· Codemax · 6 min read
Kitchen robots by 2028? What AI can already do in your kitchen today
Cooking robots will draw the crowds at TT360 next month. Most of the AI a Malaysian kitchen can use today doesn't cook at all. It runs on the cameras and records you already have.
TL;DR
Kitchen robots do physical, repetitive work such as frying, wok cooking and dishwashing, and the chief executive of TT360's main sponsor expects them to be standard equipment by 2028. The AI a Malaysian kitchen can use today is mostly software: it reads existing CCTV, POS and stock records to check hygiene at the prep line, flag outlets drifting from their baseline and answer questions from your own data.
- Robots do physical work; most kitchen AI reads data and cameras you already own.
- The 2028 robot timeline is one sponsor executive's view, not an industry forecast.
- AI on existing CCTV can check hand-washing, PPE and station hygiene by shift.
- Anomaly detection compares each outlet with its own baseline and flags drift early.
- Start with one countable problem, one outlet and a named owner for alerts.
Will robots be cooking in Malaysian kitchens by 2028? Some may, and the machines being shown this year are real. But the AI that Malaysian kitchens can run today doesn’t need a robot. It runs on the CCTV, POS and stock records you already have: checking hand-washing at the prep line, flagging an outlet whose food cost has drifted, and answering questions from your own data.
The two are worth keeping apart. Robots are hardware that does physical work. Most useful kitchen AI is software that watches, compares and reports, so people can decide faster.
Where the 2028 line came from
At the media launch of Taste & Trade 360 (TT360) on Saturday 3 October, Ethan Choong Wai Hon, chief executive of Ping Edge Technology Bhd, the show’s main sponsor, said: “I believe by 2028, AI, robotics and automation will be part of basic commercial kitchen equipment, like combi ovens or ice machines today,” as reported by the New Straits Times. He also said: “No robot will replace a good chef.”
That is one informed person’s view, not an industry forecast, and it is worth reading as such. The backdrop is familiar to anyone running a kitchen: DagangNews framed the launch around a sector facing worker shortages and rising costs. The show runs from 5 to 7 November at the Kuala Lumpur Convention Centre (KLCC), with a “Battle of the Bots” cooking-robot showcase on 6 November. The kitchen robots on show will draw a crowd, and they should.
What counts as AI in the kitchen?
AI in the kitchen (AI dapur) is software that learns patterns from your data and camera footage, then flags what looks wrong or answers questions about it. It does not need to touch food. A robotic wok station, by contrast, is kitchen automation: a machine doing a physical task the same way every time. Some robots use AI to do that task. Most kitchen AI uses no robot at all.
The distinction matters when you decide where to spend. Robotics is capital equipment for one station. AI in restaurants is usually a layer over systems you already run, and its value depends on the quality of the records underneath.
Robots earn their place where the task is physical and repetitive. Frying, tossing a wok, assembling drinks, washing dishes and running plates follow the same motions every time. Machines don’t tire at 9pm, they portion the same way on every order, and they take hot, heavy work off a thin team. For a high-volume concept with a stable menu, that is a real answer to a real labour problem.
They also come with practical questions. Think about floor space, power and ventilation, cleaning between services, what happens when the menu changes, and what the kitchen does when the machine is down. None of that is a reason to dismiss them. It is a reason to know which station you are trying to fix before you buy.
What AI can a Malaysian kitchen already run?
Here is how the two line up across the jobs a kitchen actually does.
| Kitchen task | Robotics option | AI and data you can use today |
|---|---|---|
| Frying, wok cooking | Robotic fryer arms, automated wok stations | Nothing that cooks, but camera checks at the pass can flag inconsistent portions |
| Dishwashing and clearing | Dishwashing robots, conveyor systems | Very little; this is physical work |
| Running food and drinks | Service and delivery robots | Footfall and queue-length counts from existing CCTV, so you staff the floor to real demand |
| Hygiene at the prep line | Few practical options | Hand-washing, PPE and station hygiene checks on existing CCTV, reported by shift |
| Smoke and fire hazards | Suppression systems | Smoke and fire detection on CCTV, alongside your fire alarms, not instead of them |
| Ordering and stock | Automated storage, mostly in large warehouses | Recommended order quantities from par levels and POS data; alerts when usage drifts from recipe |
| Spotting problems across outlets | None | Each outlet, product and shift compared with its own baseline, with alerts sent to managers |
| ”Why is food cost up at Outlet 7?” | None | An AI agent that answers from your own records and shows the rows behind the answer |
The pattern is consistent. Robots replace effort at one station. AI replaces the watching: the checks and comparisons no manager has time to do across every outlet and every shift. The line between what AI should watch and what people should decide still holds. The software flags; a person acts.
Why does kitchen AI depend on your records?
An alert about food cost is only as good as the recipe it is measured against. If yields are wrong, the variance is wrong. If transfers from the central kitchen aren’t recorded, an outlet looks short when it isn’t. AI reading messy data produces confident noise, and teams quickly learn to ignore it.
Cameras are the easier start. Vision AI on the cameras you already own needs sensible camera angles and calibration to each site, not a clean database. Sales, stock and recipe data take more groundwork, but they are where food cost and waste are won or lost.
What should you try first?
- Pick one problem you can count. Food cost variance on your ten best-selling items, hand-washing at the prep line, or stockouts on best-sellers. One, not five.
- Check what you already capture. Are the cameras IP cameras streaming RTSP or ONVIF? Does POS data arrive daily? Are recipes and yields current?
- Run it at one outlet with a named owner. Every alert type needs one person responsible for acting on it, or it becomes background noise.
- Measure whether people acted. An alert nobody reads is a cost, not a capability.
- Then look at robots, station by station. Once you know where the hours and waste actually go, you will know which station a machine would help most.
Where the AI-Kitchen Command Center fits
The Codemax AI-Kitchen Command Center is the software column of that table. It watches every outlet, product and shift against its own baseline and flags drift before it reaches the P&L. It puts sales, stock, food safety, footfall and social signals on live dashboards, and lets managers ask questions in plain language, with answers grounded in their own records and the underlying rows traceable. When the data doesn’t support an answer, it says so. Alerts and narrative reports go out by email, WhatsApp, Slack or Microsoft Teams. It is not a robot or kitchen hardware, and it doesn’t cook.
Codemax VisionAI adds AI to the CCTV you already have, on any IP camera of 2MP or more streaming RTSP or ONVIF, for SOP and PPE compliance, hygiene checks, queue and footfall counts, and fire and smoke hazards. It is anonymous by default, with no face recognition. Codemax RMS (Resource Management System) is the data spine underneath: recipes, stock, production and transfers on one record, so the AI is reading what actually happened. If you want to see that on your own numbers, talk to the team.
Questions worth asking this week
- Which kitchen task costs us the most hours, and is it physical work or checking work?
- Which three numbers would we want to hear about the same day, not at month-end?
- Do our cameras stream over a standard protocol, and who looks at the footage today?
- If an alert arrived tomorrow morning, who would own it?
If most answers point to checking work, start with software that runs on what you already own. If they point to one hot, repetitive station, that is where a robot belongs on your shortlist.
Frequently asked questions
Will robots replace kitchen staff in Malaysia?
Kitchen robots on show in Malaysia are designed for specific physical tasks such as frying, wok cooking, drinks and dishwashing, not to run a whole kitchen. Even the chief executive of TT360 2026's main sponsor, who expects robotics to become standard kitchen equipment by 2028, said no robot will replace a good chef. Where operators use robots, the point is to take repetitive, hot or heavy work off a short-staffed team.
What is the difference between kitchen automation and AI?
Kitchen automation is equipment or software that carries out a task without a person doing it, such as a robotic fryer or an automatic reorder. AI is software that learns patterns from data and images, so it can spot a missed hand-wash on camera or a stock variance that is out of character for one outlet. Some robots use AI, but most kitchen AI uses no robot.
Can AI use the CCTV cameras a restaurant already has?
Video AI for restaurants can usually work with the cameras already installed, as long as they are modern IP cameras streaming over a standard protocol such as RTSP or ONVIF. The models should be calibrated to each site's lighting, layout and policies before go-live, and privacy settings such as anonymous detection and data retention should be agreed and explained to staff.
What AI should a restaurant group in Malaysia try first?
A restaurant group should start with one problem it can count and that already costs money: food cost variance on its best-selling items, hygiene checks at the prep line, or stockouts. Pilot it at one outlet or region with a named owner for the alerts, and measure whether people acted on them. Expand only when the first use case is being used every week.
Do I need clean data before using AI in the kitchen?
AI in the kitchen is only as reliable as the records under it. Recipes with correct yields, sales flowing from the POS and stock movements recorded when they happen are the minimum for food cost and variance alerts. Camera-based checks need less data preparation, which is why many operators start there while they tidy their recipe and stock records.
Put the kitchen you already have to work.
See how the Codemax AI-Kitchen Command Center and VisionAI turn your existing data and CCTV into alerts your managers can act on.
Sources
- Kitchen robots set to reshape F&B operations — New Straits Times, 4 Oct 2026
- TT360 2026 ketengahkan robot dapur ketika sektor F&B Malaysia berdepan kekurangan pekerja, kenaikan kos — DagangNews, 4 Oct 2026
Related reading
- Going to TT360? Questions to ask any kitchen tech vendor Three days, 260 booths and a lot of demos. How to walk into TT360 2026 with your problems defined, ask every vendor the same questions, and leave with a shortlist you can defend.
- What a kitchen command center actually shows you Less a wall of charts than a short list of exceptions and a way to ask why. An illustrative Tuesday morning on the command deck, panel by panel and alert by alert.
- How CCTV AI checks kitchen hygiene, shift by shift What AI on your existing kitchen cameras watches from opening to overnight, how a detection becomes a closed task, and what it doesn't do.