- 4 ore fa
Categoria
📺
TVTrascrizione
00:05benvenuti in questo speciale ai club parliamo di ai fisica al poco più di una settimana dalla
00:11gtc conference dove atteso sarà lo speech di james e wang sul palco a tracciare la via per
00:18il futuro dell'ei ai fisica un nuovo capitolo legata al settore dell'industria e della manifattura c'è
00:25una nuova partnership strategica tra abb robotics e invidia che sancisce un passo di evoluzione
00:31riduzione dei costi dei tempi di sviluppo e di accesso al mercato con noi ce ne parla il presidente
00:37di abb robotics mark segura un piacere averla con noi e allora partiamo dalla notizia dicevo abb
00:45robotics e invidia hanno stretto una partnership non è la prima ma di fatto questa segna un nuovo
00:50capitolo che cosa cambia strategicamente per il business con questa collaborazione
00:56è molto felice di annunciare questa partnership con invidia oggi, yes invidia e abb robotics
01:02hanno una storia di collaborazione ma oggi è una speciale annuncemento, noi siamo
01:06joining forces per aiutare la deploymenta di fisica ai a l'industria grade e come
01:12faremo a faremo l'abbiamo faremo l'abbiamo faremo l'abbiamo
01:15robot studio software di abb robotics che è usato da più di 60,000 ingegneri ogni giorno
01:22da più di 10000 di roboti e noi saranno inizia per l'abbiamo
01:26di creare robot studio hyperreale che avvano
01:28l'abbiamo di libri di nvidia e ovni
01:31queste libri di grafico di grafico di grafico di
01:34più importante di fisica di rappresentazione dell'ambiente
01:36con questa abbiamo avviamo
01:38avviamo molto forte software che è conosco a una
01:41molto più accurate performance di simulazione di roboti in
01:44computer e poi traslare la realtà in
01:48so the high degree of accuracy from the computer to the robot,
01:53we will expand that to include things around the robot,
01:56the environment around the robot,
01:57like the real lighting or a virtual lighting projection,
02:00the shadows, the texture of the material, the material properties.
02:04So with that, we can produce a larger, more realistic simulation of the robot
02:09than what it has to produce, an environment where it produces,
02:12so you can then add costs of engineering,
02:15have less time on pilots, physical pilots,
02:17and achieve a higher performance in a shorter time.
02:22Parliamo anche però di risultati concreti.
02:24I produttori potrebbero ridurre del 40% i costi di produzione,
02:28ridurre fino al 50% il time to market.
02:32Ebbene, in quali settori e aziende lei pensa che questo risultato
02:38si possa vedere con maggior efficienza?
02:41Parliamo dell'elettronica, parliamo dell'automotive,
02:43dove vede che si riduce il gap di fatto tra automazione e physical AI.
02:52As I said, the big, big thing is that we're delivering all these benefits
02:56at industrial grade, meaning that these outcomes for the robots producing parts
03:01will be at high speed, high accuracy, high repeatability,
03:05and the industrial safety that is needed.
03:08We expect all kind of companies to really benefit from that.
03:11But if you take the large companies producing cars, producing consumer electronics,
03:15this will be a massive saving on their product development, cost, and timelines.
03:19Imagine you need to produce a new product, be it an airplane part or consumer electronics.
03:25In this platform of RoboStudio Hyperreality,
03:27you will be able to plug in early designs of the part
03:31to consider the whole installation, including the robots,
03:34and to simulate that, do experimentations there of the product,
03:37how it performs with the robots,
03:39and then translate that to the robots, directly to the physical world,
03:43without the need of too much physical testing and breaking costly and slow things.
03:48So you will reduce the cost by faster engineering,
03:52you will reduce the cost because you don't need anymore
03:54most of the physical prototypes,
03:57and you will make it in a shorter time
03:58because these critical prototypes are the ones which are taking the longer
04:01on a product development process.
04:04So the big companies will benefit from that,
04:06but the flip side of this is that we aim to do this
04:09with the highest simplicity for the customers,
04:13with autonomy from the robots, with versatility from the robots.
04:16So we expect that the small companies can easily also adopt this software.
04:20They can prompt into the software,
04:22and the software can propose the robotic solution for them,
04:25and then they can tune it and push it to the robot.
04:27So we aim to lower the barrier as well for SMEs as we go forward.
04:34Autonomia e versatility are two key words,
04:37but recently we talked a lot about the agentic AI,
04:40the SaaS apocalypse, the so-called sell-off software on the market,
04:44we have seen it.
04:45Invece here we talk about the robotic AI,
04:48we talk about the physical, agentic AI,
04:50so thinking about the robots,
04:52what can be the new path,
04:54the new path, especially in terms of automation,
04:57the robots in the industry.
04:58How does the physical AI change in this context,
05:01which actually puts in,
05:03let's say,
05:05a greater consideration of the role of the software?
05:09Yeah, of course, we are focused on industrial robotics,
05:13in making our customers more successful in producing high quality,
05:18I mean, very competitive products.
05:20So physical AI is coming basically to give robots autonomy and versatility.
05:25Physical AI is robotics with autonomy.
05:28And why is autonomy and versatility so important?
05:31Because once again,
05:32autonomy means that robots can do more things in more places.
05:35Where the environment was dynamic or unstructured,
05:38things would change.
05:39The robot could only work with extremely repetitive kind of preset tasks.
05:44But if you would have the minor variability,
05:47that would not apply.
05:47So only large customers that could invest to structure the processes,
05:52the tools, and the product,
05:53they could succeed with robots.
05:55autonomy is breaking that constraint
05:57and versatility is adding more capabilities of robots
06:00going from one task to another task to another task.
06:03If you put together autonomy with versatility,
06:06of course, it helps the large customers,
06:08but you can imagine how it can help the long tail of a smaller companies,
06:12which are the backbone of most of the economies.
06:15That's very relevant.
06:16And physical AI,
06:17which is an enabler for autonomy and versatility,
06:20is so super important.
06:23E ora, per chiudere, di fatto una domanda che guarda anche al futuro,
06:26perché se simulazione e robotica sono sempre più allineate,
06:31quando ci dovremmo aspettare di fatto che i robot impareranno autonomamente
06:35all'interno delle fabbriche, della manifattura?
06:37E poi anche una questione di tempistiche,
06:39perché giustamente questa collaborazione parla della seconda metà del 2026.
06:44Quali sono i tempi di questi passaggi?
06:47This product robot studio HyperReality
06:50will already have this closed-loop learning
06:52from reality to digital twin, digital twin to reality.
06:56That's core to the fact that this is a hyper-realistic software platform.
07:00So we are training already the robots with synthetic data,
07:04I mean, virtual data,
07:05but it's not any longer just drawing.
07:08There are virtual components,
07:10as I said, of the parts, of the lighting, of the vibrations,
07:14all the physics around a real world in virtual.
07:17But at the same time, very importantly,
07:20this software, which works on the cloud,
07:22robot studio works in the cloud,
07:24the customer will have the robots connected,
07:26will have the cameras connected.
07:27So everything that happens in real feeds back.
07:30So the AI models being trained on robot studio HyperReality
07:33are to begin with industrial-grade
07:36because they work with industrial-grade virtual data,
07:39but then they will be enhanced in real time with real data.
07:42So kind of doubling down the accuracy,
07:45the learnings already from day one.
07:47That's planned to be available from the beginning.
07:51And thank you for being with us.
07:53Grazie per essere stato con noi,
07:55Mark Segura, presidente di ABB Robotics.
07:57Thank you very much.
07:58It was a pleasure too.
07:59Voi restate con noi su Class NBC
08:01per seguire tutti gli aggiornamenti sull'AI
08:03e sulla cronaca dei mercati.
Commenti