Mosaic Venture Lab wants your startup on the factory floor

A startup can build a clever sensor and still struggle to answer a brutally ordinary question: who will put it in a product? Mosaic Venture Lab has built its business around that gap. From Taipei, the company scouts technology, organizes corporate accelerator programs and helps young businesses negotiate their way toward investment and industrial collaboration. The ambition is tangible: get something that works in a demonstration closer to something a manufacturer can use.

Consider its work with Continental. Mosaic describes a search across Taiwan, Japan and South Korea for semiconductor, chip-design, materials and display technologies. The brief starts with problems inside the automotive supplier and looks toward its five-to-ten-year development roadmap. That is a rather different assignment from collecting attractive startup logos for a presentation.

The distinction matters to both parties. A founder wants access to the people who can evaluate a component. A large company wants technology that fits something it intends to build. Mosaic says it has facilitated several technology collaborations through the relationship. The useful idea is the sequence: establish the industrial problem, find a relevant team, then pursue joint research or development. An introduction earns its keep when there is work attached.

01Taiwan is the starting point, not the boundary

Mosaic occupies an awkward but potentially useful patch of the venture market: part investor, part corporate adviser, part accelerator. Its services include technology scouting, research reports, hackathons, private demo days and fundraising support. Corporations use it to explore emerging technology and suppliers; startups use its programs to reach industrial partners and investors. These are distinct audiences with overlapping interests, rather than a single customer buying a standard subscription.

The geographic bet is equally specific. The firm searches East Asia's manufacturing and technology communities and connects them with overseas companies. A European business looking for a Taiwanese development partner and a Taiwanese founder seeking an international buyer are approaching the same commercial relationship from opposite ends. Mosaic positions itself in the middle, where technical vocabulary, business expectations and personal connections all need translating.

That work has an older lineage. Yushan Ventures, founded by Volker Heistermann in 2011, describes Mosaic as an accelerator it created in 2018 that now operates independently. Mosaic's own company history also places its origins in 2018, during the electric-vehicle boom. Its public profiles often use 2019 as the starting year. The earlier relationships help explain why a small venture lab could organize around established automotive companies.

Heistermann's connection to the island is unusually direct. “My heart beats in Taiwan,” he said in a profile republished by Mosaic. That account identifies Ken Huang as a co-founder and Johnny Yu as an operating partner; the current team page lists Yu and Heistermann as managing directors. Mosaic's staff also includes investment, program, business development and design roles, suggesting a business built to do more than review investment decks.

02Fourteen pitches. One next step.

The 2023 Audi hackathon provides a compact view of the method. Its challenges drew on Audi data and covered predictive analysis, digital paperwork and customer sentiment. The event centered women in technology and brought together participants with mentors and judges working in design, data, artificial intelligence and entrepreneurship.

Mosaic reports more than 150 signups in 17 days, more than 50 selected participants and 14 solutions pitched over two and a half days. One concept was ready for a proof of concept. Those are different milestones: interest, participation, proposals and readiness to test. Treating them separately is more informative than calling every pitch an innovation success.

There is a lesson here for anyone commissioning a challenge. Give teams a business problem with boundaries. Bring the relevant expertise into the room. Decide what a credible next experiment would require. A crowded event can be enjoyable; a defined follow-up gives a promising idea somewhere to go.

Audi hackathon, 2023 · Company-reported milestones, not commercial outcomes.

03A packaging idea reaches Kaohsiung

The LITEON relationship takes the story closer to a factory. When the electronics company announced its LITEON+ platform in 2023, it named Mosaic as a partner for the first six-month program. Selected teams would receive training, workshops, networking and a demo day, with opportunities to work with business units on proofs of concept and explore investment from LITEON and its partners.

The important organizational detail is access to operating divisions. Corporate interest can evaporate if nobody owns the next task. Mosaic's description of the collaboration emphasizes matching startups to the parts of LITEON that can help develop, manufacture and bring products to market. For hardware teams, the identity of that counterpart can matter as much as the identity of an investor.

By its 2024 demo day, LITEON+ reported five participating companies: PackAge+, Elephantech, metabiz, Sounds Great and Pulxion Medical Technology. PackAge+ had established a demonstration of its circular packaging solution at LITEON's Kaohsiung factory. Japan's Elephantech had signed a memorandum on low-carbon electronic components and sustainable processes, with scaled production planned for 2025.

A factory demonstration and a memorandum are concrete, but they are not interchangeable with a production contract. The distinction is useful for founders assessing an accelerator: ask which stage its industrial relationships have actually reached. Here, the evidence includes an on-site packaging demonstration and an agreement to develop another collaboration. That is a firmer basis for judgment than the size of a partner's logo.

Ask which stage the industrial relationship has actually reached.The question to take into an accelerator meeting

04Capital joins the manufacturing conversation

In June 2025, Mosaic and power-supply manufacturer FSP Group announced an incubation fund. The proposal combined seed capital with technical resources, manufacturing help and international commercial connections. Its focus covered semiconductors, mobility and applied AI; the announcement targeted a third-quarter 2025 launch.

The fund's own page now identifies Yu and Heistermann as investment committee members. Its technology interests include advanced packaging, sensors, AI chips, vehicle connectivity and autonomous driving. These fields make the industrial partner relevant: a founder may need technical and manufacturing decisions to move alongside financing decisions.

Mosaic's fundraising work also extends to helping other teams reach investors. In its Sounds Great case study, the company describes arranging strategic investor connections for a semiconductor-based speaker business and helping with the complexities of negotiating terms. The practical point is easy to miss: getting someone interested in investing is only one part of completing a financing.

Its business model therefore spans investment activity, corporate innovation services and training. Those roles put it alongside specialist venture firms, corporate scouting teams and accelerator operators. Its case for being hired rests on combining them around a particular industrial need. A company that already has the right internal scouting team and supplier relationships may have less reason to outsource that work.

05The same adoption problem, in a hospital

Mosaic's healthcare activity follows a related logic, with different specialists. Its CellTech program offers cell and gene therapy ventures commercialization support, preclinical and clinical expertise, individual mentoring and access to funding networks. The firm also lists the Landseed Digital Health Accelerator, established by Landseed International Medical Group to support digital and elderly-health technology.

Landseed's advertised resources include clinical trial settings, long-term community tracking and academic cooperation. A healthcare founder needs a setting in which a proposed tool can be evaluated and understood. Access to that setting serves a different purpose from another general business workshop.

At Medical Taiwan in June 2025, Mosaic participated in the M-novator demo day. The program brought together teams from Taiwan, the United States, South Korea, Belgium and Israel. Examples included RNA-drug developer InFocus Therapeutics and Pulxion, working on AI for carotid artery detection. Yu's presentation addressed the distance between laboratory work and commercial adoption, including development timelines, funding and market entry.

This broadens Mosaic's market without making the sectors identical. A component supplier and a medical technology team need different evidence and different partners. The common commercial task is finding an institution able to evaluate what the startup has built and help it take the next step.

06When the founder needs to stop talking

Mosaic's Sell Global course makes its diagnosis unusually vivid. The company describes founders leading with manufacturing specifications, underbudgeting overseas activity and letting a less-fluent CEO dominate a pitch while a stronger communicator sits silent. Its eight-week curriculum uses buyer role-play; one exercise temporarily bars the CEO from speaking. These are the firm's reported observations, not a verdict on every Taiwanese founder.

The program says it grew out of earlier trade-show training. Its advertised terms are subsidized pricing with no equity taken. The transferable exercise is simple: let someone outside the team play a skeptical buyer, and give the clearest communicator the floor.

A second product puts a number on the training business. VC Academy, launched in Taipei on August 14, 2026, advertises about ten founding-cohort seats at US$1,000 for 12 in-person sessions over six months. The curriculum covers diligence, deal terms and investment judgment in Taiwan. That is a course price, separate from startup financing or a corporate project's budget.

For founders, the fit depends on readiness. Sell Global targets teams with a product to sell; Mosaic's industrial work depends on a match with a partner's needs. An introduction cannot supply that match by itself. The most useful question to bring to Mosaic is therefore specific: which buyer, development team or clinical partner could test this product, and what would they need to see? That gives a venture lab something worth connecting.