De-Risking Deep Tech: How Smart Investors Support Innovation Beyond the Check

Deep tech isn't built in a garage over a weekend---and it doesn't succeed with capital alone. They're rooted in physical infrastructure, complex supply chains, and breakthroughs in materials science, semiconductors, robotics, and industrial AI. They may take longer, burn more capital, and are often constrained by physics rather than code.

That complexity makes deep tech both indispensable and difficult to scale---and it's where many investors struggle. Throwing money at the problem isn't enough. What's missing is operational alignment: support that goes beyond the check.

Few investors understand this terrain better than Xiaosi Yang. A semiconductor veteran with two decades of experience on both sides of the Pacific, Yang has built his career inside the trenches most investors only read about. His work spans from early design wins to public listings, with a focus on helping startups navigate the messy middle---from lab prototype to manufacturable, customer-ready product. As Yang sees it, deep tech doesn't lack funding. It lacks functional support.

The Deep Tech Misfire: Capital Without Capability

Between 2016 and 2021, disclosed deep tech funding more than quadrupled. On paper, the sector looks well-capitalized. But on the ground, progress remains uneven. Startups are frequently stuck between invention and scale, often burning through funds before product-market fit is even validated---let alone achieved.

One reason is that many investors apply a software lens to fundamentally different businesses. In software, speed is everything: ship fast, iterate, optimize growth loops. In deep tech, that mindset can backfire. It's the difference between adjusting code and trying to hire metallurgists in a tight labor market.

"You can't just look at a pitch deck, drop in funding, and expect returns in three years," says Yang. "These companies need structural support---from product design to supply chain integration to customer acquisition---just to survive, let alone scale."

In Yang's view, the better analogy for deep tech investing is engineering program management. It's about identifying friction points early---where the design may fail, where talent is missing, where the market doesn't yet exist---and building a plan to de-risk each of those nodes. That kind of support requires fluency in how technologies are built and brought to life.

Investing Like a Builder

Yang's investment philosophy has been shaped by firsthand experience building the very systems others now fund. Throughout his career, he's helped scale semiconductor and manufacturing companies from obscure early-stage ventures into category leaders.

He doesn't sit back and wait for updates---he gets involved at the ground level, working shoulder-to-shoulder with engineers and product leaders.

Rather than limit his role to capital allocation, Yang is known for engaging directly with engineering and product teams. His involvement has included refining product roadmaps, aligning development milestones with customer needs, and addressing failure points in supply chains and manufacturing. He often operates as a translator between technical constraints and commercial imperatives.

This style of hands-on support---quiet, detailed, and sustained---is often what separates stalled innovation from scalable product. The work isn't headline-grabbing, but it's fundamental: shaping how raw breakthroughs evolve into systems that perform reliably at scale and meet real-world requirements.

Building the Blueprint Before the Term Sheet

What makes Yang's model distinct is the work that happens before the deal closes. The diligence process is already operational: Yang and his team probe engineering assumptions, stress-test supplier dependencies, pressure-test hiring plans, and map out where the go-to-market motion is likely to stall.

"These conversations happen outside of spreadsheets," Yang says. "We're looking at executional friction, not just financial runway."

This is especially urgent in the U.S., where industrial policy and national strategy are converging. The CHIPS Act---projected to deploy more than $50 billion in support for domestic semiconductor manufacturing---has sparked a wave of investor interest. But federal dollars can't fabricate chips on their own. Building resilient hardware companies requires day-to-day operational fluency: knowing how to navigate vendor lead times, scale production without quality loss, and comply with a maze of regulatory standards.

XStar Capital's portfolio reflects this approach, with plans to extend its operating and investment model into the U.S. Rather than stepping in after problems arise, Yang and his team work to anticipate them---providing guidance on design decisions, supplier selection, hiring priorities, and market rollout strategies. The focus is on building resilience before scaling, not cleaning up after the fact.

The Case for Technically Fluent Investors

Yang is candid about what separates serious investors from opportunists. "The era of deep tech tourism is over," he says. "Chasing the next quantum or AI chip without understanding how any of it gets made is a losing game."

This doesn't mean every investor needs an engineering degree. But they do need a working understanding of how technology development unfolds---and how to support it. That includes knowing how to assess technical tradeoffs and distinguish between a delay that signals trouble and one that reflects a necessary iteration.

Understanding these layers is now table stakes. With geopolitical tensions and economic security rising to the top of national agendas, deep tech sectors---semiconductors, energy, aerospace, critical materials---are being recast as strategic assets. That shift demands a different caliber of capital. Not just more of it, but smarter, more embedded, and technically fluent.

"You don't invest in deep tech to chase a trend," Yang says. "You do it because you understand how innovation builds up---and you're willing to grow alongside it."