Rising Industry Penetration Does Not Equal Shareholder Returns: Reflections on Sector Investing in 2026

Data as of 2026-06-30. Past index performance does not guarantee future results.

Disclaimer: This post is solely my personal reflection and review of publicly available industry materials and business logic. It does not constitute investment advice of any kind, nor does it involve any individual stock recommendations or implied holdings. Markets carry risks; make your own independent decisions.

Nature of research: A methodological review based on public materials, not personalized investment advice. Industry data is primarily through May 2026, index data through June 30, 2026, and liquidity information through July 13, 2026.

I am revisiting China Merchants Securities' 2021 strategy report, *The Revelation of Sector Investing: Four Selection Rules and Risk Signals*. That piece explained how industry trends generate major market rallies. I still use one of its core frameworks, BEAT, which combines industry trends, earnings realization, consensus formation, and investable targets. I also keep its four categories of risk signals: thesis falsification, earnings misses, market cap exhaustion, and liquidity tightening.

Data from 2026 fills in a piece that was missing from that framework. My judgment is that rising industry penetration does not guarantee higher industry profits, much less excess returns for public equity shareholders. New energy vehicles (NEVs) are the clearest example. In May 2026, NEV sales accounted for roughly 56.9% of total auto sales, or about 61.7% on a domestic sales basis, approaching or crossing the 60% maturity threshold defined in that report. Over the same period, automotive manufacturing revenue still grew 1.4%, yet profits fell 19.8%, leaving an operating profit margin of only about 3.42%. The industry trend delivered as promised, but price competition, capacity expansion, and cost structures squeezed the profit pool. China Association of Automobile Manufacturers (CAAM), official association data; National Bureau of Statistics (NBS), official statistics

Penetration Materialized, Profits Did Not

Because of this, I have downgraded "low penetration" to a mere clue and now prioritize checking six things:

1. Whether demand has already shown up in orders, unit sales, installations, and revenue.
2. Which segment of the value chain captures the incremental profit.
3. Whether supply is expanding faster than demand, and whether prices and gross margins are deteriorating.
4. How much future growth is already implied by the current market cap.
5. Whether ETFs, indices, and passive flows have created a crowded trade.
6. Which falsifiable earnings milestones exist over the next one to four quarters.

To me, a good sector is only a starting point. Capturing excess returns requires verifying that demand is real, supply is hard to replicate, profits can be retained, and current prices have not already baked in those expectations. If even one link is missing, I will not treat a penetration story as a conclusion.

Three Things I Still Keep

Earnings Still Drive the Medium Term, and Broad Sector Labels Are Too Coarse

In January–May 2026, profits at industrial enterprises above designated size nationwide rose 18.8% year-on-year, but the divergence across industries was massive:

• Computer, communication, and other electronic equipment manufacturing profits grew 103.9%.
• Automotive manufacturing profits fell 19.8%.
• Electrical machinery and equipment manufacturing profits fell 13.7%.
• Pharmaceutical manufacturing profits fell 1.3%.
• Optoelectronic devices and semiconductor discrete devices saw profits grow 53.8% and 40.6%, respectively.
• Electronic specialty materials manufacturing profits surged 665.4%, though low base effects and statistical scope adjustments need to be kept in mind.

I still agree that the medium-term market ultimately comes down to earnings, but I now discount broad sector labels. Grouped under the same "tech" or "new energy" banner, different links in the value chain are moving in completely opposite profit directions. National Bureau of Statistics (NBS), official statistics and analysis

Rallies Can Still Accelerate After Consensus Forms

Semiconductors are the strongest piece of evidence right now. The World Semiconductor Trade Statistics (WSTS) organization projects the global semiconductor market to grow by more than 25% in 2026 to $975 billion, with both logic and memory chips growing over 30%. The SIA reported global sales of $88.8 billion in February 2026, up 61.8% year-on-year. WSTS, international industry organization; SIA, international industry association

The A-share market has been accelerating in tandem. As of June 30, 2026:

• The CSI All Share Semiconductor Products and Equipment Index is up 105.08% year-to-date and up 187.21% over the past year.
• The CSI Semiconductor Industry Index is up 129.35% year-to-date and up 256.83% over the past year.

This shows that the idea that "consensus formation does not mean an immediate top" still holds true. CSI All Share Semiconductor Products and Equipment Index Factsheet, official index provider material; CSI Semiconductor Industry Index Factsheet, official index provider material

However, the same factsheets show that the trailing P/E ratios of these two indices have reached 128.42x and 182.09x, with P/B ratios at 13.1x and 17.7x, respectively. While I agree a rally can keep accelerating after consensus forms and I do not treat consensus as a mechanical sell signal, I also apply a discount to it, because high consensus raises the earnings bar and makes potential drawdowns much more asymmetric. At this stage, I no longer just ask whether the industry trend is right; I look at whether actual profits can beat the expectations already embedded in the price.

Tradable Targets and Capital Capacity Matter More Than in 2021

At the end of 2025, total onshore ETF assets reached 6.02 trillion yuan, up 61.4% year-on-year, with equity ETFs accounting for 3.83 trillion yuan. Institutional ownership of Shanghai Stock Exchange ETFs reached 65%, up 6 percentage points from 2024. In 2025, 100 new sector and thematic ETFs were listed, covering tracks such as artificial intelligence, innovative drugs, chips, and robotics. Shanghai Stock Exchange *ETF Industry Development Report (2026)*, official exchange report

That report required a sufficiently large pool of investable stocks, and I agree with that. By 2026, the change I need to add is how consensus spreads: in 2021, investors paid more attention to sell-side reports and technical chart patterns, whereas today we also have to watch index inclusions, ETF creations and redemptions, weight concentration, and passive fund flows. The price impact of leading companies now comes from both fundamentals and index weights and liquidity mechanics.

Four Claims I No Longer Treat as Laws

"10% to 60%" Is Not a Universal Law

That article used an S-curve to describe an industry's journey from early adoption to maturity, treating 60% as a maturity threshold. I have kept the practice of using S-curves to describe new car sales substitution, but I no longer treat 60% as a universal threshold. Without a clearly defined denominator, that percentage means nothing to me.

In May 2026, CAAM-reported NEV sales were 1.496 million vehicles out of 2.629 million total auto sales, or roughly 56.9%. Domestic NEV sales were 1.049 million vehicles out of 1.7 million total domestic auto sales, putting the domestic share at about 61.7%. For January–May, the cumulative shares were roughly 47.5% and 48.7%, respectively. Over the same period, exports accounted for about 29.8% of NEV sales and roughly 31.6% cumulatively. All of these ratios are calculated from sales figures published by CAAM. NEV Data; Total Automotive Volume Data

Change the denominator, and you get a different conclusion:

• Penetration of new vehicle sales.
• NEV share of the total active vehicle fleet.
• NEV share of corporate revenue.
• NEV share of the industry profit pool.
• Domestic substitution rate in sub-segments like batteries, charging, and smart driving.

For B2B industries like semiconductor equipment, AI optical interconnects, energy storage, and robotics, a single penetration rate is even harder to apply. In those areas, checking customer capital expenditure, order backlogs, capacity utilization, ASP, gross margins, free cash flow, and customer qualification cycles is much more useful.

Consolidation Around Leaders at Maturity Still Depends on Whether the Profit Pool Survives

That report summarized the mature phase as consolidation around industry leaders. That pattern is common, but fierce competition can also compress margins across an entire industry. NEVs have entered a high-penetration stage where leading players still hold scale and brand advantages, yet total industry profits are shrinking.

As of June 30, 2026, the CSI New Energy Vehicles Index posted the following numbers:

• Down 1.61% year-to-date.
• Up 45.40% over the past year.
• Annualized return of -3.34% over the past five years.
• Total constituent market cap of roughly 4.56 trillion yuan, and index market cap of roughly 2.00 trillion yuan.

Long-term industry growth, periodic rebounds, and poor five-year shareholder returns can clearly exist at the same time. CSI New Energy Vehicles Index Factsheet, official index provider material

So while I still watch for consolidation around leaders, I will not automatically buy into a leading company when industry profits are falling. I also check unit economics, whether market share gains are bought with price cuts, capital expenditure intensity, inventory and accounts receivable, overseas profits, free cash flow, and shifts in technology routes.

"Total Market Cap Matters More Than P/E" Is Only Half Right

Total market cap does put a ceiling on runaway imagination, but it cannot be used in isolation to judge whether something is cheap or expensive. In 2026, I need to tie market cap directly to the profit pool a company can actually capture:

Supportable Market Cap ≈ Capturable Revenue × Steady-State Profit Margin × Reasonable Valuation Multiple, minus capital expenditure, financing dilution, and cyclical risk.

Right now, the constituents of the CSI Semiconductor Industry Index have a combined market cap of roughly 6.05 trillion yuan and a trailing P/E of 182.09x. Industry demand and profits are indeed growing rapidly, but the index price already demands that this high growth persist for quite a long time. Looking only at total market cap makes it easy to miss earnings quality, while looking only at P/E makes it easy to prematurely write off companies in the early stages of profit release. Both metrics have to be viewed alongside orders, capacity, and cash flow. CSI Semiconductor Industry Index Factsheet

The Kelly Criterion Derivation of "Four Sectors, Sixteen Stocks" Does Not Hold Up

That article reduced "picking a top 10% sector" to fixed odds and a fixed win rate, and from there derived a portfolio allocation of four sectors with four stocks each. That derivation cannot be copied directly for several reasons:

• Subjective win rates cannot be estimated stably.
• Returns follow a continuous distribution, not a binary bet where you win 10 points or lose 1 point.
• Sectors driven by the same macro factors are highly correlated and violate the assumption of independent, repeated trials.
• Drawdowns, liquidity crunches, gap risk, policy shifts, and valuation compression all have fat tails.
• The "scores" in that text represent grading points rather than wealth multipliers, so they cannot be plugged directly into the Kelly formula.
• The portfolio construction ignores correlation, volatility, and maximum tolerable drawdown.

I cannot accept that derivation. My current approach is to first set a risk budget for each sector, then rank research priorities by correlation, volatility, valuation, and verification strength. Without stable odds data, I would rather rely on maximum drawdown constraints and staged verification than lean on a pseudo-precise win rate.

Re-evaluating Five Sectors with the Same Set of Questions

Sector	Current Industry Evidence	Market Pricing	BEAT Stage Judgment	Key Risks	Conclusion
AI Compute, Semiconductors, Upstream Electronics	January–May electronics industry profits up 103.9%; WSTS projects global market growth over 25% in 2026	Semiconductor indices up 105% to 129% YTD, P/E around 128x to 182x	B/E/A/T all established; already high consensus	Valuation exhaustion, export controls, memory cycle, slowing CapEx slope	Strongest industry evidence, but also highest valuation and crowding; focus research on sub-segment bottlenecks not yet fully priced in rather than broad sector beta
New Energy Vehicles, Lithium Batteries	May sales penetration ~56.9%, domestic basis ~61.7%; exports ~29.8% of sales	Index down 1.61% YTD, five-year annualized -3.34%	Moved past the thematic track phase into mature competition	Price wars, falling profits, overcapacity, tech transitions, overseas barriers	Direction has materialized; shift research focus to profits, cash flow, and global share
Energy Storage, Power Grid, Energy Management	Cumulative new-type energy storage reached 144.7GW by end-2025, up 85% YoY; lithium batteries still account for over 96% of new-type storage	Energy Storage Industry Index down 4.46% YTD, P/E 26.4x, up 29% over the past year	Relatively strong B/E, A/T in place, priced lower than semiconductors	Cell commoditization, falling prices, safety, utilization hours, and revenue mechanisms	Clear installation growth and lower valuation than semiconductors; break research down into differentiated links like PCS, EMS, grid-side operations, and long-duration storage
Robotics, Embodied AI	May industrial robot output up 27.9%, robot reducers up 44.7%; policy targets 10,000-unit-scale deployment capability by end-2026	Robotics Index up 12.85% YTD, up 35.07% over the past year	Relatively strong B/A/T, E in early verification	Few actual humanoid robot orders, insufficient reliability and ROI, blurry thematic mapping	Early-to-mid-stage sector; I use orders and scenario ROI to separate real demand from demos
Innovative Drugs	May biopharmaceutical manufacturing value-added up 14.7%; January–May pharma manufacturing profits down 1.3%, investment down 9%; basic medical insurance and commercial insurance innovative drug catalog systems continue to advance	Innovative Drugs Index down 4.35% YTD, P/E 27.33x	Strong policy B, diverging industry E, A/T in place	Clinical failures, insurance price negotiations, financing and commercialization misses	Better to track product and BD realization bottom-up; do not substitute industry averages for company research

The conclusions in the table reflect my own trade-offs: strong industry evidence does not mean the price still offers a margin of safety. The index return, valuation, and market cap data for robotics, energy storage, and innovative drugs in the table are all as of June 30, 2026, sourced from official CSI factsheets: CSI Robotics Index, CSI Energy Storage Industry Index, and CSI Innovative Drugs Industry Index.

AI Compute and Semiconductors: Strongest Fundamentals, Lowest Price Tolerance for Error

According to NBS data, integrated circuit output rose 22.9% in May 2026, electronics industry profits jumped 103.9% in January–May, and high-tech manufacturing profits grew 44.7%. At the same time, semiconductor indices have experienced extreme rallies and high valuations. NBS Industrial Production Data

I agree the fundamentals here are very strong, and I also apply a discount for the sector's overall valuation and crowding. In my research, I prefer to break "semiconductors" down into the following sub-segments rather than buying the whole story wholesale:

• High-end compute chips and domestic substitution.
• HBM, DRAM, enterprise SSDs, and the caching stack.
• Advanced packaging, semiconductor equipment, and electronic materials.
• Optical chips, DSPs, 1.6T optical transceivers, and CPO.
• Power supply, thermal management, and data center infrastructure.

The key judgment here is whether profits from this boom actually land in specific links of the chain, and whether valuations have already front-run two to three years of optimistic scenarios. I wrote about this same verification chain in The Optical Interconnect Bottleneck Must Be Broken Down into Optical Chips, DSPs, and CPO Packaging, Long-Term Bottlenecks in AI Inference Costs May Shift from Compute to Cache, and Overpricing Bottlenecks Hurts AI CapEx.

New Energy Vehicles: The Industry Succeeded, and the Thematic Narrative Has Basically Run Its Course

CAAM data shows that in January–May 2026, cumulative NEV sales grew just 3.5%, with domestic sales falling 16.5% while exports increased by 1.1 times the prior-year volume. In the single month of May, domestic sales fell 4.1% year-on-year, while exports increased by 1.1 times the prior-year volume. China Association of Automobile Manufacturers (CAAM), official association data

The industry's growth engine is shifting from rapid domestic penetration gains to exports, global brand expansion, smart features, and competition over the existing fleet. I no longer treat sales penetration as the primary driver; instead, I focus on whether profits, cash flow, exports, and smart capabilities can genuinely turn into revenue. The new energy sector that sat front and center in 2021 looks much closer to a mature industry today, and what to watch in each link has changed accordingly:

• For automakers: brand, product cycle, overseas channels, and profit per vehicle.
• For batteries: global customers, energy storage share, technology routes, and free cash flow.
• For materials: cost curves, supply-demand clearing, and industry concentration.
• For smart driving: actual paid take-rates, liability boundaries, and software revenue.

Energy Storage: Installations Are Rising, Value Capture Needs to Be Unpacked Again

By the end of 2025, China's cumulative installed capacity of new-type energy storage reached 144.7GW, up 85% year-on-year and 45 times the level at the end of the Thirteenth Five-Year Plan; new-type energy storage accounted for more than two-thirds of total installed power storage capacity. National Energy Administration reprint of industry database, official platform, original stats from CNESA partial survey

This source has two limitations: the data comes from an incomplete statistical survey by CNESA DataLink, and installation growth does not equal better project returns. I accept the fact that installations are growing, but liquid lithium-ion batteries currently account for over 96% of new-type energy storage, and the battery cell segment is already mature and heavily contested, so I discount the cell layer. Going forward, I focus on who can stably capture capacity payments, peak-valley price spreads, and ancillary service revenue as power market rules evolve, and which links among PCS, EMS, thermal management, long-duration storage, and system integration possess differentiated moats. National Energy Administration, technical and industrial review

Robotics: In the Verification Phase, Not Yet in the Large-Scale Profit Phase

NBS data shows that in May 2026, output of robot reducers, industrial robots, and service robots grew 44.7%, 27.9%, and 19.8%, respectively. A joint special action plan by MIIT and SASAC sets a target of building over 100 high-value scenarios and 10,000-unit-scale deployment capabilities by the end of 2026. National Bureau of Statistics (NBS), official statistics; MIIT and SASAC, policy document

Output numbers and policy targets do support the industry direction, but "10,000-unit-scale deployment capability" is a policy goal, which does not mean companies have already realized unit sales or profits. Until return on investment is proven in specific scenarios, I keep robotics in the verification phase, tracking real-world task success rates, continuous operating hours, cost per labor hour, customer repurchases, leasing payback periods, core component gross margins, and after-sales servicing costs.

Innovative Drugs: Policy Is Advancing, Industry Profits Remain Divergent

The 2026 adjustment for the National Reimbursement Drug List (NRDL) and the commercial health insurance innovative drug catalog received a total of 818 corporate applications covering 674 generic drug names; 334 out-of-catalog basic medical insurance applications passed preliminary review, and 53 out-of-catalog commercial insurance innovative drug applications passed preliminary review. Passing preliminary review only means passing a formal compliance check, and it does not mean inclusion in the catalog or commercial success. National Healthcare Security Administration (NHSA), official notice

The policy side continues to support innovative drugs, but pharmaceutical manufacturing profits and investment under NBS statistics remain weak, so I discount the industry average. The divergence between the broader pharma sector and companies actually delivering products will continue to widen. In this sector, I mainly watch clinical data, approval progress, insurance access, overseas out-licensing, actual revenue, and cash burn.

How I Use the Four Risk Signals Today

I have kept all four of the original risk signals, but added supplementary conditions to each one.

Are the Growth Curve and the Profit Pool Slowing Down Together?

The old signal was called "penetration breaking through the threshold or thesis falsification." I kept this signal, but modified it to look at both the growth curve and the profit pool simultaneously:

• Incremental sales, active fleet share, and revenue penetration.
• Whether the incremental market is shifting toward exports or other regions.
• Whether prices, gross margins, and ROIC are falling as penetration rises.
• Whether new technologies are bypassing existing supply chain bottlenecks.

Right now, NEVs have already entered this stage; energy storage and robotics still need their denominators defined first; and semiconductors must be measured separately by specific product and domestic substitution link.

Earnings Misses Cannot Be Judged by Single-Quarter Net Profit Alone

An earnings miss is still the signal I weight most heavily, but looking solely at single-quarter net profit is no longer enough. I break it down across five layers:

1. Orders and backlog.
2. Revenue and ASP.
3. Gross margin.
4. Operating cash flow and free cash flow.
5. Management guidance and next quarter's verification milestones.

If demand is strong but gross margins are falling, the company may lack pricing power; if profits are growing but cash flow is deteriorating, there may be a problem with inventory or accounts receivable.

Implied Expectations Are Too High

The old signal was "unreasonable total market cap." I have adjusted this to "excessive implied expectations," checking at least four items at once:

• Future revenue and profit implied by the current market cap.
• The market share, pricing, and capital expenditure required to achieve that profit.
• Steady-state gross margins after competitors expand capacity.
• The difficulty of meeting implied expectations, and the magnitude of potential shortfalls, across bull, base, and bear scenarios.

The current high valuation of semiconductor indices is the clearest observation signal here. I do not treat high valuation as a sign that the trend is ending immediately, but I use it to remind myself that subsequent performance depends much more heavily on earnings continuously beating expectations.

Loose Liquidity Does Not Mean Every Sector Gets Funded

I still keep the liquidity tightening signal, but when overall liquidity is relatively loose, you cannot simply assume every sector has money flowing into it; in 2026, you have to add the layer of capital structure. At the end of May 2026, M2 grew 8.6% year-on-year, and aggregate social financing stock grew 7.7% year-on-year; the weighted average interbank pledged repo rate in May was 1.33%, and the current 7-day reverse repo operation rate stands at 1.4%. Overall liquidity is indeed relatively loose. CCTV News citing People's Bank of China data, secondary state media; China Financial Information Network, official financial media

Yet this looseness has not covered every industry. In January–May 2026, fixed-asset investment fell 4.1%, private investment fell 7.1%, and manufacturing investment edged down 0.4%. At the same time, investment in information transmission grew 30.4%, while investment in electronic circuit manufacturing, lithium battery manufacturing, and integrated circuit manufacturing grew 50.9%, 24.9%, and 11.0%, respectively. Capital is concentrating structurally rather than flowing evenly into all sectors. NBS, official investment data and analysis

The Order in Which I Evaluate a Sector

The seven steps below represent the order of research trade-offs I set for myself, not a mechanical checklist. As long as the denominator is not clearly written down, I will not move on to valuation.

Step One: Define the Denominator

First, state clearly whether penetration refers to new unit sales, active installed base, revenue, installations, customer coverage, or domestic substitution rate. If the denominator is vague, I do not talk about valuation.

Step Two: Verify That Demand Has Already Happened

Across orders, unit sales, installations, production schedules, pricing, and customer capital expenditure, I need to see at least two items materialize. Policy targets and product demos only count as research clues.

Step Three: Locate the Profit Pool

Identify clearly who in the value chain is growing revenue, who is expanding gross margins, who is bearing the capital expenditure, and who is getting squeezed on price by customers. The bottleneck and the profit pool in a value chain may not sit in the same link.

Step Four: Check the Supply Response

Verify capacity expansion cycles, yields, qualification hurdles, second sourcing, inventory, and technology substitution. Once supply expands faster than demand, profits can fall even if demand is growing rapidly.

Step Five: Calculate Implied Expectations

Work backward from the current market cap to see how much revenue, profit margin, and market share are needed to support it. When an index has already surged, the priority is calculating how much price has to be given back if earnings miss expectations.

Step Six: Observe Capital Crowding

Track index weights, ETF assets and creations/redemptions, sell-side coverage, trading concentration, margin financing structure, and shareholder stake reductions. In the ETF era, consensus signals go far beyond the number of research reports.

Step Seven: Set Falsification Conditions

Write down in advance which indicators (such as orders, mass production, gross margins, cash flow, customer qualification, or policy implementation) must appear over the next one to four quarters, and if they do not, exit the research or downgrade its priority.

Where I Stand After Writing This

If I had to condense that 2021 report into a single sentence for today, I would write it like this:

The core of sector research has shifted from "finding a low-penetration story" to "finding structural bottlenecks where demand has already arrived, supply is hard to replicate, profits can be captured, and prices have not yet been exhausted."

As of mid-2026, here is my assessment of these five sectors:

1. Semiconductors and AI Infrastructure have the strongest fundamentals, as well as the highest overall sector valuations and crowding. Research can focus on upstream bottlenecks and earnings elasticity that have not yet been fully priced in.
2. New Energy Vehicles have moved past the thematic sector narrative into mature competition. Sales penetration is no longer the primary driver; profits, cash flow, exports, and the monetization of smart features matter more.
3. Energy Storage has the clearest installation growth and is priced lower than semiconductors, but battery cell commoditization is obvious, so research needs to be broken down into systems, controls, grid operations, and long-duration technologies.
4. Robotics is moving from storytelling to order verification. Policy and output data support the direction, but return on investment in real-world scenarios has not yet been broadly proven.
5. Innovative Drugs face an improving policy environment, while industry average data remains weak. I focus on individual drugs, clinical trials, and commercialization events, ignoring the industry average.

Source Quality and Limitations

• NBS, MIIT, NHSA, Stock Exchanges, CSI: Primary official sources with high credibility; statistical industry scopes do not completely match listed company scopes.
• CAAM: Authoritative industry association with highly credible auto production and sales data; the body of its web releases is published as images, so this research read the official images and calculated the ratios, which may differ from the CPCA's retail sales basis.
• WSTS, SIA: Authoritative global semiconductor industry organizations with relatively credible market statistics and forecasts; forecasts remain subject to cyclical shifts and changes in AI capital expenditure.
• Energy storage installation data on NEA pages: Reprinted on an official platform, but the original source is an incomplete statistical survey by CNESA DataLink, which carries limitations in coverage and project scope.
• CCTV News, China Financial Information Network: Official state media or specialized national financial media, used here solely to cite PBOC data and same-day open market operations; core policy judgments should still refer to original PBOC announcements.
• Index Factsheets: Official market snapshots as of June 30, 2026; returns and valuations change with prices and constituent adjustments and should not be extrapolated as future returns.

References

• China Merchants Securities Strategy, 2021, *The Revelation of Sector Investing: Four Selection Rules and Risk Signals* (A-Share Investment Revelations XVIII): BEAT, the 60% maturity threshold, leader concentration, the comparison between total market cap and P/E, the four-sector Kelly derivation, and the four risk signals are all the subjects reviewed from that report.
• The 2026 industry and index figures come from the pages linked above from the NBS, CAAM, WSTS, SIA, CSI index factsheets, the SSE ETF report, the NEA, MIIT, and NHSA. Liquidity figures are cited via CCTV News and China Financial Information Network.