries and historical primary records, not a complete audited ledger. Do not assign a new deal date or Begin-specific ticket to the companies without a dated transaction source.

Larger follow-on rounds, without inflating Begin's cheques

Runware — three publicly reported rounds and a dating caveat

Runware is an inference platform integrating custom serving infrastructure with media-generation APIs. Its disclosed financing timeline is $3 million in October 2024 (Dealroom record), a $13 million seed announced on September 9, 2025 led by Insight Partners, and a $50 million Series A first reported on December 11, 2025 led by Dawn Capital. Together these figures reconcile to Runware's stated $66 million cumulative capital raised. They are company-round totals.

The Series A dating can cause duplication: Comcast Ventures published the financing on December 11, 2025, while Runware wrote its own extended article on January 29, 2026. These are reports of the same $50 million round, not separate 2025 and 2026 rounds. Begin Capital is named among existing backers, but the value of its holding or follow-on cheque is not public. Do not include Runware's 2026 date as a new event on top of the December 2025 Series A.

Codesphere — $18 million financing and a resolved founder name

Codesphere announced $18 million of financing on May 17, 2024, led by Creandum, with Begin among existing participants. Codesphere's own press announcement names the investors but does not identify Begin's individual cheque or the exact ownership retained in the company.

Some historical sources call Codesphere's CEO Elias Groll, while its current legal imprint and recent media call him Elias Schneider. Crunchbase explicitly identifies Groll as an alternative name for Schneider; the same co-founder should be matched across names, not represented as two separate entrepreneurs.

Dwelly — two 2026 financing packages, four underlying transactions

Dwelly, the AI-native roll-up of British letting agencies, announced two equity rounds and two distinct debt facilities in 2026. The mixed financing-package headlines must not be counted as equity:

Begin Capital participated in both equity rounds, alongside other investors. This does not establish Begin's undisclosed individual cheque amount. A February 2026 Fortune report identifies the £32 million General Catalyst-led equity component and £37 million debt facility. Dwelly's July company announcement explicitly separates the $95 million equity and $75 million debt financing.

Do not add the two headline package totals as venture-equity financing. The equity transactions remain separately denominated in pounds and dollars, and currency conversions require a rate tied to the actual date. VCT also stores each component as an independent, dated investment event, preventing both duplicate counts and confusion between debt and preferred equity.

How many Fund I investments are actually visible?

Begin describes 28 initial investments, but its current portfolio directory is not necessarily a complete historical, legally reconciled fund ledger. The January 2021 report names Samokat and Live Jam, which are not visible as normal current portfolio cards in the directory; Runware appears without the explicit S 01 label visible on most other early holdings. Conversely, Finom is labeled Season 2 despite not appearing in the October 2026 announcement's eight public Fund II investees.

That is why VCT uses the manager's 28-investment claim with attribution and treats the historical company list as a research reconstruction, not an audited schedule of 28 fund positions. A portfolio label may reflect the firm's website taxonomy, a fund-of-record holding, a follow-on allocation or editorial classification; these possibilities cannot be resolved without transaction-level disclosures.

A further distinction: Finom's €115 million Series C in June 2025 and its 2025 non-equity growth funding from General Catalyst are company financings. The available named participants in those announcements do not document Begin's specific capital contribution. Its website portfolio association should not be converted into a fabricated Begin-led financing event.

What does 'exit' mean on Begin's portfolio page?

Begin's website labels certain Season 1 names as exited, including Bulbshare, WOOM, Zadaa, Motley and Kapaga. Its historical 2021 report also described an exit from Samokat.

Those labels indicate portfolio changes or exit events reported by the manager, but generally not how much cash came back to the fund. This is distinct from a later startup raising fresh equity, which may change a valuation but is not itself an investor exit.

A rigorous investment history should distinguish acquisitions, secondary sales, changes in valuation, new company rounds and cash distributions to limited partners. VCT should leave proceeds and exit multiples unknown until documented.

Fund size and reported performance

Co-founder Alex Menn describes the first vehicle as beginning at $20 million and later growing to $72 million. He also reports a 22% hit rate and claims the portfolio ranks in the top 1% of its vintage.

These are first-person manager-reported statements. The publicly reviewed material does not define the hit-rate denominator and success threshold, identify the vintage benchmark, or provide independently auditable net IRR, DPI or TVPI for comparison. VCT should quote them as claims, not compute or imply confirmed fund returns.

There is a difference between a valuable investment and a successful fund: ownership, dilution, later reserves, fees and whether returns have been realised matter at least as much as impressive financing headlines.

Why this history matters for the $100 million Fund II

On October 28, 2026 Begin announced a second, $100 million vehicle focusing more narrowly on deep tech, AI infrastructure, hardware, silicon and physical AI. Its initial disclosed investments include ALL3, SurrealDB, Areg.AI, Crewline, xmemory, DesignVerse, Zencoder and Tomorrow Things, plus one unnamed company.

Begin says successful Fund I founders, including people from Runware, Codesphere and Dwelly, joined as venture partners and some became limited partners. The strategy is to use operator networks to find technical teams early and increase ownership in the strongest companies over time.

The challenge is that physical AI and hardware can demand longer validation cycles, field installations, manufacturing expertise and much larger later rounds. The firm's record in software demonstrates relevant early-stage experience, but it does not by itself validate returns in those new categories.

How to verify Begin Capital portfolio records

1. Start with Begin Capital's January 2021 report, which gives its own cheque sizes for six early holdings.
2. Check the official portfolio for programme membership; do not assign a funding date from a directory listing alone.
3. Compare later company financing disclosures, such as Dwelly's equity-and-debt breakdown, and capture the investor's actual role.
4. Verify proceeds and valuation changes independently before treating an exit or large new round as a return for LPs.

A more reliable way to track this portfolio

For each Begin investment, researchers should store: company name, Fund I or Fund II association, deal date, financing type, whole-company equity amount, debt amount when applicable, investor role, Begin's own cheque only when disclosed, and the original dated source.

The six cheque figures above are the clearest primary-source historical tickets we located. Most other amounts published for the companies are round totals, not Begin Capital investments. An accurate record of missing financial data is more useful than an invented comprehensive funding history.