Auf der CES 2026 enthüllt Hyundai Motor Group, wie KI-Roboter schon bald im Alltag und in der Industrie mit Menschen zusammenarbeiten sollen. Die präsentierte Strategie zeigt eine realitätsnahe, skalierbare Zukunft für humanoide Roboter, Smart Factories und adaptive Systeme.
😇 Dein Abo hilft uns: https://tublo.eu/abonnieren
✅ Source: Hyundai & Boston Dynamics
➡️ Mehr Infos: https://www.tuningblog.eu/dies-u-das/boston-atlas-764609/
Hyundai bringt künstliche Intelligenz und Robotik auf ein neues Level: Unter dem Leitmotiv Partnering Human Progress präsentierte die Hyundai Motor Group auf der CES 2026 eine konkrete Vision für den Einsatz von humanoiden Robotern in Produktion, Logistik und Alltag. Im Mittelpunkt steht eine menschenzentrierte Robotik, die durch echte Interaktion lernt, sich anpasst und produktiv mitarbeitet. Die gezeigten Systeme sind nicht nur Prototypen, sondern Teil eines skalierbaren Plans – mit Software-gesteuerten Fabriken und Robotern, die Aufgaben in Echtzeit übernehmen können. Der Mensch bleibt im Mittelpunkt, die Technologie wird zur Unterstützung.
#Hyundai #CES2026 #AIRobotics #SmartFactory #Robotik #KünstlicheIntelligenz #Industrie40 #Atlas #tuningblog - das Magazin für Auto-Tuning und Mobilität!
😇 Dein Abo hilft uns: https://tublo.eu/abonnieren
✅ Source: Hyundai & Boston Dynamics
➡️ Mehr Infos: https://www.tuningblog.eu/dies-u-das/boston-atlas-764609/
Hyundai bringt künstliche Intelligenz und Robotik auf ein neues Level: Unter dem Leitmotiv Partnering Human Progress präsentierte die Hyundai Motor Group auf der CES 2026 eine konkrete Vision für den Einsatz von humanoiden Robotern in Produktion, Logistik und Alltag. Im Mittelpunkt steht eine menschenzentrierte Robotik, die durch echte Interaktion lernt, sich anpasst und produktiv mitarbeitet. Die gezeigten Systeme sind nicht nur Prototypen, sondern Teil eines skalierbaren Plans – mit Software-gesteuerten Fabriken und Robotern, die Aufgaben in Echtzeit übernehmen können. Der Mensch bleibt im Mittelpunkt, die Technologie wird zur Unterstützung.
#Hyundai #CES2026 #AIRobotics #SmartFactory #Robotik #KünstlicheIntelligenz #Industrie40 #Atlas #tuningblog - das Magazin für Auto-Tuning und Mobilität!
Kategorie
🚗
MotorTranskript
00:00Now, our robots have a lot of talent, dancing like K-pop stars,
00:04but they are built for a higher purpose.
00:07To cooperate, to assist, to co-work, to do hazardous jobs,
00:12and many other things, for us and with us.
00:16And all of this does not happen in isolation.
00:20It's all about partnership.
00:23It takes so many talented people,
00:25my colleagues at Boston Dynamics and Hyundai Motor Group,
00:28and many other valued partners,
00:31each bringing something essential to achieve
00:34what we call human-centered AI robotics.
00:39Robots equipped with physical AI.
00:43Robots that perceive the world like we do,
00:46and work responsibly for us and with us.
00:50This is technology that is built for humanity.
00:55Technology that supports us, extends our capabilities,
00:58makes our work safer, and in some cases, more meaningful.
01:02And now, let's take a look at what that world looks like,
01:07and what it means for all of us.
01:09Every step forward opens a new path.
01:26It becomes a pulse toward worlds yet unseen.
01:43Today, robots learn from humanity.
01:47Robots learn from robots.
01:49We meet as partners to further evolve.
02:01With AI robotics at our side,
02:03our potential expands.
02:12Guiding what matters most.
02:14Giving us courage to push beyond limits.
02:31And the collective strength to share the burden.
02:34The road ahead brings us a liberated, everyday life.
02:51Safer from risks,
02:53and freed from repetition.
02:55Together, we move forward
02:57to achieve the next innovation.
02:59Every moment creates a trail of data
03:09where humanity and AI robotics
03:12open a greater cycle of progress
03:15and awaken the dawn of infinite possibility.
03:18Partnering human progress.
03:30Partnering human progress.
03:48Good afternoon.
03:49I'm Zach Joukowsky,
03:52and I'm the VP and General Manager
03:54for Humanoid Robots at Boston Dynamics.
03:56And I'm Aya Durbin.
03:58I lead our Humanoid Application Product Strategy
04:00at Boston Dynamics.
04:02Wow.
04:02It's incredible to finally show the robots
04:04we've worked so hard on
04:06at such a big global stage.
04:08How do you feel, Aya?
04:09I feel the same.
04:10I'm excited and honored to be here.
04:13For the last 30 years,
04:15Boston Dynamics has pushed
04:16the boundaries of the possible.
04:17hundreds of millions of people
04:20from around the world
04:21have watched our journey online,
04:23captivated by the future coming to life
04:25before their eyes.
04:27And while the robots are impressive,
04:29the truth is that many of us
04:30join the company for a simple reason.
04:32To change the world for the better.
04:35We believe in the future
04:36that we imagined as kits.
04:38Where robots are partners
04:39in our everyday lives.
04:41Where they're not just impressive
04:43feats of engineering,
04:44but built to make our lives better
04:46and push the world forward.
04:48And for the first time in our history,
04:50we see a very clear path
04:52to that future coming to life.
04:54I've been designing,
04:55building, and programming robots
04:57for my entire career.
04:58And I've never been more excited
05:00about what's next
05:01than I am right now.
05:03We're on the cusp
05:04of a transformational shift
05:06that will be as impactful
05:07as the smartphone.
05:09Earlier, you saw a pretty cool dance performance
05:11from our Spot robots.
05:13Spot was our first product,
05:15but the spots deployed with customers
05:17don't have time to dance.
05:19Those Spot robots collect data,
05:22reduce downtime in industrial facilities,
05:24and keep people out of harm's way.
05:27Over the past five years,
05:29we've deployed thousands of spots
05:31to hundreds of customer sites
05:32in over 40 different countries.
05:35Spot is a serious commercial success.
05:37We launched Stretch in 2023
05:40to bring total flexibility
05:42to warehouse automation,
05:44starting with truck
05:44and container unloading.
05:46This is a back-breaking task
05:48done in sweltering heat
05:49and chilling cold.
05:51Stretch has now unloaded
05:52more than 20 million boxes
05:54for our customers.
05:56These are real examples
05:57of how the robots that we build
05:59make workplaces safer
06:00and people more productive.
06:03So what's next?
06:04Our vision is a humanoid
06:07that is general purpose,
06:09that can go anywhere,
06:11understand its surroundings,
06:12and manipulate any object.
06:14We're starting
06:15in industrial environments first.
06:18We're building a tool
06:19that can drop into your facility
06:20as it exists today.
06:22It doesn't just do one job,
06:24but is flexible enough
06:25to evolve as your business changes.
06:27Over the last few years,
06:29we have visited hundreds
06:30of customer sites,
06:31walking the floors
06:32with the people
06:33who understand their operations best.
06:35And we've learned
06:36that to bring Atlas
06:37from a prototype to a product,
06:39it needs to be safe,
06:42reliable, and predictable.
06:45Atlas needs to be doing the jobs
06:47that are repetitive,
06:48exhausting,
06:49and increasingly difficult to staff.
06:52And as we did with Spot and Stretch,
06:54we are partnering
06:55with the best in the business,
06:56and we're learning directly
06:57from the people
06:58who will use our products most.
07:00But to realize this vision,
07:03we needed a robot
07:04that was built
07:04for the rigors of industry.
07:06We've been working
07:07on humanoids
07:08for more than a decade
07:09at Boston Dynamics,
07:10always keeping a close watch
07:12on when the missing pieces
07:13of technology
07:14would fall into place
07:15to make it truly
07:16commercially viable.
07:18The rapid advancements
07:19in AI
07:20over the past few years
07:21are the piece
07:23that we needed,
07:24and that moment
07:24is finally here.
07:26Now it's time
07:26to officially take Atlas
07:28out of the lab.
07:29So for the first time
07:30ever in public,
07:31ladies and gentlemen,
07:32please welcome Atlas
07:33to the stage.
07:47So as you can see,
07:48Atlas really does stand up
07:50the way that we show
07:51in our videos online,
07:52and we get tons
07:53of questions about this.
07:55Zach, while we have
07:56everyone here,
07:56do you want to provide
07:57a little bit of insight
07:58into how we made
07:58that decision?
07:59I mean, it's pretty simple.
08:01Atlas doesn't have to move
08:02like a person does,
08:03and we're looking
08:04for the most stable,
08:05efficient way to get up,
08:06and it turns out
08:07this is it.
08:08The other question
08:09we get all the time
08:09is how Atlas is moving.
08:14How is Atlas moving today?
08:15So in normal operation,
08:17Atlas is fully autonomous
08:19on a job site,
08:20but today, Sarvesh,
08:21one of our FAEs,
08:22is piloting Atlas
08:23on the stage,
08:24giving it really simple
08:25instructions like forwards
08:26and backwards
08:27just to keep things simple.
08:30This is Atlas'
08:31first time in public,
08:33and as you can see,
08:34its walk is pretty natural
08:36and pretty smooth.
08:37We've partnered
08:37with our sister organization,
08:39the Robotics and AI Institute,
08:41to push the limits
08:42on natural walking
08:43just for this performance,
08:45and I think we've done
08:46something really special.
08:48We've also been
08:49at this robotics thing
08:50for a while,
08:51and we've learned
08:52that there's more to it
08:53than just copying nature.
08:55We can pick the best parts
08:57of what nature has to offer
08:58and do better in others.
09:00So as you can see,
09:01Atlas here has joints
09:02that can move 360 degrees.
09:05This lets Atlas move
09:06even more efficiently
09:08than humans,
09:09particularly in manufacturing
09:10environments
09:11where every second counts.
09:13We've also designed
09:15Atlas' head and face
09:16very purposefully.
09:18We want folks working
09:19with Atlas
09:20to know that Atlas
09:21is a helpful robot
09:22and not a person.
09:24Seeing Atlas live in action
09:26for the first time
09:27is wildly exciting,
09:29but seeing it perform
09:31real industrial work
09:32is even more compelling.
09:34We recently took this robot
09:35down to the Hyundai Motor Group
09:37Metaplant America
09:38to have Atlas test
09:40in a real manufacturing facility
09:41where we performed
09:43autonomous material handling tasks.
09:46We proved that humanoids
09:47could be viable
09:48in a real-world application.
09:51And we collected data
09:52and learnings
09:53that we took back to Boston
09:54to make our product even better.
09:56If you would like to see
09:58what one of those applications
09:59looks like,
09:59we will be running
10:00live demonstrations
10:01all week long
10:02in the Hyundai Motor Group booth
10:03at CES.
10:05And that's not all.
10:07We have been hard at work
10:09on the hardware itself.
10:10We are building out
10:11the functionality
10:12that our customers need
10:13in a product deployed at scale.
10:16Zach, do you want to talk about it?
10:18Of course I do.
10:20The version of Atlas
10:21that you're seeing here
10:22is just a research prototype
10:24that we've been using
10:25for a couple years
10:26to understand the system.
10:28But we've been hard at work
10:29on making the actual
10:30product version of Atlas.
10:32I can confidently say
10:34that this product version
10:35of Atlas is the best
10:36and actually simplest robot
10:38that we have ever built.
10:40But rather than just talk about it,
10:42why don't we actually show you
10:43what it's going to look like?
10:44Now, obviously,
11:08this is a static display robot.
11:10We just couldn't pry
11:11the actual production samples
11:13out of our engineers' hands
11:14at the lab this week.
11:15So we're going to be
11:16showing you videos
11:17of that robot
11:18in the coming weeks.
11:20This product generation
11:21of Atlas
11:22surpasses any other
11:24enterprise-grade humanoid.
11:26This robot
11:26has 56 degrees of freedom,
11:29mostly with fully rotational joints,
11:32and it has human-scale hands
11:34with tactile sensing
11:35in the fingers and palms
11:37for dexterous manipulation.
11:39Atlas has 360-degree cameras,
11:42see in all directions,
11:44so it understands
11:45when people are approaching,
11:47which is part of our
11:48holistic safety concept.
11:50Atlas is also strong.
11:52It can lift up to 110 pounds
11:55and reach up to 7.5 feet in the air.
11:59Atlas is also water-resistant.
12:01In real industrial environments,
12:02robots need to undergo washdowns,
12:05and it can operate
12:06at its full capabilities,
12:07including strength
12:09between minus 4
12:10and 104 degrees Fahrenheit.
12:13It's going to be easy
12:14to teach Atlas new tasks.
12:16Most can be trained
12:17in less than a day.
12:19Atlas can perform these tasks
12:21at a reliable,
12:22consistent pace
12:23for about four hours
12:25using its dual-swappable batteries,
12:26and when they run low,
12:28Atlas navigates back
12:30to its charging station
12:31and swaps its own batteries,
12:33getting right back to work.
12:35Finally,
12:36Atlas robots share intelligence.
12:38Through our Orbit platform,
12:40once one Atlas learns a skill,
12:42it can be shared
12:43with any other Atlas.
12:45These industry-leading specs,
12:47along with our industry-leading
12:48foundation models,
12:50allow Atlas to take on
12:51challenging tasks,
12:53tasks that Atlas will do
12:54autonomously on its own
12:56from day one,
12:58assisting people
12:59and prioritizing
13:00a collaborative,
13:01human-centric approach.
13:03And as we continue
13:03to build out
13:04our general-purpose technology,
13:06our customers will be able
13:07to use our Orbit software
13:08to teach Atlas
13:09new and specialized skills.
13:12Do you want to wrap us up
13:13by talking about our production
13:14and deployment plan?
13:15Of course.
13:16So we've already started
13:18production of this new robot
13:19at our Boston headquarters,
13:21and our entire 2026 supply
13:24has already been allocated
13:26to Hyundai Motor Group
13:27and our new AI partner
13:29you'll hear about
13:29in a few minutes.
13:30We'll add more new customers
13:32in 2027.
13:33But for now,
13:35we're planning additional pilots
13:36at Hyundai Motor Group
13:38and building a facility
13:39that we call
13:40the Hyundai Robotics
13:41Metaplant Application Center.
13:44This center
13:45is the cornerstone
13:46of our plan
13:47to build the most complete
13:48data set in the world
13:50to train humanoid skills
13:52and manufacturing,
13:53creating an engine
13:54for deploying
13:55tens of thousands
13:56of Atlas robots
13:57across Hyundai Motor Group
13:59facilities worldwide.
14:01It's a data factory.
14:03Hyundai Motor Group
14:04isn't just deploying
14:05the Atlas robots either.
14:07They are going
14:07to build them.
14:09Together,
14:09we are building
14:10a new robotics factory
14:12capable of producing
14:1330,000 Atlas robots a year.
14:16This is what's truly required
14:18to deliver the full promise
14:20of humanoids
14:21across the industrial,
14:23commercial,
14:24and home markets.
14:26We are so excited
14:28to bring Atlas into the world.
14:30Visit the Hyundai Motor Group booth
14:31over this week
14:32in the Convention Center
14:33to see more.
14:35And now,
14:35to share how Hyundai Motor Group
14:37will accelerate Atlas deployment,
14:38please welcome
14:39Sinhyun to the stage.
14:40Thank you.
14:41Aya and Zeph.
15:05At Hyundai Motor Group,
15:07our purpose in
15:08human-centered AI robotics
15:10is simple
15:11but powerful.
15:14We are creating value.
15:16Value that advances humanity
15:18through physical AI
15:20and real-world applications.
15:23That's why we focus
15:24on what works in the field.
15:26Robots that make work
15:28safer,
15:29special,
15:30and more scalable.
15:33Technology,
15:34not for show,
15:35but for real,
15:37for measurable,
15:38positive impact.
15:39We take
15:40Boson Dynamics'
15:42top research
15:42and development
15:43with our group's
15:44production power.
15:46Among all robotics players,
15:49this collaboration
15:50is unique.
15:54Here,
15:54we follow a shared
15:55three-step development approach.
15:59We are speeding up
16:00how our robots
16:02learn physical skills.
16:04We train them on data
16:06from Hyundai Motor Group's
16:07factories.
16:09Turning data
16:10into robot motion
16:11requires a central engine.
16:15That engine
16:15is
16:16RMAC,
16:18the robot metaplant application center
16:20in the U.S.
16:23RMAC drives robot experimentation,
16:26validation,
16:26and training.
16:29It anchors
16:30our robotics business
16:31and trains
16:32the robotics workforce.
16:36Accelerating
16:37physical AI capabilities
16:39takes more than
16:40a physical hub.
16:42It also takes
16:43cutting-edge technologies.
16:46That is why
16:47we partner with
16:48cutting-edge
16:49global tech leaders.
16:51One of these
16:52valuable partners
16:53is
16:53NVIDIA.
16:56Their AI infrastructure
16:57and development tools
16:59help us innovate
17:00robots faster,
17:02streamline development cycle,
17:05and deliver
17:05scalable solutions.
17:08We are also
17:09activating
17:09our group value
17:10network affiliates
17:12and applying
17:13their mass production
17:14expertise
17:15from cars
17:16to robotics.
17:18Each affiliate
17:19plays a distinct role.
17:23Hyundai Motor Company
17:24and GIA
17:24provide
17:25manufacturing infrastructure,
17:28process control,
17:29and production data.
17:32Hyundai Mobis
17:32works closely
17:33with Boston Dynamics
17:35to develop
17:35high-performance actuators.
17:38By combining
17:39their expertise,
17:41they are helping
17:42standardize
17:43key components
17:44and build
17:45a strong
17:45global supply chain
17:47for the group's
17:48robotics platform.
17:51Hyundai Globis
17:52optimizes
17:53logistics
17:54and supply chain
17:55flow.
17:58These are just
17:59a few
17:59of our affiliates.
18:02There are many
18:03more across the group
18:04and together,
18:06they form
18:07our group
18:08value network.
18:10By combining
18:11what each does best,
18:13we build
18:13a robust
18:14end-to-end value chain,
18:16one that delivers
18:16product quality,
18:18flexibility,
18:19and excellence.
18:21finally,
18:23we are setting
18:24a new standard
18:25for customer engagement
18:27and support
18:27based on our group's
18:29customer management
18:30know-how.
18:31And we are
18:32extending it
18:33onto advanced
18:34service platform.
18:36One great
18:37platform example
18:38is our
18:39robust
18:39as a service.
18:41It's a
18:42subscription
18:43model.
18:44It lowers
18:46entry barriers,
18:49reduces
18:49upfront costs,
18:51and accelerates
18:52return on investment.
18:55Our robot
18:56as a service
18:56model
18:57is designed
18:58to be easy
18:59to use
18:59and manage.
19:01It brings
19:02everything together
19:03in one service.
19:04That includes
19:05installation,
19:07over-the-air
19:07software updates,
19:09hardware maintenance,
19:10and remote
19:11monitoring
19:12and control.
19:14This lets
19:14customers
19:15focus on
19:16using the
19:16robots
19:17and achieve
19:18greater value
19:18with them.
19:20And
19:21as robots
19:22move from
19:22the factory
19:23into everyday
19:24life,
19:25this approach
19:26will be
19:26just as
19:27important
19:27for people
19:28at home.
19:29Which means
19:30robots
19:31have to prove
19:32themselves
19:33in the
19:33factory first,
19:35then
19:36they are allowed
19:37to do
19:37housework
19:37in the
19:38home.
19:40other
19:41show stages
19:42household
19:42load demos.
19:44We prove
19:45their performance
19:46in industrial
19:47settings
19:48where every
19:49move
19:49and every
19:50result
19:51is measured.
19:54That's how
19:55we build
19:55credibility.
19:57Proven
19:57inside Hyundai
19:59Motor Group
19:59expanded
20:00to the world.
20:02By 2030,
20:03we plan
20:04to sell
20:049.8 million
20:06vehicles per year.
20:08On the scale,
20:09Atlas moved
20:10from pilot
20:11to mass
20:11production,
20:12rolling out
20:13across
20:13our global
20:14manufacturing
20:15network
20:15to support
20:16our
20:17extension
20:17targets.
20:19And the
20:20demand
20:20will go
20:20far beyond
20:21manufacturing,
20:23like logistics,
20:25energy,
20:26construction,
20:27and facility
20:27management
20:28where safe,
20:30intelligent,
20:31scalable
20:31robots
20:32add real value.
20:33Today,
20:36Spot and Stretch
20:37are already
20:37operating at
20:38customer sites
20:39outside Hyundai
20:40Motor Group
20:41in numerous
20:41use cases.
20:44In the future,
20:46Atlas will play
20:47a core part
20:48as we expand
20:49across diverse
20:50industries
20:50with a wide
20:52variety of
20:52applications.
20:54Now,
20:55I will hand
20:56over to
20:57Ung-Jae-Ri
20:57from Manufacturing
20:59Solutions.
20:59He will show
21:01how we will
21:02bring this
21:02vision to life
21:03from training
21:05to deployment
21:06to scaling
21:07human-centered
21:08AI robots
21:09globally.
21:10Please welcome
21:11my dear
21:12colleague,
21:13Ung-Jae-Ri.
21:15Thank you,
21:25Sun-Yan.
21:25Human-centered
21:28AI robotics
21:29does indeed
21:30require
21:31a bold vision
21:32and it takes
21:33two bold
21:34partners
21:34to bring
21:35that vision
21:36to life.
21:38Boston Dynamics
21:38has world-class
21:40problem solvers.
21:42Hyundai Motor Group
21:42delivers the
21:43power of scale
21:44in production
21:45and when the
21:47two come together,
21:48they create
21:49powerful impact
21:50fueled by
21:51one of the
21:51most precious
21:52resources
21:53of our time,
21:54data.
21:57At Hyundai Motor Group,
21:58we have
21:59a unique
21:59advantage.
22:01We connect
22:01mobility,
22:03manufacturing,
22:04logistics,
22:05and robotics
22:05into one
22:06integrated
22:07loop.
22:09This holistic
22:09modeling
22:10of cross-domain
22:11data
22:11is changing
22:12everything,
22:13even factories.
22:15We are shifting
22:16from hardware-driven
22:17plants
22:18to software-defined
22:19factories,
22:20also known
22:21as SDF.
22:23They run
22:23and 24-7
22:24we are more
22:25flexible,
22:26safer,
22:27and smarter.
22:29Software
22:29lets us update,
22:32improve,
22:32and share
22:33intelligence
22:34across our
22:35global plant
22:36network.
22:37That is how
22:38Hyundai Motor Group
22:39will grow
22:40a connected
22:41SDF system
22:42which is learning
22:43from the others,
22:44which is lifting
22:45the whole network.
22:47Our first
22:48SDF,
22:50HMGs,
22:50is already
22:51up and running
22:52in Singapore.
22:54We'll grow
22:54from there,
22:56connecting data,
22:57people,
22:58and technology
22:59across every
23:00plant in the
23:01Hyundai Motor Group.
23:02Our SDF roadmap
23:04has one goal,
23:06achieve the best
23:07for our customers.
23:10We aim
23:11to adapt
23:11faster
23:12to local
23:12demands,
23:13raise product
23:14quality
23:15beyond today's
23:16benchmark,
23:17upper
23:17prices
23:18through efficiency
23:19and improve
23:21employee safety.
23:24SDF
23:24is also
23:25optimized
23:26for better
23:26collaboration
23:27between humans
23:29and robots.
23:31Across our
23:31plans,
23:32robots
23:32already handle
23:33tough,
23:34dangerous,
23:35repetitive
23:36work.
23:37Automation
23:38delivers clear
23:38benefits
23:39in the press,
23:41body,
23:41and paint shop
23:42where it reaches
23:44up to 100%.
23:45It promotes productivity,
23:48strengthens quality,
23:50and most importantly,
23:51keeps people safe.
23:54It's the unique approach
23:55of Hyundai Motor Group,
23:57human-centered
23:58automation.
23:59Even in a highly
24:00automated era,
24:02people remain
24:03essential,
24:04supervising,
24:05guiding,
24:05and maintaining
24:06robot colleagues,
24:08always retaining
24:08full control.
24:10As automation
24:12grows,
24:13the human role
24:13will become
24:14even more
24:15important.
24:17Robots will
24:17take on
24:18physical strain
24:19so people
24:20can focus
24:21on oversight,
24:22decision-making,
24:24and problem-serving.
24:26People
24:26will lead
24:27through software
24:28using digital twins
24:30to test scenarios,
24:32adjusting workflow
24:33with data-driven
24:34precision.
24:36But this
24:37doesn't happen
24:39overnight.
24:40the task
24:41must be trained
24:42systematically,
24:44safely,
24:45and with
24:46real-world context.
24:48That's the purpose
24:49of Almec,
24:51which will be located
24:52here in the U.S.
24:56Almec is where
24:57robots learn
24:58to work with people.
24:59They learn
25:00precision,
25:01repetition,
25:01and safety
25:02by mapping
25:03human movement
25:04into robot learning.
25:07Every lead,
25:08turn,
25:08and recovery
25:09motion
25:10all captured
25:11as real
25:13operational data.
25:15No company
25:15has more
25:16real-world factory data
25:18than Hyundai Motor Group
25:19because our SDM network
25:21is global
25:22and well-connected.
25:24This data
25:25creates better
25:25training,
25:26better training,
25:28creates safer,
25:29smarter robots,
25:30and smarter robots
25:32create better factories
25:33for people.
25:36Big input,
25:37great output.
25:39Our roadmap
25:39is clear.
25:41The Almec opens
25:42this August
25:43and creates
25:44the basis
25:44for future robots.
25:45by 2028,
25:48the Atlas
25:48global rollout
25:49begins within
25:50our SDF network.
25:53Atlas
25:53will start
25:54to perform
25:54high-precision
25:55sequencing tasks.
25:58By 2030,
25:59robots
25:59will handle
26:00complex assembly.
26:03Together,
26:04we saw
26:04robot colleagues
26:05working safer,
26:07lighter,
26:08and more meaningful.
26:10Because our robots
26:11don't compete
26:12with humans.
26:14They compete
26:14for humans.
26:18For our safety,
26:20efficiency,
26:21and well-being.
26:23This is Hyundai Motor Group's
26:25robotics philosophy.
26:27Robots expand
26:28the human reach
26:29and partnering
26:30for human progress.
26:33This concludes
26:34our look
26:35at the manufacturing
26:36power of Hyundai Motor Group
26:38in AI robotics
26:39to learn
26:40how we are
26:41advancing
26:41this vision.
26:43Please welcome
26:43Boston Dynamics
26:45and our new
26:46valued partner,
26:47Google D-Mine.
27:00Hey, Carolina.
27:03Hello.
27:05My name is
27:05Alberto Rodriguez,
27:06and I lead
27:07the Atlas Behavior
27:08Team at Boston Dynamics.
27:10And I'm Carolina Parada,
27:11and I lead robotics
27:12at Google D-Mine.
27:14Well, Carolina,
27:15finally,
27:16we are here.
27:17The last session
27:18of this presentation,
27:19and I'm so excited
27:20to finally be able
27:21to announce this.
27:22Very excited.
27:25At Boston Dynamics,
27:26we've been building
27:27humanoids
27:28and other mobile robots
27:29for more than
27:30two decades now.
27:31And a central focus
27:32of our work
27:33has always been
27:34physical intelligence.
27:35So imbuing our robots
27:37with the mastery
27:37of force and motion
27:39that has allowed us
27:41to demonstrate
27:42incredible athletic
27:43feats,
27:44as well as deploy
27:45them in real-world
27:47environments
27:48where they work
27:49for months at a time
27:49without failure.
27:51And all that work
27:53has led us to here,
27:54to today,
27:55to the new generation
27:56of Atlas,
27:57the world's most
27:58physically capable
27:59humanoid robot.
28:00But making Atlas
28:02into a product
28:03requires more
28:04than athletic performance.
28:06For humanoids
28:07to really deliver
28:08on their promise,
28:09they have to be able
28:10to interact
28:12with people naturally.
28:14They have to be able
28:16to understand
28:17and adapt
28:17to ever-changing
28:19real-world environments.
28:21And they have to be able
28:22to accomplish
28:22real complex
28:24dexterous manipulation tasks.
28:26And I hope you agree
28:27with me
28:28that recent advancements
28:30in AI
28:31have created
28:32a clear path
28:33to get to those capabilities,
28:35which has spurred
28:36our commercialization
28:37journey for Atlas.
28:39And that's why
28:39I'm so excited
28:40about announcing
28:42today our partnership
28:43with Google
28:44DeepMind's robotics team.
28:53This partnership
28:54sits at the world's
28:56leading intersection
28:57of robotics
28:59and AI.
29:00And we believe
29:01it will take us
29:02to a new era
29:04of human-centered
29:05AI robotics.
29:07What do you think, Carolina?
29:08Excellent.
29:09At Google DeepMind,
29:10our mission
29:11is to build AI
29:12responsibly
29:13to benefit humanity.
29:15We've always believed
29:16in the potential
29:17for AI
29:17to transform robotics.
29:19And we've seen
29:19firsthand
29:20how robotics
29:21can shape
29:22the frontier of AI.
29:23for AI.
29:25Robotics brings
29:26many different domains
29:27of machine learning
29:28together,
29:29from 3D perception,
29:31multimodal understanding,
29:32advanced reasoning
29:33and planning,
29:34and all of this
29:36in the physical world.
29:37That makes robotics
29:38a frontier application
29:40of AI,
29:41and one that will help us
29:42shape how we evolve
29:43AI beyond our screens
29:44and into the physical world.
29:47Over the last couple
29:49of years,
29:49we've seen remarkable
29:50progress on our Gemini
29:51and other foundation models
29:53that operate
29:54across text,
29:55images,
29:56and video modalities.
29:58And at Google DeepMind Robotics,
29:59we've been making
29:59rapid progress
30:00to bring all
30:01of Gemini's
30:02multimodal world
30:03understanding
30:04into the physical world
30:05by also adding actions
30:07as a new modality
30:08in Gemini.
30:09These models
30:11called Gemini Robotics
30:12bring advanced
30:13and body reasoning
30:14and generate actions
30:16that can directly
30:17control a robot
30:18in a way that is
30:19general,
30:20interactive,
30:21and dexterous.
30:23But these
30:24highly intelligent
30:25general purpose
30:26robot brains
30:27are most powerful
30:28when combined
30:29with a highly capable
30:30generalist body.
30:32And that is
30:33what brought us together.
30:35Boston Dynamics
30:35has defined
30:36the leading edge
30:37of robotics
30:37for decades,
30:38and we can't wait
30:39to work with them.
30:40We want to bring together
30:42the remarkable robots
30:43and know-how
30:43with our advanced AI models
30:45to build truly
30:46transformative robots.
30:49That's our vision
30:50for this collaboration
30:51with Boston Dynamics.
30:52We're looking to integrate
30:53our cutting-edge
30:54AI foundation models
30:55with Boston Dynamics
30:57new Atlas robots.
30:58And we'll aim
30:59to develop
31:00the world's most advanced
31:01robot foundation model
31:02to fulfill the promise
31:04of true
31:05general purpose human beings.
31:07Exactly.
31:07We're going to bring
31:08together
31:08the deep expertise
31:10in robotics and AI
31:11from both teams,
31:12and we're going to develop
31:14next-generation humanoids
31:16that combine
31:17the world-leading
31:18generalization
31:19and reasoning capabilities
31:21of Gemini
31:21with the strength,
31:23speed,
31:24fluidity,
31:25and reliability
31:26that characterize
31:27Boston Dynamics robots.
31:29The DeepMind robotics team
31:31has already shown
31:32that robots
31:33can generalize
31:35to newer scenarios
31:36and that they can perform
31:37manipulation tasks
31:38just from simple
31:39language instructions.
31:40And one of the core goals
31:42of this partnership
31:43is to significantly expand
31:45the scope
31:46of that generalization
31:47to include
31:48a much larger set
31:49of tasks,
31:51complex real-world tasks,
31:52like the ones
31:53that you'll see
31:54Atlas performing
31:55this week
31:56at our booth
31:56at CES.
31:58That's right.
31:59Rather than having
32:00a set of predefined
32:01loaded tasks
32:02onto the robot,
32:03we think robots
32:04should understand
32:05the physical world
32:06the same way we do.
32:08They should be able
32:08to learn from
32:09their experience,
32:10they should be able
32:10to generalize
32:11to new situations
32:12and get better
32:13over time.
32:15So whether it is
32:15to assemble
32:16a new car park
32:17or to tie
32:18your shoelaces,
32:19robots should learn
32:20the same way we do
32:21from a handful
32:22of examples
32:23and then get better
32:23very quickly
32:24with a little bit
32:25of practice.
32:26And building
32:27these models
32:28is critical
32:29both for reliability
32:30and scale,
32:31like the ability
32:32to easily
32:33onboard new tasks
32:34or the ability
32:35for our robots
32:36to improve
32:37their performance
32:37over time
32:38on their own
32:39or to be able
32:40to react
32:41to any unexpected
32:42things that the world
32:44puts in front of them
32:45are all key
32:46to successful
32:47large-scale
32:48deployments
32:49of this technology.
32:51Robots,
32:51they should be able
32:52to onboard
32:53new tasks
32:54in one day
32:55and they should be able
32:57to perform
32:57those tasks
32:58for hours on end.
33:00And we think
33:00that Atlas
33:01with its endurance,
33:03with its strength
33:03and its unique mobility
33:05gives us an opportunity
33:07to even go beyond
33:08human performance
33:09in many cases.
33:11So how do we get there?
33:12Well, we're currently
33:13building a very large fleet
33:15of Atlas robots
33:15and soon
33:17we'll start deploying
33:18them both
33:19in Boston Dynamics offices
33:21and in Google DeepMinds labs
33:23across the world.
33:25So stay tuned
33:26for more information
33:26to come.
33:27We're very excited
33:28about pulling our teams
33:29together
33:29and getting to work.
33:30Yeah.
33:31So that brings us
33:32to the end
33:32of today's presentation.
33:34Today marks
33:35an important first step
33:37on the path
33:38towards AI-enabled humanoids
33:40that delivered
33:41real-world impact.
33:42As robots move
33:44from industry
33:45to our daily lives,
33:46they will have to
33:47earn their trust
33:48through safety,
33:49reliability,
33:50and performance.
33:52Human-centered AI
33:53will make our lives
33:54easier and safer
33:55and ultimately
33:56more fulfilling.
33:58Together with
33:59Hyundai Motor Group,
34:00with our AI partners
34:01and with all of you,
34:03we have the opportunity
34:04to shape a future
34:05where humanity
34:06and robotics
34:07they thrive side by side.
34:09Thank you for being here
34:10with us today.
34:11Please join us
34:12on this journey
34:13to see what partnering
34:14for human progress
34:15looks like in action.
34:16Thank you, everyone.
34:17Für mehr Videos
34:20einfach abonnieren.
Kommentare