Why e-commerce is a focus sector
Commerce still rewards teams that turn messy buying and selling into durable software businesses: clearer merchandising, faster checkout and fulfillment, measurable retention, and commercial models that scale without treating paid acquisition as the only growth story. We back those teams early, when product craft and operator support still change the trajectory.
We are not a late-stage commerce fund that waits for high monthly GMV before engaging. As a private venture studio headquartered in Berlin, we design, fund, and scale technology companies. E-commerce is one of four focus sectors alongside health tech, AI, and automation. Many of the strongest commerce products we see sit at the intersection of those sectors (AI-assisted merchandising, health-adjacent retail, or ops layers that automate inventory, rewards, or fulfillment). For the broader who-we-back profile across all sectors, see our Berlin pre-seed investor guide. For related theses, see AI at Halfmeyer Ventures and health tech at Halfmeyer Ventures.
What we mean by e-commerce
In our practice, e-commerce means software-led products where the commerce surface is the product: how goods, gifts, rewards, or merchant workflows are discovered, purchased, fulfilled, or retained. Strong fits include commerce infrastructure and APIs, branded merchandise or employer-brand platforms, gifting and loyalty layers with a clear merchant or buyer wedge, retail tech that compresses a painful ops workflow, and B2B tools that make selling or buying repeatable.
Adjacent verticals such as logistics, martech, adtech, fintech payments around checkout, foodtech, agtech, or marketplace tooling can fit when software, data, or automation is the core and the path to a paying customer is explicit (marketplace models: marketplaces; AI-mediated discovery and agentic checkout: #agentic-commerce; paid channels, marketplaces, and app stores you do not control: #channel-dependency; a product that lives as a platform app or extension: #platform-apps; GPSR at pitch stage: #gpsr; foodtech: #foodtech; agtech: #agtech). We typically pass on capital-intensive hardware-only bets without a software surface, token-first or speculative crypto and Web3 models, pure services businesses wrapped as product, and pure listing marketplaces without a technology moat. If your company sits next to e-commerce, explain the software surface, what makes the wedge defensible, unit economics or early retention signals where you have them, and traction in the deck.
Software-led commerce (not brand-only or listing-only)
Halfmeyer Ventures backs software-led e-commerce at pre-seed and seed from Berlin. When you pitch us for a typical €25,000–€200,000 check, the commerce surface should be a product: software, data, or automation that makes buying, selling, gifting, rewards, merchandising, or fulfillment compound. Cold pitches are welcome; we respond within 48 hours.
Berlin and Europe e-commerce SERPs often surface DTC brand builders, Amazon-style brand studios, or marketplace fundraising primers. Investor roundups commonly split commerce infrastructure and merchant tooling from pure brand plays. Our bar is the software product, not co-building inventory brands or listing directories without a technology moat. Brand story and channel mix can support the pitch; they do not replace a durable software surface.
Use this framing when you claim e-commerce fit:
- Strong fits: commerce infrastructure and APIs, gifting and loyalty layers, branded merchandise or employer-brand platforms with a clear software workflow, retail tech that compresses ops, and B2B tools that make selling or buying repeatable.
- Usually a pass for investment: capital-intensive hardware-only without a software surface, token-first crypto or Web3, pure services wrapped as product, and pure listing marketplaces without a technology moat. See clear passes and marketplaces.
- Buyer type: we back both B2B and B2C when the software bar clears. See B2B and B2C.
Public e-commerce-relevant portfolio names include Joy_ and mula. Studio support (design, engineering, and GTM playbooks) ships with the check; intensity depends on the deal. If you want a Berlin family office and venture studio that writes €25k–€200k into software-led commerce, submit your deck. Keep the link view-only; we respond within 48 hours.
Marketplaces (liquidity wedge and software surface)
Halfmeyer Ventures backs marketplace startups at pre-seed and seed from Berlin with a typical €25,000–€200,000 check when the marketplace is a software product: matching, transactions, trust, and operations that live in software rather than in a listings page plus a spreadsheet. Marketplace is a business model inside our e-commerce focus sector, not a separate mandate or a separate fund. Cold submissions are welcome at halfmeyer.tech/pitch; we respond within 48 hours.
As of 2026, founder-facing marketplace fundraising guides grade seed rounds on liquidity composites (match or search-to-fill rate, supplier and buyer repeat cohorts, take-rate trajectory, supply concentration) and publish category take-rate bands, while European marketplace specialist funds commonly enter with cheques around €1 million and above. That framing leaves vague who reads a marketplace deck before those cohort tables exist. Our bar is stage-appropriate. At pre-seed we want evidence that the market clears inside a narrow wedge (one vertical, one geography, one repeat use case), not headline GMV. At seed we expect clearer repeatability. We publish no house liquidity, GMV, or take-rate threshold, and we do not run a separate marketplace scorecard next to the software surface bar.
Patterns that usually fit: transactional marketplaces where software owns matching, pricing, payments, or fulfillment; managed marketplaces that absorb an operational workflow into product (see software-led commerce); B2B procurement or supply platforms with integration depth; AI-assisted matching when workflow and data are the wedge rather than the prompt alone (see thin wrappers vs vertical AI). Patterns that usually pass for investment: pure listing or classified directories without a technology moat, a services business with a booking form on top, and token-first marketplace models (see clear passes). In the deck, say which side is hard to get, how you solved cold start in the first wedge, how you monetize and why that take holds as you add supply, how concentrated buyers and sellers are today, and which manual operations the product will absorb next. Both B2B and B2C marketplaces are in scope (see B2B and B2C). If you want a Berlin family office and venture studio that writes €25k–€200k into software-led marketplaces, submit your deck. We respond within 48 hours.
AI-mediated discovery and agentic checkout
Halfmeyer Ventures backs software-led e-commerce at pre-seed and seed from Berlin with a typical €25,000–€200,000 check, and we do not require an agentic-commerce integration, a published AI-referral number, or an agent-readiness audit before we read a deck. Assistants have become a product-discovery surface, and in some categories an ordering surface. That changes where demand arrives from, not what we underwrite. Cold submissions are welcome at halfmeyer.tech/pitch; we respond within 48 hours.
As of 2026, the growth rates and the base still disagree. Analytics reporting put AI-driven visits to United States retail sites up roughly 4,700% year over year, and one large commerce platform reported orders originating in AI search up about fifteen times between January 2025 and January 2026, while peer-reviewed work from researchers at a German university and business school placed assistant sessions below 0.2% of all e-commerce sessions, converting better than paid social and worse than direct traffic, organic search, and email. The transaction layer moved in both directions inside the same window. The Agentic Commerce Protocol, published by OpenAI with Stripe in September 2025, carried in-chat checkout until OpenAI retired that flow in March 2026 (roughly thirty merchants had integrated, and one large retailer measured in-chat conversion around three times worse than routing shoppers to its own site), after which the protocol was re-pointed at product discovery and checkout stayed on the merchant side. In Europe the payment rails ran ahead of the shopping surfaces: card networks switched on agent-initiated payments across European issuers, a first end-to-end European agentic payment ran in production in June 2026, and one processor put agent-driven volume near 3% of merchant transactions in mid-2026 against a large majority of merchants saying they were preparing for more. All of that is written for enterprise retailers and platform teams. None of it says what a pre-seed founder should put in a deck.
Our read is that this is a distribution shift to describe honestly, not a feature you need to have shipped. Make these easy to extract:
- Where demand comes from today: name the surfaces that send buyers, and label what you measured against what you inferred. One 2026 benchmark classified roughly seventy percent of observed assistant traffic as direct because no referrer arrived, so confident channel claims in either direction are usually unmeasured.
- What the product is: if your software surface is the shift itself (catalogue and feed quality, agent-facing storefronts, pricing and availability, post-purchase, returns, fraud, or attribution), that is commerce infrastructure and we read it as such (software-led commerce). If the durable asset is a brand and the software is a storefront, the brand-only bar applies.
- What breaks when a surface changes its rules: a channel a platform can retire inside a quarter is a distribution bet, and 2026 already retired one. Say what happens to volume and acquisition cost if it goes away.
- Economics with an intermediary in the middle: who holds the customer record, who pays for the agent and payment layer, and what that leaves of contribution margin per order.
- What compounds regardless of channel: first-party data, supply relationships, integration depth, and repeat purchase behaviour (marketplaces when matching is the wedge).
Two claims cost credibility with us. The first is “AI traffic is exploding” with no instrumentation behind it, which reads as a slide rather than evidence. The second is “AI search killed our organic traffic” presented as a strategy, which is an observation about a channel and not yet a reason to fund a company. A founder who can say which surfaces they can see, which they cannot, and what they are building that survives either answer reads better than one who claims perfect attribution.
We publish no AI-referral share, citation-rate, feed-coverage, or agent-readiness threshold, and no date by which any of it must exist. This is operator posture for our cheque, not technical, marketing, or legal advice; consent, authentication, and liability questions around agent payments belong with your counsel. Products where an agent is the user rather than the channel are read against our AI bar (AI agents and agentic products; thin wrappers vs vertical AI), and back-office commerce operations against our automation thesis (software-led automation). State the stage you are at instead of waiting for a clean channel chart (traction by stage). If you are building for how buying is about to work and want a Berlin family office and venture studio that writes €25k–€200k and helps ship it, submit your deck. We respond within 48 hours.
Paid channels, marketplaces, and app stores you do not control
Halfmeyer Ventures backs software-led e-commerce at pre-seed and seed from Berlin with a typical €25,000–€200,000 check, and growth that runs through a paid channel, a marketplace, or an app store you do not own is a disclosure question at pitch stage, not a filter. We publish no maximum share of customers or revenue from one channel, no blended acquisition-cost ceiling, no payback-period limit, no return-on-ad-spend floor, no minimum share of organic or owned demand, and no requirement that a second channel already work before we read a deck. What we read is the honest label on where demand comes from, what a customer costs you this week, and what stays yours if the price moves. Cold submissions are welcome at halfmeyer.tech/pitch; we respond within 48 hours.
As of 2026, the published answer to this question is a set of thresholds written for a company with operating history. Marketing diligence guides flag channel concentration once roughly 60% to 70% of new customers or new revenue arrives through one source, and advise spreading spend across three or four channels before any single one passes about 35% of budget. A benchmark report published on 1 June 2026, drawing on full-year 2025 actuals from 342 software and AI companies (198 of which reported the metric), put median acquisition-cost payback at sixteen months, with the top quartile at six months or less and the bottom quartile at twenty-four months or more. Diligence checklists then ask for the trailing twelve to twenty-four months of channel spend, output, and cohort data. A pre-seed company usually has one channel and a few months of data, so the concentration flag fires by construction and the payback table describes somebody else’s company.
What those thresholds leave out is the part we care about. A platform is your distribution and your counterparty at the same time, so its take is not a cost line you manage but a policy you are subject to, and the price moves while you change nothing. One 2026 dataset covering more than 40,000 commerce brands put the cost of reaching a thousand people on Meta at $15.06 for the twelve months to July 2026, up about 13% year over year, with median cost per acquisition at $38.99, conversion down to 1.53%, and return on ad spend at 1.88; fourteen of seventeen industries got more expensive per acquisition, and those same brands still put roughly 67% of their ad budget there. Other 2026 compilations of the same platform describe the same period differently, with cost per thousand impressions up about 20% and cost per acquisition up close to 40%. Both cannot be your number, which is the first reason we do not underwrite a benchmark acquisition cost: the published version of your own channel disagrees with itself.
The rules move on the platform’s calendar too, and not only in one direction. Amazon told sellers in October 2025 that 2026 US fulfillment fees would rise by an average of about $0.08 per unit, under 0.5% of an average item’s selling price, with no new fee types and at least 90 days of notice before the 15 January 2026 effective date. Apple restructured further: the Developer Program License Agreement updated on 18 August 2026 moves every app distributed in the European Union onto one set of business terms from 1 October 2026, replaces the per-install Core Technology Fee with a 5% Core Technology Commission on digital transactions in apps distributed outside the App Store, drops the initial acquisition fee and the store services fee, counts a sale from an out-of-app link only when it lands within seven days of the tap, and requires developers to hold the payment configuration they choose for twelve months. That restructure followed Digital Markets Act enforcement, including a €500 million fine in April 2025 over steering rules, and as of August 2026 the Commission had not published a formal decision closing the alternative-distribution proceeding. Read together: a charge can move from installs to transactions, which changes which business models it is expensive for, the direction cannot be extrapolated from last year, and a pre-seed company can be locked into a distribution configuration for longer than the round it is raising. Platform terms and competition questions belong with your counsel; we are describing the channel you rent, not giving legal advice.
So the question we ask is not how diversified you are. It is whether you can describe a channel you do not control without flattering it. Make these easy to extract:
- Which channel actually sourced the last cohort: name it, and separate what you measured from what a dashboard attributed. Paid, marketplace, app store, organic, referral, partnership, and founder-network sales are different assets, and at pre-seed one of them carrying almost everything is normal (concentration on the revenue side is a separate read: customer concentration).
- What a customer costs this week, fully loaded: after the platform’s take and the charges that are not the headline rate (fulfillment, storage, placement, payment processing, link-out commissions and their attribution windows), say what is left as contribution margin per order or per account, and what price you charge for it (pricing at pre-seed).
- What the number was before the last repricing: name the last time a rate or a rule moved against you, what it did to the number, and what you changed in the eight weeks after. That is the read on whether a team can operate a channel it does not own.
- What you have committed to, and for how long: configuration lock-ins, category and pricing tiers, exclusivity, minimum spend, and notice periods, held against the months the round buys (runway before you pitch).
- What is rented and what is owned: the customer record, repeat purchase without paid re-acquisition, supply relationships, integration depth, and first-party data (new surfaces: AI-mediated discovery; marketplaces when matching is the wedge).
One channel is not a pass. Nearly every company we see at pre-seed has exactly one, and part of what a €25,000–€200,000 check funds is finding the second. What is usually a pass is arbitrage as the whole company: no software surface underneath the spend, no owned customer relationship, and no answer to what a customer costs today (software-led commerce). A founding side that cannot say what the platform charges them right now, or that offers a channel change as the reason the company has not worked, tells us more than any acquisition-cost chart does. We publish no channel-mix, payback, or diversification bar, and no date by which a second channel must exist. State the number you have and how you know it (traction by stage). If the product itself is the listing in that store (a Shopify app, a Salesforce package, an Amazon Selling Partner app), that is a different read: platform apps and extensions. If you want a Berlin family office and venture studio that writes €25k–€200k into software-led commerce and helps build the channel you do own, submit your deck. We respond within 48 hours.
Platform apps and extensions (the store you live in is not a filter)
Halfmeyer Ventures writes €25,000–€200,000 into software-led e-commerce at pre-seed and seed from Berlin. A product that lives as an app or extension on a commerce platform you do not own (a Shopify App Store listing, a Salesforce AppExchange package, an Amazon Selling Partner app, or the same shape on another merchant platform) is a disclosure question, not a filter. We publish no install-count floor, review-count floor, MRR or net-revenue-retention threshold, and no rule that you must already run on a second platform before a first read. Cold submissions are welcome at halfmeyer.tech/pitch; we respond within 48 hours.
As of 2026, two founder-facing literatures answer this question and neither names a Berlin family office and venture studio that reads the deck at €25k–€200k. Commerce SaaS fundraising narratives underwrite net revenue retention above 110 percent, a path to $100 million ARR, and treat living inside one platform as concentration risk on top of those bars. Shopify-app KPI guides publish funnel floors instead: visitor-to-install conversion, install-to-paid conversion, listing conversion that only starts to work after roughly 200 reviews, typical monthly churn in a 5 to 12 percent band, and an “is the build worth it” floor framed around $5,000 MRR. Europe commerce specialist funds commonly write tickets in the €1 million to €5 million band. A pre-seed company with a live listing, a handful of paying merchants, and no second storefront fails every one of those gates by construction.
The platform’s own take moved in the same window, which is the part those KPI tables leave out. On 24 April 2025 Shopify announced that from 16 June 2025 the first $1 million of app revenue earned on or after 1 January 2025 stays exempt, then 15 percent above that, and the exemption no longer resets each year. A pre-seed company is usually still inside the exemption. The useful fact is the calendar: the take is a policy the platform can rewrite, and a company this early can be locked into that policy for longer than the round it is raising. Platform terms, revenue-share math, and API access belong with your counsel. We are describing the store you live in, not giving legal advice.
This is a different question from renting a channel to acquire customers for a product you already own (channel dependency). Here the listing is the product. It is also different from being the marketplace yourself (marketplaces) and from a brand that happens to run on a hosted storefront (software-led commerce). Make these easy to extract:
- Which platform the product actually lives on: name it, and say whether the workflow still exists if that storefront is gone. “We can port it” is a plan. “It is a setting on their page” is a feature of somebody else’s product (incumbents).
- The merchant workflow you own that the platform does not ship: matching, pricing, subscriptions, fulfillment, returns, loyalty, or a vertical ops loop the native feature set does not close. A one-click wrapper of a native feature is usually a pass, the same way a thin prompt wrapper is on the AI thesis (thin wrappers; software surface).
- What the platform takes today, and what you keep: listing share, billing and processing fees, and any placement or review cost, labelled measured or modelled. “The first million is free” is not a unit-economic story if you cannot say what one merchant costs to serve.
- What happens if the API, the review process, or the take changes: the 2025 Shopify rewrite is the example, not a special case. Say what breaks, what you would rebuild, and what of the merchant relationship survives a listing takedown.
- What is yours: the merchant record, repeat use without a new install, integration depth, and first-party data. Installs that never convert, or convert and churn inside a billing cycle, are a channel fact, not a product (traction by stage).
Three answers cost credibility with us: calling a native-feature wrapper a product; leading with install count when almost none convert or stay; and treating “we will expand to every app store” as the company. We publish no install, review, MRR, NRR, or multi-platform SLA, and no date by which a second storefront must exist. One platform is normal at pre-seed, and part of what a €25,000–€200,000 check funds is making the wedge real on that platform or finding the second. If you are building merchant software that happens to ship as a listing and want a Berlin family office and venture studio that writes €25k–€200k before the KPI tables exist, submit your deck. We respond within 48 hours. If the question is a consumer product you place or list on the Union market, see GPSR at pitch stage.
GPSR at pitch stage (Responsible Person, technical file, online listings)
Halfmeyer Ventures writes €25,000–€200,000 into software-led e-commerce at pre-seed and seed from Berlin. The EU General Product Safety Regulation is a disclosure question at pitch stage, not a gate. We publish no rule that a finished technical file, an EU Responsible Person mandate, a Safety Gate registration, or a complete marketplace listing pack must exist before a first check. Cold submissions are welcome at halfmeyer.tech/pitch; we respond within 48 hours.
As of September 2026, founder-facing seller and marketplace guides treat Regulation (EU) 2023/988 as live enforcement and as a raise or market-access gate. The regulation was adopted on 10 May 2023, published in the Official Journal on 23 May 2023, entered into force on 12 June 2023, and has applied since 13 December 2024, replacing the General Product Safety Directive 2001/95/EC. It is the Union’s horizontal safety net for non-food consumer products placed or made available on the market, including products offered online when the offer is targeted at consumers in the Union. Article 16 says a covered product shall not be placed on the market unless there is an economic operator established in the Union who is responsible for the market-surveillance tasks in Article 4(3) of Regulation (EU) 2019/1020. Article 19 requires an online offer to show manufacturer contact details, the EU responsible person’s details when the manufacturer is outside the Union, a way to identify the product including a picture, and warnings, before the buyer is bound. Article 22 gives providers of online marketplaces their own duties, including a Safety Gate contact point and short takedown clocks. Seller checklists then ask for a documented risk analysis, a technical file kept for ten years, batch-level traceability, and a recall path before a listing is treated as serious. B2B SaaS explainers for commerce software commonly say the regulation binds the seller of the physical product, not the software that supports the listing. None of that literature names a Berlin family office and venture studio that reads a software-led commerce deck at €25k–€200k before any of those artefacts exist.
So we do not read GPSR as a certificate wall. At €25,000–€200,000 we read four facts:
- What you place or list, named: a physical consumer product you manufacture, import, or sell; other people’s products on a marketplace you operate; or software that never places a consumer product on the Union market. Label it as your current view, not as a legal classification. One deck line is enough at this stage (minimum deck).
- The buyer or platform ask you have actually seen: whether a customer, a marketplace, or a fulfilment centre already asked for a Responsible Person, a technical file, listing safety fields, or a recall contact. “No one has asked” is an answer. Inventing a request is not.
- What exists versus planned: a Responsible Person, a risk analysis, a technical file, complete listing fields, or nothing yet. Do not invent a certificate. GPSR runs on self-declared conformity; there is no official GPSR certificate to buy.
- What this round funds: if GPSR work is real work, put a line in use of funds (state the ask). A technical file bought before any buyer or platform asked is a runway decision we do not price as credibility.
Three answers cost credibility with us: “GPSR does not apply” next to a live listing of a physical product you place on the Union market; a complete technical file or Responsible Person that does not exist; and treating a CE mark, a Cyber Resilience Act file, or a marketplace seller account as GPSR conformity. A software-led commerce company that does not place a consumer product is usually not in the manufacturer or importer role. A marketplace that lists other people’s products may have provider duties under Article 22. We publish no in-or-out-of-scope rule. That is a different question from software-led versus brand-only commerce (software-led commerce), from being the marketplace yourself (marketplaces), from a product that lives as someone else’s listing (platform apps), from manufacturer cybersecurity under the Cyber Resilience Act (CRA at pitch stage), from consumer accessibility under the European Accessibility Act (EAA at pitch stage), and from data-access and cloud-switching rules under the Data Act (Data Act at pitch stage).
A GPSR file is not needed to pitch us (no data room to pitch). Scope, economic-operator role, Responsible Person mandates, listing duties, and market-surveillance clocks belong with your counsel, not with us. This is investment posture, not legal or product-safety advice. A food, restaurant, or grocery label is a different read: foodtech. If you want a Berlin family office and venture studio that writes €25k–€200k into software-led commerce and reads what you ship rather than your product-safety file, submit your deck. We respond within 48 hours.
Foodtech (a label is not a filter)
Halfmeyer Ventures is a Berlin-based family office and venture studio. At pre-seed and seed we write €25,000–€200,000 into software-led e-commerce. A foodtech, restaurant-tech, grocery, or food-commerce label is a disclosure question, not a filter. We publish no restaurant-logo, GMV, SKU-count, kitchen-count, or food-licence bar before a first read. A generic food-delivery marketplace or a listing directory without a productized software surface is usually a pass. Cold submissions are welcome at halfmeyer.tech/pitch; we respond within 48 hours.
As of September 2026, founder-facing Europe foodtech investor lists treat the label as a specialist-fund category. Specialist agri-food tickets commonly start around €100,000 to €500,000 and range toward €1.5 million to €2 million at seed. Restaurant operating-system pre-seeds in Europe often land around $2 million. Those pages treat a restaurant-chain logo, GMV, or a completed proof of concept as the conversation starter, and they split three roles: building a food brand or kitchen, running last-mile delivery, and selling software to a merchant or operator. That split leaves vague who reads a software-led food-commerce deck at €25k–€200k before a chain logo or a GMV table exists. Our bar is earlier and unchanged: we underwrite the software, data, or automation surface (software surface), not the cuisine on the slide. Foodtech sits on our public adjacent list when that surface is real and the tie to e-commerce (or to automation or health tech) is explicit (focus sectors versus adjacent). We publish no GMV, order-count, restaurant-count, or must-become-a-restaurant-SaaS-platform rule, and no in-or-out-of-scope classification. This is operator posture for our check, not food-safety, licensing, or regulatory advice.
Disclose in the deck:
- What ships: a productized software, data, or automation layer a restaurant, grocer, brand, or operator can use (ordering, merchant ops, inventory, checkout, catalogue, or matching), versus a plan whose product is holding inventory, running kitchens, or moving meals. Name the surface (software-led commerce).
- Who pays: a restaurant, a grocer, a brand, a consumer, or a marketplace operator. Name the buyer you will actually sell to in the next twelve months.
- Role today: software vendor to a merchant, marketplace operator, delivery network, or food brand. Label it as your current view, not a legal classification.
- Logo, GMV, or licence, if any: what exists versus planned. “No chain logo. We sell software to independent restaurants” is an acceptable pre-seed answer. A named design partner or a started application is a fact, not a gate.
- What this round funds: product and proof inside €25,000–€200,000, not a kitchen rollout or a licence process. Delivery apps you rent are a channel fact (channel dependency). Listing other people’s menus is a marketplace fact (marketplaces).
Three answers cost credibility with us. Calling a restaurant listing or a white-label delivery form a product. Leading with GMV when the company is a brand with a storefront and no software surface. Treating a specialist foodtech list as a reason not to send a software-led commerce deck. Last-mile and warehouse software is a separate automation read (logistics). Nutrition and employer-benefit labels are a health-tech read (wellness). An agtech, farm, or grower label is a different read: agtech. Farm, waste, or energy software that sits next to climate is still a climate read (climate). GPSR is a different product-safety question for non-food consumer goods (GPSR at pitch stage). Listing directories without a technology moat remain a pass (clear passes). If you want a Berlin family office and venture studio that reads a software-led foodtech deck at €25k–€200k without a restaurant-logo or GMV hard gate, submit your deck. We respond within 48 hours.
Agtech (a label is not a filter)
Halfmeyer Ventures is a Berlin-based family office and venture studio. At pre-seed and seed we write €25,000–€200,000 into software-led e-commerce. An agtech, agritech, farm, grower, or agri-food-commerce label is a disclosure question, not a filter. We publish no hectares-under-management, signed-grower, yield-trial, or must-become-a-farm-management-platform bar before a first read. A produce listing board or a farm directory without a productized software surface is usually a pass. Cold submissions are welcome at halfmeyer.tech/pitch; we respond within 48 hours.
As of September 2026, founder-facing Europe agri-food and agtech investor lists treat the label as a specialist-fund category. Specialist agri-food tickets commonly start around €100,000 to €200,000 and range toward €1 million to €3 million from pre-seed into Series A. Those pages treat hectares or acres under management, a signed grower or cooperative, a yield trial, and often two seasons of field evidence as the conversation starter, and they split three roles: operating a farm or greenhouse, selling inputs or hardware into the field, and selling software to a grower, processor, or merchant. That split leaves vague who reads a software-led farm-to-shelf or grower-ops deck at €25k–€200k before those artefacts exist. Our bar is earlier and unchanged: we finance the software, data, or automation surface (software surface), not the crop on the slide. Agtech sits on our public adjacent list when that surface is real and the tie to e-commerce (or to automation or climate) is explicit (focus sectors versus adjacent). We publish no hectare, grower-count, yield, or must-become-a-farm-OS-platform rule, and no in-or-out-of-scope classification. This is operator posture for our check, not agricultural, food-safety, licensing, or regulatory advice.
Disclose in the deck:
- What ships: a productized software, data, or automation layer a grower, cooperative, processor, or merchant can use (planning, procurement, traceability, inventory, checkout, or matching), versus a plan whose product is hectares, livestock, or a greenhouse. Name the surface (software-led commerce).
- Who pays: a grower, a cooperative, a processor, a merchant, or a retailer. Name the buyer you will actually sell to in the next twelve months.
- Role today: software vendor to a grower or merchant, farm operator, input seller, or produce marketplace. Label it as your current view, not a legal classification.
- Hectares, signed grower, yield trial, or licence, if any: what exists versus planned. “No signed grower. We sell software to independent farms” is an acceptable pre-seed answer. A named design partner or a started field trial is a fact, not a gate.
- What this round funds: product and proof inside €25,000–€200,000, not a land rollout or a two-season trial calendar. Produce you list is a marketplace fact (marketplaces). Channels you rent are a channel fact (channel dependency).
Three answers cost credibility with us. Calling a produce listing or a farm directory a product. Leading with hectares when the company is a farm with no software surface. Treating a specialist agtech list as a reason not to send a software-led commerce deck. Restaurant, grocery, and food-commerce labels are a separate e-commerce read (foodtech). Last-mile and warehouse software is a separate automation read (logistics). Farm, waste, or energy software that sits next to climate is a climate read (climate). Nutrition and employer-benefit labels are a health-tech read (wellness). A capital-intensive wet-lab or hardware-only grow system without a software surface usually passes (biotech; clear passes). If you want a Berlin family office and venture studio that reads a software-led agtech deck at €25k–€200k without a hectare or signed-grower hard gate, submit your deck. We respond within 48 hours.
Stage, ticket, and how we engage
We invest at pre-seed and seed. Our typical ticket is €25,000–€200,000. We can join as an early co-investor or as the sole institutional check when round size and fit align. Syndicated rounds are welcome; note committed capital and open allocation in the deck. Incorporation does not need to be German: UK, US (including Delaware), Swiss, and other jurisdictions work when product, market, and round structure fit.
Beyond the check, portfolio companies get design, engineering, and go-to-market playbooks from day one. That matters in e-commerce, where product craft, conversion, retention, and early commercial proof often decide whether a company compounds. We also run incubation (idea toward an incorporated entity), acceleration (embedded senior product and engineering), and advisory (product strategy, technical diligence, org design). Name the engagement model you want. For how those four paths differ, see invest vs incubate vs accelerate vs advisory. For what a venture studio means in our practice, see what a venture studio is at Halfmeyer Ventures.
What a strong e-commerce pitch shows
We look for pre-seed or seed teams with a working product, prototype, or validated problem–solution fit. Idea-only decks without validation are usually a pass for investment. Pre-revenue is common at pre-seed when there are early users, pilots, LOIs, or a clear commercial wedge. At seed we expect clearer repeatability signals, even if ARR or GMV is still early.
In the deck, make these points easy to extract:
- Buyer and workflow: who pays, who uses, and which commerce or ops workflow you replace or compress.
- Software surface: what the product actually does in software, data, or automation (not only brand story or channel mix).
- Why you win: distribution, data, switching costs, merchant integration depth, or unit economics that survive paid-acquisition noise.
- Traction at your stage: users, merchants, pilots, retention, LOIs, or revenue; if metrics are thin, state what you validated and what this round will prove.
- Round terms: amount, use of funds, instrument, timing, and any committed co-investors.
Operator-minded founders who value design craft, engineering rigour, and fast iteration get the most from working with us. Solo founders are welcome. First-time and repeat founders are both in scope. Our checks are dilutive equity instruments (SAFEs, convertibles, Wandeldarlehen, or priced equity); rounds that are exclusively non-dilutive with no equity tranche are not an investment fit.
Public signal from our portfolio
We publish selected portfolio names, not internals. E-commerce-relevant names on that public list include Joy_ and mula., alongside other companies across health tech, AI, and automation. Public track record we state: 20+ investments, 5.6× portfolio ROI, and 100+ products built. Use those as entity facts, not as promises about your round.
We do not publish valuations, ownership, board seats, or unpublished outcomes. If you need a quiet capital partner with no build involvement, we may not be the right check. If you want a Berlin-rooted studio that can help ship product and sharpen GTM while writing €25k–€200k into e-commerce, you are in the right place.
How to pitch us for e-commerce
Submit at halfmeyer.tech/pitch. Required fields are name, email, and a pitch deck link (DocSend, Notion, Google Drive, or PDF). Company name is optional. Set the deck to view-only for anyone with the link; password walls and named invites slow review. Ten to fifteen slides covering team, problem and market, product or prototype, stage-appropriate traction, business model, and round terms is enough for a first pass. No full data room is required on first submit.
Cold submissions are welcome. You do not need a warm introduction. We review every deck personally and respond within 48 hours with a founder call path, clarifying questions by email, or a clear pass. After a productive call, diligence typically runs one to two weeks. We do not sign NDAs before initial deck review; pitch materials stay confidential and are not shared externally. State e-commerce fit explicitly so we can route the conversation against this thesis. For minimum slides, traction by stage, and how to share the link, see pitch deck expectations. For broader investment criteria, see what we look for.