Conversations with 29 Tech Investors: There Are Treasures, But No Treasure Map

By Lili Yu, Lixin He, Qian Ren, Muxin Xu

Edited by Jing Liu, Lili Yu

In a KrNote published six months ago, 23 investors made predictions to "Waves" about the future: market mania would not return; the RMBization of the primary market would intensify; policy would become the foremost factor in investment, among others.

In hindsight, these judgments were almost eerily accurate. This may itself be a defining feature of today's primary market: change has gradually become uniform.

But one shift stands out even more. Let's call it the "tech revolution." Over the past six months, China's primary market has taken on an unmistakable flavor of "no tech, no investment." And even within the narrow definition of tech investment, the definition wasn't narrow enough. We felt this acutely during interviews for this KrNote: nearly 90% of firms proactively requested to discuss only "new energy."

To some extent, "tech" is a pseudo-concept — what corner of our world contains no technological element whatsoever? But what drives the world forward is often sentiment. The "technologization" of the primary market is itself a kind of sentiment.

"Waves" decided to split the difference. Within so-called "tech investment," we selected five sectors for discussion: new energy, semiconductors, biotech, AI, and commercial space. Because their transformations in 2023 may be most visible, and because we believe they best illustrate the characteristics of tech investment: the "dual logic" in chip investing; technology route disputes in upstream energy; the "near-total lack of consensus" in AI and commercial space; and the de-financialization and strong industrialization features present in every one of these sectors.

Through this article, we hope to explore with readers: what exactly is "tech," what characterizes tech investment, and where the money is flowing.

We are honored that in this roundtable focused primarily on investors, we revisited many old friends and added some new faces. In investment careers unlike any that came before: "May your road be long, full of miracles, full of discovery."

Our thanks to the 29 investors who shared their insights: Zongyi Bai, Founding Partner of Glory Ventures; Weiguang Chen, Managing Partner of Lanchi Ventures; Hua Feng, Partner at PineVC Capital; Zhilong Fu, Partner at Yunxi Capital; Xiangyuan Huang, Managing Director at Shangqi Capital; Ming Jin, Managing Partner at Gaohe Capital; Jianwei Li, Founding Partner of Zhencheng Investment; Dan Liu, Senior Partner of Innovation and Growth Investment at CDH Investments; Kai Liu, Partner at 5Y Capital; Lei Mi, Founding Partner of CASSTAR; Rui Ma, Partner at FreeS Fund; Bobing Ren, General Manager of Frontier Tech Fund at Sinovation Ventures; Xuesong Shi, Managing Partner at Chuanliu Investment; Huadong Wang, Partner at Matrix Partners China; Wenrong Wang, Partner at Fortune Venture Capital; Xingshi Wang, Partner at Source Code Capital; Shi Xu, Founding Partner at Hillhouse Capital (Shanxing); Kun Yang, Partner at Gaorong Capital; Yunxia Yang, Partner at Hongshan; Xinhua Yu, Partner at IDG Capital; Zhiyun Yu, Partner at Matrix Partners China; Yingzhe Zeng, Partner at Linear Capital; Gaonan Zhang, Managing Partner at Huaying Capital; Jinjian Zhang, Founding Partner of Oasis Capital; Feng Zhao, Partner at Harvest Capital; Yang Zhao, Partner at Xianfeng Changqing; Zhanxiang Zhao, Partner at Yunxi Capital; Wei Zhou, Founding Partner of CCV; and Zhifeng Zhou, Partner at Qiming Venture Partners. (Listed alphabetically by surname)

The Energy Revolution

Beware the "Beautiful Trap" of New Energy Investment

If this year's KrNotes could preserve only one sector, the answer would undoubtedly be: new energy.

"Dual carbon" is fundamentally a revolution of renewable energy resources (light, silicon, hydrogen, etc.) replacing carbon-based energy (coal, oil, gas, etc.). It primarily transforms four links in human energy utilization: production, transmission, storage, and application.

According to IT Juzi statistics, China's primary market new energy sector saw 285 investment and financing events in 2022, with over 20 companies achieving unicorn status, mainly in power batteries, new energy vehicles, and energy storage. Based on disclosed funding amounts, the average single financing round in new energy reached RMB 378 million.

In investors' view, because the new energy industry is sufficiently large, every segment within it can produce major companies. IDG Capital's Xinhua Yu notes, "Looking at solar ten years from now, we may see 1,000-1,200 GW of annual installations — roughly 5x today's market space. The same applies to electric vehicles: when market penetration reaches 80-90%, what scale of market will that represent?" In this process, many segments that currently appear small could represent markets of tens of billions of RMB. The flow of primary market capital into increasingly niche tracks within new energy confirms this trend.

"Made in China" is another source of certainty for the new energy industry. "New energy is fundamentally manufacturing." This was the consensus expressed by multiple investors. For China, which holds 38% of global manufacturing capacity, breakthroughs in core technologies for any niche segment can be combined with massive manufacturing scale to crush competitors on cost and volume.

Source Code Capital's Xingshi Wang adds that new energy has transformed the traditional energy industry, which was heavily dependent on natural resource attributes, into one dominated by manufacturing attributes — making it "inevitably subject to capacity cycles, inventory cycles, and downstream economic cycles."

Xianfeng Changqing's Yang Zhao draws a comparison with semiconductors: 50 years ago, Silicon Valley defined the regional distribution of the global semiconductor supply chain and core patents, creating long-term "chokepoint" problems. But new energy is different: "China has relatively complete independent intellectual property rights in hydrogen energy storage, automotive downstream applications, and many technologies are world-leading."

CCV's Wei Zhou believes the biggest difference between today's new energy industry and 15-20 years ago is that technology has reached the point of independent, mature commercialization. "When photovoltaic conversion efficiency was low and production processes imperfect, the positive energy effects generated over its lifecycle didn't match the emissions from production and subsequent processes — operating costs didn't pencil out. But now, even without subsidies, it can generate positive economic returns, enabling natural maturation."

Linear Capital's Yingzhe Zeng similarly views ROI as central to the industry. For example, the widening gap between peak and valley electricity prices directly improves the economics of energy storage, creating an inflection point for the sector.

For 2023 specifically, CASSTAR's Lei Mi believes that with capacity expansion, supply tightness in upstream raw materials will ease. Downstream, new energy vehicle market share is already relatively high, so absolute growth rates will slow, and capital market performance will be more stable than the past year.

Some divergent views exist. 5Y Capital's Kai Liu believes that while semiconductor inventory digestion is largely complete, "new energy, especially power batteries and related sectors, may be at a point of overcapacity — 2023 could be a peak, with improvement in 2024." PineVC Capital's Hua Feng, meanwhile, states that "the new energy supply chain's momentum may continue through 2023, with real demand still present."

Worth noting: a good company in a good industry does not automatically equal a good investment.

This problem shadows energy investment everywhere: return cycles mismatch with venture capital. A year ago, when we interviewed multiple veteran energy investors, nearly everyone raised this point. Though industry and capital environments are now vastly different, the problem persists.

"New energy investment has never been a field of abundant opportunity. From the last cycle onward, much of it was infrastructure-type investment," says Zhou. For VC's relatively limited range, some new energy sub-sectors require capital beyond what VC can supply. For example, lithium battery companies may start at RMB 10 billion valuations from day one — early-stage institutions invest heavily without securing proportional influence. VC should instead "use small inputs to leverage large innovation."

Some new energy investment may be better suited for PE and state-backed capital. Shanxing Capital's Shi Xu also points this out, emphasizing that we are indeed in a transition between old and new cycles, and new energy will remain a persistent investment theme, "but some specific areas aren't really suitable for VC, especially heavy-asset deployments where capital efficiency is low for venture capital."

New Energy Vehicle Growth Peaking, VC Attention Shifts Upstream

The leap of most industries from static to dynamic states often stems from downstream chain pressure. This feature is especially pronounced in new energy.

It is fair to say that new energy vehicles were the spark that ignited the past two years' energy revolution, and thus the most significant component of the energy sector. In 2022, over 30 new energy vehicle brands completed funding rounds in the primary market. Among these, GAC Aion's Series A of RMB 18.3 billion set a record for the largest single private financing in China's new energy vehicle industry.

Investors generally agree that new energy vehicle manufacturing has entered its second half. CPCA data shows that by late last year, domestic new energy passenger vehicle retail penetration had exceeded 30%, with domestic brands' penetration surpassing 50%.

IDG Capital's Yu states that electrification, connectivity, and intelligence are the core directions for future mobility development. The first step — electrification replacing fuel vehicles — has largely been achieved. "But that's only part of it. The more important development in mobility is intelligence. The true future car will be a 'mobile intelligent terminal.'"

Matrix Partners China's Huadong Wang further elaborates on the critical nodes of vehicle intelligence. He believes electrification provides the foundation for intelligence, which is the competitive core of the future automotive industry. The electronic/electrical architecture is the foundation of intelligence, with automotive semiconductors as its core component.

Gaohe Capital's Ming Jin believes China's new energy vehicle industry's global position closely resembles that of the mobile phone industry. "Apart from Tesla, other overseas manufacturers lag significantly behind China's leading OEMs." Product competitiveness is globally leading, but international relations challenges constrain overseas expansion.

Shangqi Capital's Xiangyuan Huang tells us that expected sales growth for new energy vehicles this year is lower than 2022's. "In fact, after two years of massive growth, the industry has entered a relatively mature stage — though compared to other industries, it's still high growth."

Kai Liu of 5Y Capital also points out that in 2023, downstream new energy industries represented by new energy vehicles are undergoing consolidation, with price competition being one manifestation. Feng Hua of Songhe Capital similarly believes that polarization in the new energy vehicle market will accelerate — "leading OEMs will get stronger," with greater flexibility on pricing, "but whether smaller manufacturers' supply chains can withstand losses is where real strength gets tested. The market ultimately won't have too many players left." Gaonan Zhang of Huaying Capital states that in the coming year, his focus on the new energy vehicle track will shift toward incremental supply chain opportunities.

Put simply, venture capital in pure new energy passenger vehicle OEMs has reached its twilight — the market is large and mature, targets are expensive, and the mismatch with primary markets is widening further. Entrepreneurs and investors are migrating toward new energy commercial vehicles, auto supply chains, and upstream segments of the new energy industry.

Upstream Energy: Bright Prospects, Bumpy Road

For primary markets in 2022, power batteries, energy storage, and hydrogen energy were all key themes in the new energy track.

On one hand, expanding new energy vehicle production capacity accelerated upstream power battery investment, with pure electric passenger vehicles contributing the bulk of installed capacity. Primary market capital flowed increasingly into niche segments, evolving from investing in battery manufacturers, to cathode/anode materials and electrolytes among other main and auxiliary materials, to aluminum-plastic film for battery casings and battery recycling.

On the other hand, with continued growth in wind and solar installations and accelerated 5G base station construction, demand for energy storage batteries grew rapidly. According to incomplete statistics, since 2022, over 120 energy storage industry chain companies have announced IPO progress, while more than 100 domestic companies had financing activity, with total fundraising exceeding 100 billion yuan. Capital concentration also triggered explosive growth in company numbers — nearly 30,000 new energy storage-related enterprises were established in 2022.

"As global green transition trends become certain, new energy transportation represented by lithium batteries and new energy power represented by solar and wind are gradually beginning to displace traditional energy in existing capacity. New energy is moving from backup to primary energy source." Matrix Partners China believes that in recent years, alongside the普及 of electrification and intelligence, dual-carbon policies, grid parity for solar and wind, and electricity cost spikes from the Russia-Ukraine war, demand for both power batteries and energy storage batteries has exploded. Among these, cell factory expansion and main/auxiliary battery materials present enormous opportunities.

"Energy storage is the fastest-growing segment across all new energy industries." Lei Mi of CAS Star notes that among three energy storage approaches, those suitable for VC investment are mainly: compressed air energy storage on the generation side (hydropower stations aren't suitable for technology investment), hydrogen energy storage on the transport side among chemical storage methods, and battery energy storage — electrochemical storage — on the user side.

Furthermore, future energy storage will advance toward "full-time-domain storage," with "long-duration storage" potentially becoming the next breakthrough. Yingzhe Zeng of Linear Capital believes that while most attention currently goes to electrochemical storage, new-type power systems for society as a whole will require storage at scales of seconds, hours, days, months, even years. For example, hour-level batteries can be applied to passenger EVs and other transport, while heavy transport, aerospace, and metallurgical industrial functions require day-plus scales. "Second-level frequency regulation technology is already relatively mature, mobile storage technology can basically meet requirements; pulling back to a longer view, there's currently a gap in long-duration storage."

On technical implementation, most investors believe there's no one-size-fits-all solution. Liquid, solid/semi-solid, sodium batteries, hydrogen fuel cells and other routes will coexist. Yang Zhao of K2VC told Anyong Waves, "Looking back over the past decade, beyond industrial applications, the biggest new energy application has been electric vehicles — because lithium battery technology advanced rapidly, costs dropped, and supply chains became complete enough to support EVs' rapid development." Meanwhile, beyond lithium batteries, "there's much discussion of alternative categories like sodium batteries, hydrogen fuel cells — various storage technologies are iterating quickly."

Among these, policy tailwinds at the top level directly propelled hydrogen energy investment into a hot primary market track. In 2022, national ministries issued over 50 relevant policies, with over 300 supporting local policies. Hydrogen fuel cell vehicles in particular will become the next heavily subsidized category after pure electric and plug-in models.

"The hydrogen energy industry chain is still in a relatively early stage, much like lithium batteries ten years ago — very suitable for early-stage investment." Zhao notes hydrogen's inherent advantages: high energy density, abundant natural reserves, low cost — hydrogen's average cost in later stages could very likely fall below gasoline. Hydrogen is among the best technical paths in new energy; it can not only replace high-carbon energy in traditional industry, but also serve as a storage means to smooth volatility from solar and wind, forming clear complementarity with lithium batteries.

Xiangyuan Huang of Shangqi Capital is similarly optimistic about hydrogen's future prospects. On sub-segments, current focus is on ion exchange membranes, target materials, and equipment materials for hydrogen supply and hydrogen production. On application scenarios, commercial vehicles and heavy trucks are priority landing points. The construction speed of hydrogen refueling stations and other supporting infrastructure is a key constraint.

Of course, as with any "trend effect" in venture capital history, some investors believe the hydrogen track has excessive hype, the industry is immature, and may not achieve scalable, sustainable application for a long time. This is evident from Elon Musk's repeated public skepticism of hydrogen and hydrogen fuel cells on social platforms, and multiple hydrogen energy IPOs being terminated.

Zhilong Fu of Yunxiu Capital notes that many technical barriers in upstream materials science are also constraints. For hydrogen energy stacks, upstream raw materials like membrane electrodes; for photovoltaic HJT cells, upstream raw materials like low-temperature silver paste — core material formulations remain in the hands of European, American, and Japanese companies, limiting the industry's leapfrog development. Hydrogen's practical application and corporate self-sustaining profitability remain to be tested over a longer period.

Also worth mentioning is controlled nuclear fusion. To multiple investors, nuclear fusion may be the ultimate energy solution — Lei Mi of CAS Star believes it's "like electric vehicles 20 years ago, with impact that will grow over time." But for such a "sci-fi" flavored track, investors are generally in a wait-and-see posture at this stage.

From "ignition success" to "practical application value," controlled nuclear fusion remains咫尺天涯. At least under current technical conditions, achieving the commercial threshold of "output energy > input energy" still requires clearing numerous hurdles. To borrow a globally universal quip: controlled nuclear fusion is always 50 years away from commercialization.

Semiconductors

Continued Pressure for the Next Two Years

After two years of breakneck growth, semiconductor investment in 2022 began exiting its supercycle and entering deeper waters.

"The craziness of several-fold annual gains is gone — even doubling is rare." Zhanxiang Zhao of Yunxiu Capital told us. Over the past two years, this financial advisory firm has served numerous major chip deals including Moore Threads, Axera, and InnoGrit.

Data from Juzi Technology shows that as of November 22, 2022, China's primary market saw 3,587 chip and semiconductor investment events, with total financing of 69.6414 billion yuan. Compared to 2021, both deal count and total amount declined.

Despite billion-yuan-plus semiconductor financings completed that year across CPU/GPU, automotive chips, semiconductor equipment/materials and other sub-segments, it was still palpable that chip design — previously highly sought-after — cooled in 2022. Wenrong Wang of Fortune Venture Capital believes that fewer and fewer quality chip design projects are visible above water, with many moving upstream toward equipment and materials, while equipment startup development patterns have already begun taking shape. Feng Zhao of Harvest Capital states that in 2022 they actively slowed their investment pace, and believes 2023 and 2024 will remain "years of pressure."

The cooling first stems from supply-demand adjustments. Most Chinese chip companies produce consumer electronics chips, and with smartphone and other consumer electronics shipments dropping sharply in 2022, chip demand plummeted accordingly. In fact, due to supply-demand imbalances in prior years, chips in some segments rapidly went from shortage to saturation. Multiple institutions note that in H1 2023, the chip sector will continue digesting existing inventory, with reasonable levels only reached in H2.

Second are high valuations. In 2022, GigaDevice Semiconductor and Advanced Optical Materials tied for top primary market semiconductor financing at 4.5 billion yuan each. Zhao says that due to excessive valuations, from 2021 the semiconductor sector began seeing many PE firms investing at VC stages, and VC firms at angel stages. He predicts this phenomenon will continue this year, though some PE firms, having been taught by the market, will "return to their proper places."

This Year Will Be Critical for Silicon Carbide

As a bulk strategic product combining both strategic and market characteristics, chips concern both information security and national security, while also needing to demonstrate value through market transactions — making their investment more intensely embody a subtle duality: certain demand comes from understanding of domestic substitution and autonomous control, while greater commercial imagination comes from the market.

In 2023, chip and semiconductor investment opportunities continue along these two paths. Under the former logic fall areas Wenrong Wang favors: "military electronics," "core semiconductor components including some consumables," and the Xinchuang (IT innovation) direction Zhao favors.

Military is a very prosperous segment in recent years. Regarding chips within military electronics, Wang believes "though the market ceiling isn't particularly high, revenue certainty is high with good margins, and the sector concept is strong." His investment in Chengchang Technology, which makes phased array T/R chips, exceeded 10 billion yuan market cap after its 2022 listing.

In Zhao's view, "key industries involving national welfare and people's livelihoods will need to adopt domestic chip hardware and software operating systems in the future," so "Xinchuang will be a long-term opportunity." In H2 2022, Xinchuang concepts already clearly became a secondary market theme.

Ming Jin of Gaohe Capital believes that in 2023, in advanced silicon-based semiconductor process nodes, trends toward line-width shrinkage and 3D structures have spawned multiplied thin-film equipment demand, where thin-film equipment has strong domestic substitution potential. Additionally, he favors key materials and components around lithography machines, and compound semiconductor industry chains. Gaonan Zhang of Huaying Capital believes that in new energy vehicles, photovoltaics, 5G and other third-generation semiconductor downstream applications, "China is the world's single largest market," with considerable domestic substitution space.

Under the second logic is another terminal that will help chips and semiconductors complete their grand narrative after consumer electronics and industrial applications: automotive. In 2023, directions Zhao favors include: smart vehicle supply chains, mainly autonomous driving, some pre-controllers, and the silicon carbide industry chain.

Lei Mi of CAS Star states that in 2022, while consumer chips cooled, automotive chips began rising. "After automotive new energyization comes intelligence, so large numbers of power chips and optical chips will be used in vehicles," and currently the pull from "IGBT, silicon carbide and other power chips" is very rapid.

As a third-generation compound semiconductor material, silicon carbide's properties — high voltage tolerance, high temperature resistance, low losses — give it distinct advantages in new energy vehicles, photovoltaics, rail transit, and smart grids. Currently, "design, substrate, equipment and other industry chain segments are all in shortage." In Zhao's view, if in 2023 "mainstream enterprises get financed and launch capacity expansion," "competition difficulty increases for later entrants." To him, this year will be critical for the silicon carbide industry.

Feng Hua of PineVC also believes that "silicon carbide is indeed a massive opportunity," but material integration innovation typically follows decade-long cycles, requiring ample patience and risk assessment.

Because new energy vehicles are expected to gain market share over the next three to five years, combined with chip shortages in recent years, domestic chips have been given opportunities to enter vehicles. Automotive semiconductors—represented by in-vehicle computing chips, power semiconductors, and sensor chips—along with autonomous driving, smart cockpits, and intelligent chassis, have become hot sectors for investment and financing in the semiconductor industry.

In Zhao Zhanxiang's view, "to avoid future supply chain security issues," new energy vehicles will "still cultivate domestic leading manufacturers for some key components."

Zhao Feng of Harvest Capital similarly believes that the next stage of opportunities will definitely go deeper into the foundational layer, the deep-water zones—such as equipment, materials, instruments, software, and other underlying technologies involved. In Wang Wenrong's view, while automotive chips do have high technical barriers and represent a compelling narrative, their cycles are relatively long. In the capital environment of recent years, numerous startups have emerged with inflated valuations, intensifying competition. Ultimately, only a handful will survive; who comes out on top remains to be seen.

Bai Zongyi of Glory Ventures also believes that if one were to position in the automotive intelligence industry chain now, investment decisions would be relatively difficult. Because in areas surrounding perception, decision-making, and execution—including "LiDAR, 4D millimeter-wave imaging radar, ADAS SoC, cockpit SoC, L3-L4 autonomous driving decision algorithms, automotive-grade MCU, and others"—relatively leading companies have already emerged. Although these companies still need time to commercialize, "the scale of funding raised and the import of commercialization certifications" have already formed certain competitive barriers.

Biotechnology

Life Sciences: Spring Is Coming Back, But a Different Spring

After the investment frenzy of 2020 and the sharp downturn in the second half of 2021, the biotech industry in 2022 remained stuck at a freezing point.

According to China Renaissance statistics, in 2022, there were 2,012 investment events in China's healthcare sector, with disclosed total investment of 179.83 billion yuan—the latter figure was 329.25 billion yuan in 2021, a year-on-year decrease of 45.4%.

For innovative drugs in particular, it was undoubtedly a year of "disillusionment." However, amid the painful forced transformation, some encouraging changes were also taking place in the innovative drug space.

Liu Dan of CDH Investments observed that companies are shifting from pure R&D-driven innovation to a "begin with the end in mind" clinical perspective, paying more attention to how much market application value a differentiated product can ultimately deliver. In Yu Zhiyun's view at Matrix Partners China, the innovative drug industry will transition from its current Biotech phase (focused on R&D, with little attention to revenue) into a Pharma era, where commercialization must generate substantial profits for shareholders.

Multiple investors agreed that innovative drugs will warm up in primary and secondary markets in 2023. "Because the industry and market still need genuine innovation," said Yang Kun of Gaorong Capital. "Through global COVID-19 prevention and control, whether vaccines or innovative drugs, people have seen the speed and power of technological innovation." However, unlike the previous wave of capital market dividends, "this investment warming may place greater emphasis on new product development driven by new technologies and new supply," Ma Rui of FreeS Fund told us.

Beyond innovative drugs, which sub-sectors present more investment opportunities?

Liu Dan believes that due to breakthroughs in technical approaches (such as vector delivery technologies breaking bottlenecks), the cell/gene therapy space is gradually moving into deeper waters. He defines the keyword as new modality—distinct from large and small molecules, representing an important transformation in technical pathways.

Yang Kun stated that Gaorong continues to focus on early-stage, truly internationally innovative targets, including breakthrough medical devices, drug molecules or platform technologies with actual clinical demand, and applications of life sciences in synthetic biology, agricultural technology, and other fields.

Matrix Partners China also mentioned that beyond continued investment in globally innovative technology platforms in biopharmaceuticals, medical devices, and diagnostics, they are paying attention to opportunities at the intersection of healthcare and other industries, with forward-looking positioning in life science tools and synthetic biology upstream in the industrial chain.

Cross-disciplinary frontier science was a term investors mentioned with particular frequency this year. In Ma Rui's view, neuroscience is also reaching an inflection point. The next ten years will be a golden period for developing innovative drugs and therapies targeting central nervous system diseases.

A signal mentioned by Matrix Partners China's Yu Zhiyun: in January 2023, Biogen/Eisai's Alzheimer's drug Lecanemab was finally approved by the FDA. Their previous-generation Alzheimer's drug, though approved, was hugely controversial, while Lecanemab's approval is widely recognized in the industry as a milestone event.

For a long time, brain science—extremely difficult to crack—has been a "holy grail" of life science research, with brain-computer interfaces being the hottest area in the past two years. Since 1924, when German psychiatrist Hans Berger first collected electrical signals from the human brain, brain-computer interfaces have been a "century-long project." However, with advances in AI, electrode materials, and information acquisition capabilities, brain-computer interface technology is beginning to move from the laboratory into reality.

Liu Dan defines brain-computer interfaces in China as being at a stage of "ice-breaking application exploration." In the past, only non-invasive brain-computer interface companies existed, with relatively mature application scenarios; now the field is moving toward serious scientific applications, namely invasive brain-computer interfaces. Elon Musk's Neuralink is a typical example of the latter.

Hongshan is currently investing at high frequency in the brain science track. In Yang Yunxia's view at Hongshan, brain-computer interfaces already have very clear application scenarios, such as deep brain stimulation (DBS), commonly known as brain pacemakers. "The fastest-growing application area right now is actually in the field of decoding the human brain. We look forward to the day when humans can map out a 'brain atlas,'" Yang Yunxia said.

Overall, biotech investment in 2022 trended toward greater sobriety, and across nearly all sub-sectors, the trend of investing early and small remained pronounced.

Synthetic Biology: Will Remain Depressed, Investors No Longer Pay for Concepts

One branch of the life health industry is synthetic biology—"lukewarm" on the industrial side, yet scorching hot as a concept. It is essentially an underlying technology, a new mode of production. Its applications are extremely broad, including chemical products, food ingredients, environmental treatment, bioenergy, pharmaceutical intermediates, and more.

Since 2018, China's synthetic biology has seen a new wave of financing highs. In 2022, domestic synthetic biology companies announced at least 43 funding rounds, with some completing two rounds in a single year.

But as 36Kr once统计, this industry with a current global market size of less than 100 billion yuan saw total annual funding in 2022 of less than 10 billion yuan, with over 40% of funding created by the top four companies—a highly pronounced "winner-take-most" pattern.

The core constraint on synthetic biology's development may lie in its costs. Zhang Jinjian of Oasis Capital stated, "Cost is a very important factor." Shi Xuesong of Chuanliu Investment noted that limited by cost and technology, synthetic biology's commercial applications remain relatively narrow, mainly holding advantages in high value-added products and areas where chemical synthesis is difficult. "From a development stage perspective, synthetic biology is roughly equivalent to the internet in 2000; reaching the widely anticipated broad applications may still require technological progress."

Multiple investors expressed to us that the synthetic biology industry remains in an early stage, with most companies yet to reach the critical phase of industrialization. However, in areas such as bulk new materials, primary market companies have reached a critical point of achieving commercial closed loops.

What about synthetic biology's future? Mid- and downstream synthetic biology companies may become investors' "main battlefield," said Liu Dan of CDH. "Previous investment mainly concentrated on mid- and downstream companies; downstream is closest to monetization, midstream depends on R&D capabilities."

Though synthetic biology is currently rather depressed, many investors still recognize its commercial value and long-term potential.

Liu Dan believes synthetic biology technology is rapidly maturing, with commercial application scenarios emerging. "We continue to be optimistic about the broad application prospects of efficiently engineered cell pathways in new drug R&D, cell engineering, and high-end biomanufacturing," Yu Zhiyun of Matrix Partners China said.

Ma Rui of FreeS Fund believes that for synthetic biology companies to become profitable and sustainably so, the road ahead is long and difficult. But regardless, the reconstruction of over one-third of the molecular world, and the infinite possibilities behind it, is thrilling. After all, whichever country primarily controls raw materials, molecules, commodities, and manufacturing can control inflation, interest rates, and finance.

Artificial Intelligence

Investors Divided, Breaking Through Commercialization Is the Only Consensus

Over the past decade, AI investment waves have risen and fallen repeatedly, but have consistently struggled to land due to the commercialization challenge. In 2022, as RMB funds with stronger preference for commercialization capability gained strength, AI's further cooling in the Chinese market was hardly surprising.

Data shows that as of November 10, 2022, there were 400 financing events in China's AI industry in primary markets this year, down 50% year-on-year; estimated financing transaction value was 77.04 billion yuan, down 61% year-on-year.

But across the ocean, strange flowers still bloomed. In late 2022, ChatGPT, born from the continuous scaling of Transformer algorithms, reached millions of C-end users within a week, providing new narrative space for AI-generated text.

AIGC, or AI-generated content, by modality can be divided into audio generation, text generation, image generation, video generation, and other sub-sectors, as well as cross-modal generation such as text-to-image and text-to-video.

2022 is considered the inaugural year of AIGC. Dark Waves once interviewed multiple investors about ChatGPT; one view held that AIGC might become a platform-level track.

ChatGPT's birth has, to some extent, shaken domestic investors' strategy of investing in AI ToB. ChatGPT does not directly target developers, but first targets ToC scenarios, allowing developers to see C-end enthusiasm for the product, after which more people will start businesses based on it. Chen Weiguang of Bluerun Ventures believes that AI entrepreneurship must think about business model integration from day one, rather than thinking about it after scaling traffic.

But new commercialization thinking for ToC scenarios still has problems. This is also why CCV's Zhou Wei intensively reviewed AIGC projects but ultimately did not pull the trigger: first, beyond OpenAI, no AI company has a large lab, and beyond ChatGPT, no product shows differentiation. Second, paying users are unclear, payment willingness is uncertain, and payment methods are undefined. And he believes that if AIGC goes ToB, it must seize the industry's greatest pain points and produce explosive, bottleneck-breaking multiplier effects for the other party's payment willingness to be strong.

Regarding investment in the AIGC field, more people believe there will be more opportunities on the infrastructure side, perhaps also the tier with the greatest short-term commercial returns.

In Zhou Wei's view, the application layer finds it too difficult to compete with major tech companies for traffic, while underlying technology is the source of differentiation; therefore "large models" as infrastructure have become the key direction in most investors' minds.

But large models have their own "route disputes": for instance, people are still debating whether the future will bring general-purpose large models, or foundational models tailored to each specific domain. In the view of Kai Liu of 5Y Capital, the "battle of large models" at the current stage of AI development can be analogized to the evolution of cloud computing. In its early days, cloud computing's upfront costs were prohibitively high, so companies running public clouds handled unified operations and maintenance — a model that naturally suited smaller companies with needs but limited resources. The bottleneck emerged when client companies grew to a certain scale, at which point building private clouds became more cost-effective. However, for AI's current early stage, Liu predicts that unified large models will remain the norm for the next several years.

Still, Bobing Ren of Sinovation Ventures notes that China currently has very few investable infrastructure companies related to large models, and that most application-layer ecosystems still lag behind the US to some degree.

Another problem with large models is cost. OpenAI, which raised $1 billion last year, is the primary contributor to large models. Its GPT-3 series was trained on staggeringly diverse data — the entire English Wikipedia comprised just 0.6% of its training corpus — keeping costs persistently high. GPT-4, widely expected to debut in 2023, could push costs into the hundreds of millions of dollars.

Reducing cost has become a crucial investment direction on the infrastructure side. Liu says 5Y Capital made several bets in 2022 on platform-collaboration and tooling companies within the infrastructure layer, all sharing the common goal of driving down costs and lowering barriers to entry.

For AI investment, finding vertical application scenarios is also critically important. Zongyi Bai of Glory Ventures told us: "AI is a new form of productive force. We need to view industry development opportunities through the lens of complete industrial chains — from data perception, transmission, storage, and computation to data/AI infrastructure, as well as vertical industry applications including robotics, autonomous driving, enterprise SaaS, and the digital and intelligent transformation of high-end manufacturing in sectors like new energy and consumer electronics." In the infrastructure domain, greater emphasis is placed on a founding team's technical moat and tight human-resource fit, plus their know-how in AI+vertical industry applications, productization, market expansion, and delivery capabilities.

In interviews, many investors mentioned that AI is merely middleware. Weiguang Chen shares this view, believing that AI cannot be examined in isolation. He is more convinced that the "triple wave" of AI + Web3 + 3D interaction will define the next major technology cycle.

Another frequently heard phrase is the arrival of the "democratization era." Jinjian Zhang of Oasis Capital believes that AIGC will penetrate more deeply into everyday life scenarios in 2023 — this is the year of landing and permeation. But more general-purpose artificial intelligence as we understand it (such as autonomous driving) is more likely to emerge between 2025 and 2027.

Zhifeng Zhou of Qiming Venture Partners argues that the core significance of the generalization capability provided by large models is that they allow everyone to use them directly, enabling AI capabilities to land more effectively across various scenarios. Large models are like centralized infrastructure laying down the foundational "utilities" — water, electricity, gas — that entrepreneurs and enterprises can tap into directly. They need only focus on deploying end applications for niche scenarios, rather than the previous high-R&D-difficulty, vertical "chimney-style" entrepreneurship paradigm that AI once demanded.

One industry consensus about artificial intelligence: although the current wave of AI has been running for ten years since the birth of neural networks in 2012, it remains in an early stage. AI technology penetration across all industries is still below 30%, with most applications limited to relatively superficial commercial nodes.

Qiming Venture Partners remained one of the most active investors in the AI sector in 2022. For Zhifeng Zhou, AI continues to be one of the most exciting investment themes, because the new technical paradigm of large models plus fine-tuning — a breakthrough in underlying technology and engineering — allows startups to access AI capabilities through low-cost cloud service APIs rather than building models from scratch. This dramatically reduces development costs and enables faster, more effective application deployment. Zhou therefore believes that AI has clearly reached a new inflection point at this moment, a critical turning point after more than a decade of deep learning development.

Wei Zhou believes that for AI entrepreneurs to raise funding smoothly in 2023, the key remains commercial progress — AI technology must help industries solve critical bottleneck problems, not merely be "nice-to-have" additions. Bobing Ren points out that confidence in 2023 should be more grounded, for two reasons: first, AI as "technological infrastructure" continues to iterate rapidly, penetrating further across industries while becoming more democratized; second, the creativity, comprehension, and generalization capabilities based on large models will continue to improve. He believes 2023 will be another year of explosive growth for large model technology and AIGC applications.

Commercial Space

More Landmark Events, the Singularity Moment Approaches

Since China began encouraging private enterprises and capital to enter the space sector in 2015, an increasing number of investors have come to recognize commercial space as one of the hard-tech arenas where Chinese companies have a genuine shot at claiming significant global standing.

Yet what hangs over this field is more "non-consensus." Dark Tide Waves' research found three critical questions in commercial space that remain unresolved (and perhaps temporarily unanswerable): orbit-first versus reusability-first, solid versus liquid fuel, small rockets versus large rockets.

With SpaceX as the precedent, and cost reduction as the ultimate goal, making rockets "reusable" became an early industry consensus. But whether startups should first develop rockets capable of successfully delivering satellites to their designated orbits, or first experiment with reusability technology, has split the field into two camps.

The former approach is represented by Galactic Energy. On January 9, its Ceres-1 Y5 carrier rocket launched five satellites into orbit, marking the successful first commercial space launch of 2023. This was the fifth consecutive successful launch for Galactic Energy's self-developed commercial carrier rocket, having delivered 19 commercial satellites to orbit for clients — prioritizing commercial self-sufficiency over complete reliance on external funding. The latter approach is represented by Deep Blue Aerospace, which has completed three consecutive liquid rocket recovery experiments at ten-meter, hundred-meter, and kilometer altitudes.

In rocket technology, the long-standing "solid-liquid debate" persists: whether solid-fuel or liquid-fuel rockets are better suited for space launch. Solid rockets have a longer history and mature technology, require no fueling, offer high reliability, and have significantly shorter launch cycles than liquid rockets — but they cannot be recovered. Liquid rockets offer higher specific impulse and greater payload capacity advantages, but their required engines are far more complex, demanding substantial R&D investment and high launch costs.

A world-class "curse" haunts the space industry: after SpaceX, every private commercial space company's first liquid rocket launch has ended in failure, including LandSpace, which had the highest probability of success domestically. This means that for the more difficult liquid rockets, no Chinese commercial space company has yet achieved normal orbital insertion.

The small-versus-large rocket debate stems from differing market demands: lower-cost small solid rockets can capture fragmented launches for micro and nano satellites; medium-to-large liquid rockets are needed for low-orbit constellation deployment. As the number of satellites on the same orbital plane increases, with launch requirements frequently reaching dozens of tons, rocket payload capacity becomes critical.

Multiple investors believe that with policy support and technological breakthroughs, commercial space will reach a critical juncture in the next two years. Jianwei Li of Zhencheng Investment judges that "solid rockets capable of orbital insertion are already moving toward commercialization; 2024 should see liquid reusable rockets achieve orbital insertion, with recovery possible by year-end."

Zhencheng Investment began studying SpaceX's rise in 2015 and has since evaluated nearly every Chinese commercial space company, but Li has always "held reservations" about solid rockets. He believes solid rockets cannot achieve the ultimate goal — reducing per-kilogram orbital insertion costs. "Solid costs are 100,000 RMB per kilogram; liquid rockets might drop to 60,000 or 70,000. But the fundamental change with reusable rockets is that returning the rocket's most valuable component, the engine, saves 70% of costs and dramatically improves margins."

In the 1990s, countless investors suffered heavy losses in the space sector — Bill Gates, for instance, invested substantial sums only to be deeply disappointed. The impasse was broken by Elon Musk's SpaceX, whose approach was simple in description: reduce costs, by 10x, by 100x. Chinese commercial space companies currently pay over $15,000 per kilogram for launch costs, five times that of SpaceX's Falcon 9.

This means that whoever first completes liquid rocket orbital insertion plus recovery will trigger a crushing transformation in cost structure. And the ranking of leading commercial space companies will face a major reshuffling: companies that have yet to achieve successful launches will be partially disproven and disappear, with investment rapidly differentiating.

Li believes the ideal target is: liquid rocket orbital insertion plus recovery. "Everyone shouldn't be redundantly developing many models." Shi Xu of Shanchuan Capital sees demand-side signs gradually emerging in commercial space. The State Council issued Guiding Opinions on Encouraging Social Investment in Innovation in Key Investment and Financing Mechanisms as early as 2014. After eight years of development, a preliminary ecosystem has formed, with a certain number of enterprises in satellite manufacturing, rocket R&D and production, launch services, and ground measurement and control.

Li also believes that "the next three years will be the singularity for Chinese commercial space." Following current R&D trajectories, before 2025 there is a high probability that domestic commercial space companies will achieve orbital insertion with recoverable liquid rockets. At that point, payload orbital insertion costs could drop from the current 60,000–100,000 RMB per kilogram to 20,000–30,000 RMB per kilogram, dramatically reducing orbital costs. Meanwhile, as recovery technology matures, rocket launch frequency will also increase. Subsequently, triple-core heavy reusable rockets capable of carrying larger payloads will enter the agenda.

Commercial space is the arena that most viscerally reflects how technology and technology investment will transform humanity. The "fuzzily correct" and "precisely wrong" judgments that entrepreneurs and investors encounter here represent the ultimate experience of confronting technological revolution. This process is undeniably thrilling, but it is destined to be long.

Image source | Visual China

Layout | Du Meng