Main主页 / roboindex / rbi-system-ce-i
RBI-system-ce-IRBI-system-ce-I
The fully invested capital-efficiency portfolio, with the Thesis Gap redistributed to eligible names.
- RESEARCH CUT 2026-08-27研究截点 2026-08-27
ANALYST-REVIEWED INDEX · 17 HOLDINGS分析师审阅指数 · 17 项持仓
RBI-system-ce-I
Thesis Gap redistributed pro rata to names with headroom. Measures what an investor could actually own today.
CAPITAL POSITION资本配置
100.00%equities股票0.00% cash · 100.00% accounted for0.00% 现金 · 100.00% 已计入Annualized return年化收益+14.74%-3.93% vs SPY-3.93% 对比 SPY
Annualized volatility年化波动率32.95%SPY 12.70%
Maximum drawdown最大回撤-23.00%SPY -9.13%
Sharpe ratio夏普比率0.45SPY 1.47
Correlation to SPY与 SPY 的相关性0.67daily returns日收益率
PORTFOLIO HEATMAP投资组合热力图
Where the weight sits.权重分布在哪里。
Every position, largest first. Darker tiles carry more portfolio weight.按权重从高到低展示全部持仓。颜色越深，投资组合权重越高。
TER 10.00% ISRG 10.00% EH 10.00% AMPX 10.00% 9880 9.45% 300024 8.42% CGNX 6.45% AMBA 5.87% 002747 5.00% 9660 5.00% HSAI / 2525 4.91% 2498 4.91% ROK 2.00% TSLA 2.00% 002415 2.00% NVDA 2.00% 300750 / 3750 2.00%
Lower weight较低权重Higher weight较高权重
LAST 13 MONTHS过去 13 个月
Growth of $100100 美元的增长
RBI-system-ce-ISPY
2025-07-28 to至 2026-08-27. SPY observations are aligned by date in the chart.图表中的 SPY 观测值按日期对齐。 Published weights held as a fixed quarterly-rebalance target through the same backtest engine as the variants below -- not a live track record.
PORTFOLIO EXPOSURE投资组合敞口
Where the portfolio is concentrated.投资组合集中在哪里。
Exposure is calculated across the complete portfolio shown above.敞口根据以上展示的完整投资组合计算。
Geography地域
CN51.69%
US48.32%
Robotics purity机器人业务纯度
High42.87%
Medium-low / robotics segment within test equipment group10.00%
Medium / battery pure play, robotics indirect10.00%
High / LiDAR pure play9.82%
High / industrial machine vision6.45%
Medium / edge AI across endpoints5.87%
Medium-low / core business autonomous driving5.00%
Low / diversified industrial automation2.00%
Very low / current robotics option2.00%
Low / subsidiary exposure opaque2.00%
Low / highly diversified2.00%
Very low / broad battery proxy2.00%
RELATIVE PERFORMANCE相对业绩
Against every construction and benchmark.与所有构建方法和基准对比。
| Portfolio投资组合 | Ann. return年化收益 | Ann. vol年化波动 | Max DD最大回撤 | Sharpe夏普比率 | Corr / SPY与 SPY 相关性 |
|---|---|---|---|---|---|
| RBI-system-ce-T | +10.7% | 23.70% | -16.94% | 0.45 | 0.68 |
| RBI-system-ce-I | +14.7% | 32.95% | -23.00% | 0.45 | 0.67 |
| RBI-MCW | +34.4% | 23.83% | -17.97% | 1.44 | 0.59 |
| RBI-HP | +3.0% | 31.93% | -21.09% | 0.09 | 0.46 |
| RBI-US | +46.2% | 37.85% | -22.43% | 1.22 | — |
| RBI-JP | +40.7% | 43.70% | -24.39% | 0.93 | — |
| RBI-CN | -8.2% | 30.16% | -29.54% | -0.27 | — |
| SMH | +81.9% | 38.06% | -24.62% | 2.15 | 0.77 |
| SPY | +18.7% | 12.70% | -9.13% | 1.47 | — |
CAPITAL POSITIONS资本持仓
17 holdings · 100.00% invested.17 项持仓 · 已投资 100.00%。
| Ticker代码 | Company公司 | Country国家 | Portfolio weight投资组合权重 | Robotics purity机器人业务纯度 | Research stance研究立场 |
|---|---|---|---|---|---|
| TER | Teradyne | US | 10.00% | Medium-low / robotics segment within test equipment group | avoid回避 |
| The segment loss is far smaller than the FY2025 framing implies, and it still does not make P17 a moat. The 10-Q for the period ended 2026-06-28 reports robotics segment income before income taxes of −$2.6m in Q2 2026 and −$3.5m in H1 2026, against −$18.5m and −$55.7m in the year-ago periods: within $2.6m of quarterly breakeven on about $400m of annualised revenue. But segment gross margin fell from 53.8% to 52.0% in H1 while revenue grew 32.9%, so 58% of the $52.2m improvement in loss before taxes is cost and charge reduction and 42% is volume at a lower margin — none of it is the margin expansion an attach-driven annuity would produce. Two disclosed robotics restructuring actions in 2025, about 10% of global robotics staff in January (roughly 140 people) and a further 14% of the remainder in November, are cumulatively about 24%, roughly 330 people on a derived base of about 1,400; the widely cited '~400 layoffs' is not reconciled to those two actions. The $416m of goodwill remains unverified — no impairment charge was found in FY2025 or H1 2026, but absence of search results is not absence of an impairment, so read the goodwill note directly. The falsification test is cheap and specific: if gross margin expands above roughly 55% over the next two quarters while revenue grows, the annuity reading gains real support and this verdict should be revisited. Return −7.76%. | |||||
| ISRG | Intuitive Surgical | US | 10.00% | High | pass通过 |
| Return +12.54%, confidence 85. PMF is mature; the risks are valuation, regulation and competition. Margin of safety is better below $400. | |||||
| EH | EHang亿航智能 | CN | 10.00% | High | avoid回避 |
| Only four units delivered in Q1 2026, with revenue of RMB 25.7m and an operating loss of RMB 127.9m. Commercialisation and funding needs have not cleared the hard gates. | |||||
| AMPX | Amprius Technologies | US | 10.00% | Medium / battery pure play, robotics indirect | avoid回避 |
| Q2 2026 revenue $34m (+126%) but a $5.1m net loss; customer concentration plus Asian contract manufacturing. Verify cycle life, yield and cost. | |||||
| 9880 | UBTECH优必选 | CN | 9.45% | High | avoid回避 |
| FY2025 revenue was RMB 2.001bn (+53.3%) with humanoid revenue of RMB 821m (+2,203.7%) on 1,079 full-size humanoid units, a net loss of RMB 790m (narrowed 32%), gross margin of 37.7% (+9pts), an adjusted EBITDA loss of RMB 438m and period-end cash of RMB 4.888bn; Walker S2 entered mass delivery in November 2025. Two flags. The claim that Walker S2 monthly capacity exceeds 1,000 units is design capacity, not output. And the same source reports both 1,079 units sold and full-year deliveries exceeding 500 units — sold, delivered and recognised are three different numbers, and shipments may include channel or internal units — so neither unit figure should be relied on without the annual or interim report. The H1 2026 interim was not published as of 2026-08-14 and is due at the end of August. Return −18.91%; monitor dilution. | |||||
| 300024 | SIASUN新松机器人 | CN | 8.42% | High | avoid回避 |
| FY2025 revenue RMB 4.12bn, net loss RMB 398m, gross margin 12.5%, and the Q1 2026 loss widened. Return −23.51%; watch as a turnaround only. | |||||
| CGNX | Cognex | US | 6.45% | High / industrial machine vision | watch观察 |
| The conclusion — wait for a better price — may stand, but the stated reason no longer does. Q2 2026, reported 2026-08-12, was a record quarter: revenue $291m (+17%), adjusted EPS $0.45 (+80%), adjusted EBITDA margin 32.2% (up 1,150bp year over year), an eighth consecutive quarter of both revenue growth and margin expansion, OneVision at general availability with hundreds of customers, and data-centre revenue at a low-single-digit percentage of sales but growing more than 30%. This is not a company sitting at the bottom of its cycle, so any decision to wait must now rest on valuation alone rather than on cycle position. Return +2.28%; revisit below $55. | |||||
| AMBA | Ambarella | US | 5.87% | Medium / edge AI across endpoints | watch观察 |
| Still GAAP loss-making, and robotics revenue is still not separately disclosed. Q1 FY2027, the quarter ended 2026-04-30: revenue $100.4m (+16.9%), GAAP net loss $18.1m ($0.41 a share) against $24.3m a year earlier, non-GAAP net income $5.0m ($0.11), and GAAP gross margin 58.4% against 60.0%. FY2026, ended 2026-01-31, carried a GAAP net loss of $75.9m against $117.1m in FY2025, with non-GAAP net profit of $26.9m. The GAAP-to-non-GAAP gap is $23.1m, roughly 23% of quarterly revenue. The loss is narrowing but has not crossed over, and gross margin fell 160bp year over year while revenue grew 17% — margin, not the loss, is the number that tests the P04 pricing-power thesis. Return +4.04%. | |||||
| 002747 | ESTUN埃斯顿 | CN | 5.00% | High | avoid回避 |
| Recast before quoting any profit figure. The H1 2026 preannouncement of 2026-07-15 puts attributable net profit at RMB 150–180m, up 2,144.74% to 2,593.68% year over year, but recurring net profit at only RMB 60–75m. In Q1 2026 attributable net profit was RMB 97.84m (+674.64%) with non-recurring items of RMB 86.54m and recurring net profit of RMB 19.36m, reported as more than 88% of Q1 net profit coming from non-sustainable items; note that 19.36 plus 86.54 does not equal the stated 97.84, most likely a gross-versus-net-of-tax presentation — resolve it from the quarterly report rather than interpolating. Q1 2026 revenue was RMB 1.217bn, down 2.22% year over year: revenue is declining. FY2025 revenue was RMB 4.888bn (+21.93%) with attributable net profit of RMB 44.97m, recurring net profit of RMB 7.28m and a net margin of 0.93%. The company is pursuing a Hong Kong listing. Goodwill original cost of RMB 1.34bn is still unverified. Return −22.09%.Recast before quoting any profit figure. The H1 2026 preannouncement of 2026-07-15 puts attributable net profit at RMB 150–180m, up 2,144.74% to 2,593.68% year over year, but recurring net profit (扣非) at only RMB 60–75m. In Q1 2026 attributable net profit was RMB 97.84m (+674.64%) with non-recurring items of RMB 86.54m and recurring net profit of RMB 19.36m, reported as more than 88% of Q1 net profit coming from non-sustainable items; note that 19.36 plus 86.54 does not equal the stated 97.84, most likely a gross-versus-net-of-tax presentation — resolve it from the quarterly report rather than interpolating. Q1 2026 revenue was RMB 1.217bn, down 2.22% year over year: revenue is declining. FY2025 revenue was RMB 4.888bn (+21.93%) with attributable net profit of RMB 44.97m, recurring net profit of RMB 7.28m and a net margin of 0.93%. The company is pursuing a Hong Kong listing. Goodwill original cost of RMB 1.34bn is still unverified. Return −22.09%. | |||||
| 9660 | Horizon Robotics地平线 | CN | 5.00% | Medium-low / core business autonomous driving | avoid回避 |
| The H1 2026 headline is non-operating, but the driver is the CARIAD convertible loan, not the D-Robotics deconsolidation. The company's profit alert puts H1 2026 revenue at RMB 1.93–2.08bn (+24.8% to +34.5%) and net profit at RMB 3.5–4.0bn, which it states is primarily a non-cash fair-value gain on convertible loans previously issued to CARIAD, Volkswagen Group's software arm, driven by share-price movement and with no material effect on core operating performance or cash flow; adjusted (non-IFRS) net loss is expected to widen to as much as RMB 1.7bn from RMB 1.33bn in H1 2025, and Horizon has moved to redeem the CARIAD convertible loan early to curb dilution. That is a swing of roughly RMB 5.4bn between reported net profit and adjusted net loss on about RMB 2.0bn of revenue. The dilution of D-Robotics to roughly 40% and its deconsolidation in 2026 are real and separate; they are not what drives the H1 headline. Return −9.74%. | |||||
| HSAI / 2525 | Hesai禾赛 | CN | 4.91% | High / LiDAR pure play | watch观察 |
| Best risk-adjusted return on the board at +19.35%, but the earnings behind it are not repeatable at that level. FY2025 revenue was RMB 3,027.6m (+45.8%) and net income RMB 435.9m; the largest single non-operating item is Q3 2025 other income of RMB 172.9m, which the company states came mainly from a gain on disposal of an equity investment in an early-stage technology company — 39.7% of full-year net income from one line. Q1 2026 revenue was RMB 681m (+30%) with GAAP net income of only RMB 18m (non-GAAP RMB 48m), a fourth consecutive GAAP-profitable quarter; net margin fell from 14.4% in FY2025 to 2.6%. The recast this position required has in effect been performed by the company's own Q1 print: recurring earnings power looks like 2–3% of sales, not 14%. Q2 2026 results are scheduled for 2026-08-18 — do not re-size the position before that print. | |||||
| 2498 | RoboSense速腾聚创 | CN | 4.91% | High / LiDAR pure play | watch观察 |
| Top five customers were 83.4% of revenue in FY2024 — still the latest verified concentration disclosure, and the date must stay attached to it; no FY2025 or H1 2026 five-largest-customer disclosure was located. FY2025 revenue was RMB 1,941m (+17.7%) with the net loss narrowed to RMB 145.0m from RMB 481.8m, gross margin 26.5% against 17.2%, a first-ever quarterly net profit of RMB 103.7m in Q4 2025, and robotics and other non-ADAS revenue up 257.7% to RMB 709.8m. H1 2026 lidar sales were 719,200 units (+169.6%), of which robotics 282,600 (+510.4%) and ADAS 436,600 (+98.0%); Q1 2026 was the first quarter in which robotics units exceeded ADAS units, and the company says it serves more than 3,400 robotics clients. Concentration has very likely fallen materially, but an inference is not a disclosure — settle it from the Major Customers note in the FY2025 annual report before moving the 60/100 confidence score. Return +10.8%. | |||||
| ROK | Rockwell Automation | US | 2.00% | Low / diversified industrial automation | avoid回避 |
| Good quality but richly valued; robotics is not separately disclosed and the exposure is essentially industrial-capex beta. Return −11.26%. | |||||
| TSLA | Tesla | US | 2.00% | Very low / current robotics option | avoid回避 |
| Q2 2026 disclosed the first Optimus line but zero robotics revenue; group operating margin 1.4% and free cash flow −$1.09bn. Treat it as an unvalidated option, not as current earnings. | |||||
| 002415 | Hikvision / Hikrobot海康机器人 | CN | 2.00% | Low / subsidiary exposure opaque | avoid回避 |
| Hikrobot is no longer undisclosed. Hikvision's H1 2026 revenue was RMB 46.823bn (+11.97%) with net profit attributable of RMB 7.996bn (+39.57%), an interim dividend of RMB 5.041bn, and a separately reported robot-business revenue line of RMB 4.028bn (+28.34%). On 2026-03-31 the company restarted preparations to spin off Hangzhou Hikrobot onto the Shenzhen ChiNext board; the application has been with the exchange since March 2023, through two rounds of enquiry and four suspensions for stale financial data. Hikrobot's standalone revenue was RMB 3.90bn in FY2022 (net income RMB 640m), RMB 4.94bn in FY2023 (RMB 800m), RMB 5.929bn in FY2024 and RMB 6.452bn in FY2025 (+8.82%), with a planned raise of about RMB 6bn; the regulator's stated concerns are related-party transaction fairness, business independence and the reliability of financial data, and Hikrobot has relied on the parent for contract manufacturing. Do not merge the parent's robot-business segment series with the Hikrobot legal-entity series until scope is reconciled — +28.34% and +8.82% are not the same base. Sanctions and export-control exposure and the governance discount are unchanged. Return −1.19%. | |||||
| NVDA | NVIDIA | US | 2.00% | Low / highly diversified | pass通过 |
| Return +18.04%, confidence 85. Robotics is still not separately disclosed and remains immaterial against data centre: hold it as free optionality, never as a robotics pure play. The reporting structure now invites exactly the error that rule forbids — beginning Q1 FY2027 NVIDIA combined the former Gaming, Professional Visualization and Automotive segments into a single Edge Computing segment, which also captures revenue from devices for agentic and physical AI, AI-RAN base stations and robotics. Edge Computing growth is mostly PCs, consoles and telecom base stations; it is not a robotics number and must never be quoted as one. Jetson AGX Thor delivers 2,070 FP4 TFLOPS with 128GB at 40–130W, and the company states 7.5x the AI performance and 3.5x the efficiency of AGX Orin, but the volume T3000 and T2000 modules announced in July 2026 do not ship until Q1 2027 — any 2026 P04 revenue thesis is about design wins, not shipments. | |||||
| 300750 / 3750 | CATL宁德时代 | CN | 2.00% | Very low / broad battery proxy | watch观察 |
| FY2025 revenue RMB 423.7bn, net profit RMB 72.2bn and operating cash flow RMB 133.2bn, driven by EV and storage. The robotics business must not be valued, and the reason is now stronger rather than weaker. In June 2026 CATL and Galbot signed a global strategic cooperation agreement, and CATL states that the Galbot S1 — which it describes as the first humanoid in regular operation powered by CATL batteries, with a 50 kg dual-arm payload and a claimed up to eight hours of continuous operation — is deployed inside a CATL cell plant; earlier, CATL commissioned a humanoid-enabled battery-pack line in Luoyang using 'Xiaomo', built by Spirit AI. CATL is therefore simultaneously the cell supplier, the deployment site and the commercialisation partner, and the duty-cycle claim is a supplier's claim about its own cell in its own building. No robotics revenue, order value, unit count or segment disclosure exists. The H1 2026 interim is due by 2026-08-31. | |||||
| Total合计 | 100.01% | ||||
- Ticker代码
- The exchange symbol used to identify the listed security.用于识别上市证券的交易所代码。
- Country国家
- The issuer's primary-market or headquarters classification in this snapshot.本快照中发行人的主要市场或总部分类。
- Portfolio weight投资组合权重
- The percentage of total capital assigned to the position; cash is included in the 100% total.分配给该持仓的总资本百分比；100% 合计中包括现金。
- Robotics purity机器人业务纯度
- A qualitative label for how directly the issuer's business is exposed to robotics.用于描述发行人业务与机器人领域直接相关程度的定性标签。
- Research stance研究立场
- The current analyst disposition, not a personalized investment recommendation.当前分析师判断，不构成个性化投资建议。
INDIVIDUAL METHODOLOGY单项方法论How RBI-system-ce-I is constructedRBI-system-ce-I 的构建方式Expand +展开 +
01
Hypothesis假设
The bottleneck thesis can be translated into a fully invested public-equity portfolio without hiding which exposures were originally unavailable.
02
Universe样本范围
The same analyst-reviewed eligible universe as RBI-system-ce-T.
03
Weighting权重方法
Starts from the CE-N system targets, then redistributes the Thesis Gap pro rata to eligible names with remaining capacity.
04
Caps权重上限
10% per company, 25% per system, 2% for low robotics purity, and 5% for low earnings quality.
05
Cash现金
Fully invested; target cash is 0% after redistribution.
06
Rebalance再平衡
Quarterly, following the documented research cut.
07
Benchmark基准
SPY for broad-market comparison and RBI-system-ce-T for the cost of investability.