Consumer and home robots 2026 explained: 34.2 million units sold, two Chinese makers above RMB 10bn a half-year, home humanoids still pre-scale
Consumer service robots, mostly domestic machines such as floor cleaners, sold almost 34.2 million units in 2025, up 37%, about 137 times the number of professional service robots, according to the International Federation of Robotics.
The business is now run from China: Roborock and Ecovacs each booked more than RMB 10 billion of revenue in the first half of 2026, while iRobot, the company that launched the Roomba, filed a Chapter 11 plan that hands it to its Chinese contract manufacturer.
Consumer service robots, which the International Federation of Robotics (IFR) describes as mostly domestic machines for the mass market, sold almost 34.2 million units in 2025, up 37%. That is about 137 times the 250,000 professional service robots sold in the same year and more than 1,000 times the IFR's estimate of 7,000 full-size humanoids.
The home robot is therefore already a mass-market product, but not the one most investors talk about. It is a floor cleaner or a lawn mower, it is mostly built and sold by Chinese listed companies, and the American pioneer of the category has just been handed to its Chinese supplier. This explainer sets out who leads, what their filings show and why the venture money flowing to home humanoids is betting on a different business.
34 million units, and the IFR counts only part of the industry
The IFR's World Robotics 2026 service robot figures, released on Sept 30, put consumer robot sales at almost 34.2 million units, against an implied 25 million in 2024 (34.2 million divided by 1.37, our calculation). The federation names the trends as automating household tasks, adding artificial intelligence and connecting robots to smart-home systems. It does not split out vacuum cleaners, window cleaners or mowers in the release.
The count is also a floor. The IFR's market presentation says its service robot numbers come from 238 of the 1,001 makers it knows of and are not projected to the whole industry, and it warns against comparing survey waves. The direction, though, is unambiguous, and it raises the question of who sells those machines.
Two Shanghai-listed makers above RMB 10 billion in six months
The clearest answer comes from Shanghai. Ecovacs, based in Suzhou, reported first-half 2026 revenue of RMB 10,340.9 million, up 19.2%, in its half-year report published on Aug 22. Roborock, the Beijing maker listed on the STAR Market, reported RMB 10,084.4 million, up 27.6%, on Aug 25. Together that is RMB 20.4 billion in six months, and the gap between them is 2.5%.
- Ecovacs
- Shanghai 603486
- H1 2026
- RMB 10,340.9m
- +19.2%
- RMB 1,247.9m profit (+27.4%)
- Excluding non-recurring items RMB 805.4m (-6.4%)
- Roborock
- Shanghai STAR 688169
- H1 2026
- RMB 10,084.4m
- +27.6%
- RMB 986.4m profit (+45.6%)
- R&D 7.14% of revenue
- iRobot
- Nasdaq (delisting under Chapter 11 plan)
- 9M to Sept 27, 2025
- USD 375.0m
- -26.4%
- USD 131.6m loss
- Equity to Picea under Dec 2025 plan
- OneRobotics (SwitchBot)
- Hong Kong
- IPO Dec 30, 2025
- n/a
- n/a
- n/a
- Gross IPO proceeds HK$1,640.01m
Figures as reported in each company's filing; no currency conversion. Change = reported year-on-year change. OneRobotics revenue not in MEKAWIRE's verified records.
As of Oct 17, 2026
The quality of the growth differs. Roborock's net profit attributable to shareholders rose 45.6% to RMB 986.4 million, and its profit excluding non-recurring items rose 60.7%, so the gain came from operations. Ecovacs's headline profit rose 27.4% to RMB 1,247.9 million, but profit excluding non-recurring items fell 6.4% to RMB 805.4 million, which means non-recurring items made up 35.5% of its first-half profit, and its operating cash flow fell 85.8% to RMB 206.3 million. Roborock spent 7.14% of revenue on research and development, down from 8.67%.
Both companies' net margins sit near 10% to 12% (Roborock 9.8%, Ecovacs 12.1%, our calculation from the reported figures). That is a consumer electronics margin, not a software one, and it frames what happened to the company that started the category.
iRobot ended up owned by the factory that built its robots
iRobot, which sold the first Roomba in 2002, filed for Chapter 11 in Delaware on Dec 14, 2025 under a restructuring support agreement that gives 100% of the reorganised equity to Shenzhen PICEA Robotics and its Hong Kong unit, its own contract manufacturer. Picea had bought USD 190.7 million of iRobot's loan claims from Carlyle affiliates and was owed USD 161.5 million for manufacturing, USD 352.2 million in all, according to iRobot's filings. Shareholders receive nothing.
The operating decline explains the outcome. iRobot's revenue for the nine months to Sept 27, 2025 fell 26.4% to USD 375.0 million, its nine-month net loss widened to USD 131.6 million, and it warned of substantial doubt about its ability to continue as a going concern. The Amazon takeover that might have rescued it was abandoned in January 2024.
The same direction shows elsewhere in our records. OneRobotics, the Shenzhen company behind SwitchBot home robots that sells mostly in Japan, raised HK$1,640.01 million in a Hong Kong IPO on Dec 30, 2025. LDROBOT, a Shenzhen sensor and robot mower maker, told the Hong Kong exchange on Oct 7, 2026 that it will buy the German mower maker SABO-Maschinenfabrik for no more than EUR 6 million to reach its 1,400-plus dealers. The question left is whether the next home robot will follow the same path.
Home humanoids are a different bet with a different price tag
The venture-funded home robot looks nothing like a vacuum. 1X Technologies opened orders for its NEO humanoid in October 2025 at USD 20,000 or USD 499 a month, and said in April 2026 that its Hayward, California plant can build up to 10,000 a year, aiming for more than 100,000 a year by the end of 2027. Sunday raised a USD 165 million Series B at a USD 1.15 billion valuation in March 2026 for Memo, a household robot it trains on data from people wearing its capture glove, with an invite-only beta from this autumn. LG Electronics showed its CLOiD home robot at CES 2026 without a price or date.
- 1X Technologies
- NEO
- USD 20,000 or USD 499/month
- Hayward plant up to 10,000 a year; aims for 100,000+ by end-2027
- Series B USD 100m (Jan 2024)
- Sunday
- Memo
- Invite-only home beta from autumn 2026
- Shipping at scale 2027-2028
- Series B USD 165m at USD 1.15bn (Mar 2026)
- LG Electronics
- CLOiD
- Shown at CES 2026, no price
- No date given
- Listed group
Company announcements as recorded in MEKAWIRE's verified rows. 1X's disclosed rounds in our records total USD 123.5m (Series A2 23.5 + Series B 100); later rounds reported by media are not counted.
As of Oct 17, 2026
The scale gap is the point. A full year of 1X's stated Hayward capacity equals 0.03% of the consumer robots the IFR counted in 2025, and its rental price would take 40 months to reach the purchase price. These are pre-scale products sold on what they might learn, not on a task they already do.
MEKAWIRE analysis
The home robot that already works is a Chinese consumer electronics business with 10% margins; the home humanoid is a venture bet that a general-purpose machine can escape those economics before the vacuum makers learn to build arms.
The evidence for the first half is in the filings. Two Chinese companies each sell more in six months than iRobot did in nine, they are growing at 19% to 28%, and the most famous Western brand now belongs to its supplier. A single-task floor cleaner competes on cost, navigation sensors and retail distribution, all advantages of Shenzhen and Suzhou.
The strongest counter-argument is that the categories will converge: the vacuum makers have the supply chains, the retail shelf space and the installed base in tens of millions of homes, and a robot arm on a cleaner is a smaller step than a humanoid. If that happens, the venture-backed home humanoid makers would compete with companies that already sell at consumer prices, and their valuations, such as Sunday's 7.0 times step from its Series B amount to valuation (USD 1.15 billion divided by USD 165 million), would rest on software and data alone.
- Consumer service robots
- about 34.2 million
- +37%
- Professional service robots (incl. AMRs)
- almost 250,000
- +24%
- Medical robots
- 15,300
- +19%
- of which surgical robots
- 8,400
- +57%
- Full-size humanoids (above 140 cm)
- about 7,000
- n/a
Ratio consumer to professional = 34.2m / 250,000 = 137x. Implied 2024 consumer sales = 34.2m / 1.37 = about 25.0m. IFR sample of 238 of 1,001 makers, not projected.
As of Sep 30, 2026
Bull case: home humanoids reach useful autonomy by 2027 and command prices a vacuum cannot, so 1X, Sunday and Figure build a premium category with recurring revenue. Bear case: autonomy arrives slowly, remote operators fill the gap, and the Chinese cleaner makers add manipulation at a fraction of the price.
Signals to watch: Roborock's and Ecovacs's third-quarter reports in late October 2026, which will show whether Ecovacs's recurring profit recovers; the first disclosed NEO delivery numbers from 1X, which said deliveries start in 2026; and the IFR's next consumer robot update in its World Robotics 2027 report. What would change our view is a Chinese cleaner maker announcing a mobile manipulator at a consumer price, or a home humanoid maker disclosing paid units in the thousands.
- Consumer service robots sold almost 34.2 million units worldwide in 2025, up 37%, the IFR said on Sept 30, 2026; professional service robots sold almost 250,000 (+24%) and medical robots 15,300 (+19%).
- The IFR describes consumer robots as mainly domestic robots for the mass market and does not split out robot vacuums or lawn mowers in its release; its figures come from 238 of 1,001 known makers and are not projected to the whole industry.
- Roborock reported first-half 2026 revenue of RMB 10,084.4 million, up 27.6%, and net profit attributable to shareholders of RMB 986.4 million, up 45.6%, in its Shanghai half-year report.
- Ecovacs reported first-half 2026 revenue of RMB 10,340.9 million, up 19.2%, and net profit of RMB 1,247.9 million, up 27.4%, but profit excluding non-recurring items fell 6.4% to RMB 805.4 million.
- iRobot filed for Chapter 11 on Dec 14, 2025 under a plan giving 100% of the reorganised equity to Shenzhen PICEA Robotics, its lender and main contract manufacturer, which held USD 352.2 million of claims.
- iRobot's revenue for the nine months to Sept 27, 2025 was USD 375.0 million, down 26.4% from USD 509.8 million a year earlier.
- 1X Technologies prices its NEO home humanoid at USD 20,000 or USD 499 a month and says its Hayward plant can build up to 10,000 a year; Sunday raised USD 165 million at a USD 1.15 billion valuation for its Memo household robot.
- One year of 1X's stated Hayward capacity equals 0.03% of the consumer robots the IFR counted in 2025.
Why it matters
The consumer segment is by far the largest robot market by units, and its economics are already visible in public filings, which makes it the best available test of what happens when robots become consumer products. The answer so far is consumer electronics margins and Chinese manufacturing dominance.
For investors in home humanoids, the vacuum market is both the proof that households will buy robots and the warning about what happens to price and margin once a product is commoditised.
Rival analysis
Roborock and Ecovacs are close in size, RMB 10.08 billion against RMB 10.34 billion in the first half, but Roborock is growing faster (27.6% against 19.2%) and its profit growth is operational, while 35.5% of Ecovacs's first-half profit came from non-recurring items and its operating cash flow fell 85.8%. On these filings Roborock has the momentum.
iRobot is no longer an independent rival, and its new owner Picea is a manufacturer rather than a brand. OneRobotics competes in smart-home devices with a Japan-heavy footprint. Among home humanoids, 1X has disclosed price and capacity, Sunday a valuation and a beta, and LG only a prototype.
Valuation context
Sunday's USD 1.15 billion valuation is the only disclosed home-robot venture valuation in our records, 7.0 times the size of its Series B. The listed cleaner makers are valued on earnings; we do not quote their market capitalisations here because we have not recorded them as of this date.
iRobot's equity went to zero against USD 352.2 million of supplier and lender claims, a reminder of where value sits when a hardware brand loses its cost position.
Supply-chain implications
The cleaner business runs on lidar and camera navigation, brushless motors, batteries and contract manufacturing concentrated in southern and eastern China, with Picea's factories in China and Vietnam serving iRobot. Home humanoids need actuators, hands and onboard compute, and 1X says it makes its own motors, batteries, hands and structures in-house.
That difference matters for tariffs and for scale: the cleaner supply chain is mature and price-driven, while the humanoid chain is still being built, with in-house parts that do not yet benefit from volume.
Signals to watch
1. Roborock and Ecovacs third-quarter 2026 results, due by Oct 31 under Shanghai rules, for growth and Ecovacs's recurring profit.
2. 1X's first disclosed NEO delivery volume in 2026, and whether the USD 499 monthly option dominates.
3. Sunday's beta results and any disclosed price for Memo before its 2027-2028 shipping window.
Analyst view
Thesis: the working home robot is a Chinese consumer electronics business, and home humanoids are a venture bet on escaping its economics. Confidence: medium. The filings make the first half solid; the second half depends on autonomy progress that has not yet been shown in paid deployments.
We would raise confidence if a home humanoid maker reported thousands of paid units with low remote-operator involvement, and lower it if a cleaner maker launched a capable mobile manipulator under USD 5,000.
Questions you should be asking
For Roborock and Ecovacs: do you plan robots with manipulation for the home, and at what price point? What share of first-half revenue came from robot cleaners rather than other appliances?
For 1X and Sunday: how many units have been delivered to paying households, and what share of tasks still needs a remote operator?