Tang Rui studied automation engineering and did not move directly into finance after graduation. He spent nearly eight years in product and engineering management at an enterprise software company, including three years running a product line built for manufacturing customers. That experience gave him a reading of the word “technology” that differs from most investors: he breaks it into three layers — can it actually be built, can it be delivered at scale once built, and will customers keep paying for it once it is delivered at scale. Most investors stop at the first layer. His operating background makes the second and third layers instinctive to him.
Moving from operations into investing, Tang Rui first spent three years in the technology group of an international investment bank, advising technology companies on mergers and financing, before joining Havrion Capital to focus on technology investments. His education also includes a master's degree in technology management from the National University of Singapore, coursework that turned his earlier, more intuitive product judgment into a structured framework for assessing technology companies — and became the starting point for the technology-investment methodology he later helped establish inside Havrion Capital.
The technology-investment work he leads covers systematic assessment of companies, markets, technology trajectories, competitive dynamics and business models. Tang Rui repeats one point to his team often: the most dangerous mistake in technology investing is not missing an opportunity, it is mistaking a moment of market enthusiasm for a durable competitive barrier. He asks the team to be able to answer two questions clearly for any technology company under review — why is the cost of switching away from this product high enough to matter, and why can competitors not replicate the core capability within two or three years. Without clear answers, growth numbers alone, however impressive, do not justify a long-term investment.
Tang Rui has developed his own way of separating fashionable technology from durable technology franchises. The former tends to rest on a single feature or a short-lived shift in market sentiment, and loses customers quickly once a better alternative appears. The latter is usually built on accumulated customer data, workflow embedding or network effects, and its moat strengthens the longer it is in use. His judgment rarely comes from published reports; it comes more from direct conversations with founders, engineering leads and frontline customers — a habit left over from his operating years. He believes the real state of a technology is usually visible in product details and customer complaints, not in pitch materials.
Published Insights
Where AI Adoption Is Real in the Industrial Economy
Most industrial AI deployments remain pilots. A smaller set has moved into repeatable, cost-accountable operation. What separates the two has less to do with model sophistication than with data access, deployment capability and ownership of the workflow — the three factors that determine whether a vendor compounds or stalls at the next budget review.
7 min readFrom Project to Product: The Structural Shift in China's Enterprise Software Market
China's enterprise software market is going through a change in economic model — from custom project work billed as a one-time license, toward standardized product sold on subscription. The idea is simple to state and full of traps in execution: most companies that claim to have made the shift still carry a heavy shadow of project economics inside their revenue mix. This piece examines why mid-market vertical depth travels further than broad horizontal ambition, why retention is the one number that does not lie, and where this shift is furthest along, and where it has stalled.
8 min read