# COMPUTEX 2026 CEO Keynote: Qualcomm

https://www.youtube.com/watch?v=BH6pg0LY8Fw

[16:30] AI once lived behind screens.
[16:34] Now it steps into the real world.
[16:40] Intelligence is no longer confined to the cloud.
[16:44] The brain is everywhere.
[16:49] AI now has a body sensing, acting, transforming the world it touches and it's binding every intelligent device into one living network.
[17:03] It sees, it moves, it connects.
[17:09] This is where the next era begins.
[17:13] Global visionaries, 500 startups, 23 nations of innovation.
[17:22] Computex 2026.
[17:25] We make it happen together.
[17:29] Ladies and gentlemen, let us welcome our
[17:32] Host, Mr. James Huang, chairman of Taiwan External Trade Development Council, Titra, to the stage.
[17:52] For five centuries, Michelangelo's creation of Adam has shown God given life to man, the first intelligence on earth.
[18:04] In the 21st century, we have created a new intelligence.
[18:11] The hand is ours.
[18:14] The intelligence reaching back is AI born from human hands.
[18:21] Good afternoon, ladies and gentlemen and friends from the international media.
[18:27] Welcome to Taipei.
[18:27] Welcome to Computex 2026.
[18:38] Computex convenience in a time of two great waves.
[18:43] The first is geopolitics.
[18:46] The postwar order is unwinding and until a new one takes shape.
[18:53] Uncertainty will be the norm.
[18:56] The second is technology.
[18:58] AI is not just another tool.
[19:03] It is the beginning of a new kind of civilization.
[19:07] Language, memory, initiative, society being assembled inside the very machines that fill these holes.
[19:21] Computex witnessed the rise of the PC, internet, mobile, and cloud.
[19:30] each wave making machines more powerful and more central to human life.
[19:36] Now AI has evolved from tools to actors.
[19:41] This is why it is more than a technology story.
[19:45] It is a human story.
[19:50] The world used to come to Computex to get updated on new tech.
[19:57] Today the world comes to see where humanity goes next.
[20:03] Computex is no longer just a tech show.
[20:07] It is a conduit to the new world.
[20:13] That is the spirit of this year's theme AI together.
[20:18] The next error will not be defined by any one country, company or AI model.
[20:25] It will be built together.
[20:31] AI and humans together.
[20:37] Taiwan and the world together.
[20:40] Step out of this hall and you walk into
[20:44] The future being built.
[20:46] This year is truly exceptional.
[20:50] The largest Computex in our history across AI and computing, robotics and mobility and next-gen tech.
[21:02] The future is taking shape around us in Nang Exhibition Hall One, Hall Two, the Taipei World Trade Center and also at the Taipei International Convention Center where Ambidia will host GTC Taipei.
[21:23] Our CEO keynotes bring four global chip leaders to one stage.
[21:29] Quaccom's Cristiano Amo, Marvel's Matt Murphy, Intel's Deepam, and NXP's Rafael Sotomayo.
[21:43] Together, they map the full landscape of
[21:47] AI computing from age devices to data center connectivity to the silicone powering AI's next era.
[21:59] all the way to physical AI moving into the real world.
[22:06] These four CEOs are not flying here to make announcements.
[22:12] They are flying here because the conversations that matter for the next decade of AI happen in this city this week.
[22:25] And this year, for the first time, Computex dedicates an entire pavilion to robotics.
[22:34] the AI and robotic drone at Taipei World Trade Center where leaders from Intel, E Inc., Highwing, Solomon, and Texas Instruments
[22:48] and many more are showing the world what physical AI looks like when it leaves the lab and walks into the world.
[22:59] Right inside the same building on the morning of June the 2nd, four forum sections take the stage in a row.
[23:09] All on physical AI with the senior executives running robotics at Nvidia, Quaccom, ABV and NXP.
[23:22] Four decision makers, one room, one morning.
[23:27] That density does not exist anywhere else in the world.
[23:33] Beyond robotics, our forum gathers 26 global leaders across 28 sections.
[23:43] Microsoft, Google Mind, Seammens, Sypnosis and many more.
[23:51] covering the full stack of AI
[23:54] from chip design to data centers to
[23:58] industrial development.
[24:01] And at Innovex,
[24:03] nearly 500 startups from 23 countries
[24:07] are showing where AI goes next
[24:11] in robotics, healthcare, materials, and
[24:14] industrial intelligence.
[24:17] These are not demos anymore.
[24:20] They are deployments waiting for scale.
[24:25] All of this is happening here because
[24:28] Taiwan plays an indispensable role in
[24:31] this AI revolution.
[24:34] Computex has turned Taiwan into the AI
[24:39] hub of the world.
[24:43] Last year I asked when AI is everywhere,
[24:47] what will the world need?
[24:51] This year the question deepens
[24:54] When we have given life to a new intelligence.
[24:59] What kind of civilization do we want to build together?
[25:04] Pope Leo the 14th put it well.
[25:04] The AI challenge is not technological but anthropological.
[25:14] AI may be fast and precise but without purpose.
[25:14] Humanity carries memory, morality and meaning, but it may struggle with the bottlenecks of the old world.
[25:31] The future must be built by both machines expanding what we can do.
[25:36] Humanity deciding why it matters.
[25:43] Intelligence anchored in trust.
[25:47] Power placed in service of life.
[25:52] As you walk through these halls,
[25:55] Look beyond chips and machines.
[25:59] Look at the beginning of a new relationship between humanity and intelligence.
[26:05] Look at the future we are being called to build together.
[26:10] Thank you and once again, welcome to Computex 2026.
[26:16] Thank you, James.
[26:21] Next, let us welcome our co-host, Mr. Jason Chen, chairman of Taipei Computer Association, TCA, to the stage.
[27:04] Hi everyone, disc distinguished guest, the members of international media, industry leaders, ladies and gentlemen.
[27:14] Good afternoon.
[27:14] Welcome to Taipei.
[27:17] I'm Jason Chen.
[27:20] I'm the chairman of Taipei Computer Association.
[27:22] First, I would like to express my sincere gratitude to our media friends from around the world for longstanding coverage of Computex year after years.
[27:36] Your reporting enables the global technology industry to identify the next important direction emerging from the event.
[27:44] The theme of Computex this year is AI together.
[27:48] This is far more than a slogan.
[27:52] It reflects a real transformation taking place across the global technology industry.
[27:58] AI is rapidly moving from an innovation showcase to a real world deployment.
[28:07] From isolated breakthrough to system integration and from individual competition to collaboration around the global ecosystem.
[28:15] This year we are proud to host more than 1,500 exhibitors across 6,000 booth spanning nang is hipp center the world trade center and the taipe international convention center.
[28:38] This expansion not just about scale.
[28:42] It signals the global AI ecosystem is gathering in Taipei in an accelerating pace.
[28:47] Today AI is no longer just a race for models or computing power.
[28:51] It is a competition of complete system capacity.
[28:55] from AI chips, servers and PCs to high speed connectivity, data governance, thermal management, energy efficiency
[29:08] and smart application.
[29:11] AI is fast integrating into real world scenarios including manufacturing, healthcare, retail, transportation and urban governance.
[29:19] The trend reflects this year official best a best choice award reinforced the ship clearly demonstrating that global AI competition is moving from a race for computing power to a quest for such a value.
[29:41] If I will summarize the core focus of computer tax this year in five key words, they will be computing, connectivity, storage, efficiency and application.
[29:58] Dear media friends, computer tax 2026 promise to be extraordinary.
[30:05] We sincerely invite you to explore the exhibition halls attending forums. Visit
[30:11] the best choice of world display areas.
[30:14] and emerge yourself in innovand.
[30:21] how AI is moving from model to systems.
[30:24] This year we're also extending the experience experience beyond the venue.
[30:30] into the city.
[30:33] Tomorrow night with Taipei 101 as a international Denmark stage, thousand MIT drones will light up the night sky, bringing together technology and art for audience around the world through a live streaming.
[30:52] On June 3, the Innovax night party will bring together global startups, investors, and industry partners, making Taipei an important platform for global innovation exchange.
[31:07] In closing, AI together is not just our theme.
[31:11] It is shared direction for the
[31:14] entire global technology industry.
[31:17] Thank you once again.
[31:20] Welcome to Taipei and welcome to Computex 2026.
[31:24] Thank you very much ladies and gentlemen.
[31:28] Thank you.
[31:28] Thank you Jason.
[31:30] Ladies and gentlemen, this marks the conclusion of our global press conference.
[31:37] We're truly grateful for your continued support and for being with us today.
[31:39] We will now proceed to the Qualcomm opening keynote speech.
[31:45] Let us once again invite Hitra Chairman James Huang to the stage to officially introduce our opening keynote speaker.
[31:52] Welcome, James.
[32:06] Ladies and gentlemen, as we move into our keynote program, I'm pleased to welcome back a familiar face, a good
[32:15] friend.
[32:17] For the third consecutive year, Cristiano Arm has chosen Computex as the stage to share Qualcomm's vision with the world.
[32:31] The continuity speaks for itself.
[32:32] This year, everything Qualcomm has been building toward AI8H, intelligent devices, the connected world is becoming reality.
[32:46] There is no better person to open our keynote series.
[32:51] But first, please turn your attention to the screen for a special opening from Quarkcom.
[33:04] We're stepping into a world where AI feels less like a tool.
[33:08] Good morning. Here's your daily brief and more like a teammate.
[33:15] It's day five of strength training, the dreaded leg day.
[33:19] Okay, let's do it.
[33:22] Always ready, always learning.
[33:28] You're approaching your personal best.
[33:31] Push yourself.
[33:34] Built to keep up with real life, not slower down.
[33:37] Nice work.
[33:39] Intelligence that moves with you across the moments that make up your day.
[33:44] working quietly in the background so everything else feels easier in the foreground.
[33:51] And when the how is handled, thanks.
[33:54] You get more freedom to focus on the why.
[33:56] More flow, more creativity, more time for what you care about.
[34:03] This Qualcomm technology adapts to you.
[34:06] Your meeting canceled, so I freed up your schedule.
[34:07] You mentioned needing a new shirt for the party this weekend.
[34:11] Want to pick one up on the way to the office?
[34:12] Sounds good.
[34:17] Responding to your life, your needs on your terms.
[34:21] It's more than just a helping hand.
[34:23] It's an extra set of eyes and ears that understands the world around you.
[34:28] I've prepared a few color matches with your wardrobe, that sees what you see.
[34:33] The white t-shirt would also go well with your tan jacket.
[34:36] Here's what you hear.
[34:37] I'll take both.
[34:39] Those shirts are selling well.
[34:41] I've drafted a social post to drive engagement and placed an additional order with the wholesaler.
[34:48] Technology that meets you where you are and takes you where you need to go.
[34:57] Your ride should be about 35 minutes today.
[35:00] Okay.
[35:01] I've summarized the latest product design feedback to review on the train.
[35:04] Great.
[35:06] Killed that photography podcast I mentioned earlier.
[35:08] Done.
[35:11] A network of agents working around the clock to create a brighter future for everyone.
[35:18] This is intelligent computing everywhere.
[35:21] Driven by Qualcomm innovation.
[35:25] Smarter, faster, more seamlessly connected.
[35:29] Transforming the devices you rely on into partners you trust.
[35:36] It's computing that never misses a moment.
[35:41] Looks good.
[35:41] So neither do you.
[36:02] Ladies and gentlemen, Chris, president and CEO of Qualcomm.
[36:06] Cristiano Arm.
[36:14] Thank you.
[36:15] Good to have you back here.
[36:16] Thank you so much.
[36:17] Thank you.
[36:23] Beautiful.
[36:23] Thank you.
[36:24] Thank you very much.
[36:26] Hello everyone.
[36:29] Very very happy to be here at Computex.
[36:31] Um, this is a special time I think for Qualcomm, a special time for technology,
[36:37] and it's such an honor, I think, to be able to speak with you all here at Computex.
[36:41] So, thank you so much for being here.
[36:44] Um, as I said, I think we are Qualcomm incredibly excited about the future.
[36:49] Is there some Qualcomm folks here in the room?
[36:52] All right.
[36:52] Thank you.
[36:56] Thank you so much.
[36:57] uh before I start this presentation uh this is this presentation is is not about Qualcomm it's not about Qualcomm
[37:05] it's really at the end of the day I think you know evolution of technology is really a combination of companies not one company is never responsible for everything it's really about partnership
[37:17] so first thing I wanted to say a big thank you to all uh our Taiwan partners
[37:25] first all of our suppliers ers all of
[37:28] our development partner a special thank
[37:30] you to TSMC which has been an incredible
[37:32] partner on this journey. So thank you to
[37:35] our suppliers uh and our developers and
[37:39] then also a big thank you uh to the
[37:44] entire ecosystem in the partnerships and
[37:47] many of you which are here today. Thank
[37:50] you so much. This is not about Qualcomm.
[37:52] This is all about you and an incredible
[37:55] ecosystem that exactly drives uh
[37:58] everything in technology with a
[38:01] wonderful partnership companies that are
[38:04] leaders in technology today. Uh they are
[38:07] leaders because of those partnerships.
[38:09] So thank you very much and thank you all
[38:12] for having me here speaking to you.
[38:21] Okay, we are incredibly excited because
[38:24] a lot of the things that we have been
[38:26] talking about it is really started to
[38:29] become real and it's very easy to see.
[38:33] I have one job today. One job today. Uh
[38:37] and by the way, I won't talk about a
[38:39] data center. So if you came here to hear
[38:41] about our data center pro projects, I
[38:44] will tell I'll tell a little bit a
[38:46] little bit but we'll talk about this at
[38:48] the end of the month. But I have one job
[38:50] today. I want you to understand how AI
[38:56] is going to evolve for all the devices
[38:59] because it's the new form of computing
[39:01] that is touching every single device.
[39:05] And if you walk out of this keynote
[39:08] today understanding what's going to
[39:10] happen to all the devices of the edge
[39:12] and how how much we're excited about it,
[39:16] I have done my job. So I'm going to do
[39:17] my best uh to tell you how this future
[39:21] is going to look like. It's very clear
[39:23] to us. We have conviction what's going
[39:25] to happen and we're going to make it
[39:27] very concrete for you. But it's an
[39:29] incredibly exciting time for all of the
[39:33] computing technology and all the devices
[39:36] uh that we use every day and the ones
[39:38] we're going to use it. So about two
[39:40] years ago, we talked about how how AI
[39:43] will change the human computer interface
[39:47] and as a consequence will change the
[39:50] architecture of all of our personal
[39:51] computing devices and that is starting
[39:54] to become a reality in 2026. That's why
[39:57] we call it Qualcomm. 2026 is the year of
[40:00] agents and it's now how AI is really
[40:04] evolving
[40:05] and is going to get to incredible amount
[40:08] of scale. It is evolved from simply
[40:11] answering prompts and work as a tool
[40:15] that augments how we as humans work with
[40:19] our computers into something that can
[40:21] take action on their own. I'm going to
[40:23] give you some examples. I think what you
[40:25] see today in the screen at home you're
[40:28] going to have agents going to basically
[40:31] update everyone on your activity and
[40:33] your schedule all of the things that you
[40:35] need to do at work will be with you and
[40:39] think about what are the projects that
[40:41] you need to do what are the tasks that
[40:43] you need to do it flags the decision and
[40:46] it's going to be working with you and in
[40:47] your personal life um will give you
[40:51] information that is relevant to you all
[40:54] of the updates that you need work with
[40:56] you on a schedule. Everything's going to
[40:57] be prioritized and it's going to be a
[40:59] little different. It's going to feel
[41:00] more autonomous. It's going to feel more
[41:03] personal and it's going to feel like
[41:06] really a companion. And all of this is
[41:09] going to happen proactively across all
[41:12] of your devices.
[41:14] And for that to be understood, we have
[41:17] to go back to a very uh basic element of
[41:20] this transition. We have been talking
[41:22] about this, right?
[41:25] which is the phone today is at the
[41:29] center of your digital life. And I I
[41:33] remember when I presented something
[41:35] similar like this uh when we talk about
[41:38] it uh at Snapdragon summit I actually
[41:40] talk about here at Computex that the
[41:43] phone is the center of a digital life
[41:44] and therefore everything is around the
[41:47] phone is how we interact with the phone
[41:49] with the OS and the applications even
[41:52] the other wearable devices they are
[41:55] around this phone ecosystem
[41:59] but now this is different those agents
[42:02] become the center of your digital
[42:04] experience. So including the phone uh
[42:08] and the other devices they now around
[42:11] the agent it's not about an extension of
[42:13] the phone and the digital ecosystem is
[42:15] no longer at the phone itself in the oss
[42:19] and the applications
[42:21] and what happened those devices become
[42:24] end points for agents and agents is not
[42:29] they are not trapped they are not
[42:32] attached to a single ecosystem.
[42:36] and everything that connects you with an
[42:39] agent becomes an 8 point for AI and
[42:42] that's a very important change once you
[42:45] understand that change you understand
[42:47] how the whole mobile industry is going
[42:50] to change uh the things that we take for
[42:53] granted today uh are going to be
[42:56] different tomorrow and there's one thing
[42:59] I can tell you coming from Qualcomm I've
[43:01] been in the company for 30 years uh One
[43:04] thing is guaranteed the mobile industry
[43:06] is not static. It changes all the time.
[43:11] So those end points for AI. I am going
[43:14] to talk about what is really the surface
[43:18] area. I'm going to remind you of what
[43:21] are the scale of those devices. They're
[43:23] about six billion phones today. Two
[43:25] billion um personal AI devices which is
[43:29] the evolution of wearables. And that is
[43:31] exactly a small number when you think
[43:34] about those things as an end points of
[43:36] agents. Two billion PCs, half a billion
[43:41] connected cars and people will
[43:43] experience and interact with those
[43:45] agents on this devices.
[43:49] And and as I said in the beginning, the
[43:52] role of this conversation today is help
[43:54] you understand the role of those devices
[43:57] at the edge for this next phase of AI.
[44:02] And we at Qualcomm, we actually told you
[44:05] about those devices running AI before it
[44:08] was popular, but now it's very easy to
[44:10] see. So once you understand the scale,
[44:15] one important message that I want to
[44:18] deliver to you is that today's device
[44:21] were not designed for those experiences.
[44:23] They're not. It will require a different
[44:26] kind of device when agent AI becomes the
[44:32] number one AI function uh in our daily
[44:36] lives especially for personal computing
[44:39] as we interact with those devices.
[44:42] All of those devices today they have
[44:45] been built for actions initiated by the
[44:49] user not by the agent. And I'm going to
[44:52] break that down for you. The agents are
[44:55] different. They operate it all the time.
[44:58] They carry context forward. They will
[45:01] orchestrate multiple task reliable and
[45:04] they and securely. But they are going to
[45:06] do that on your own, not necessarily
[45:09] with human intervention. That alone give
[45:12] you a glimpse how those devices are
[45:14] going to change and how the architecture
[45:16] of those devices are going to change.
[45:19] And when you think about that,
[45:21] the agent acts on intent. As you have
[45:25] now AI taking a role in the man machine
[45:30] interface, you understand human
[45:32] intentions. We'll proactively understand
[45:35] uh the text that it needs to react with
[45:37] the human. We'll take a go. We'll break
[45:39] it into several steps. will coordinate
[45:42] across systems, across data sets. Your
[45:45] own data set that lives on the device,
[45:48] the data set that is unique to you on
[45:50] your personal graph and the data set is
[45:52] going to be in the cloud. It plans, it
[45:54] executes, you'll verify, you will keep
[45:57] interacting it until the task is done.
[46:00] And that by definition changes the
[46:03] hardware with different different kind
[46:06] of hardware to be able to do this
[46:08] because it's not just what is initiated
[46:11] by the user the agent is going to be
[46:13] running on its own is also change the
[46:15] operating system and changing the
[46:17] application
[46:18] and I think that's the incredible
[46:20] opportunity that we have in computing
[46:22] across this entire safer surface area of
[46:26] devices which are going to be upgraded
[46:29] for this age of AI. time you need to be
[46:32] able to tight integrate across all the
[46:35] compute domains.
[46:37] You have strict power and latency
[46:40] constraint. I cannot emphasize enough
[46:44] the importance of power. Think about
[46:45] your phone. If it is challenging to make
[46:48] your phone last all day with you
[46:52] operating it, what happens when you and
[46:54] the agent is operating it? That's an
[46:56] incredible engineering challenge from
[46:59] power and latency.
[47:02] Those devices needs to be able to
[47:04] support planning, reasoning,
[47:06] coordination across the system. And for
[47:08] that you can start to understand how the
[47:11] compute is going to evolve. You need a
[47:15] very strong and power efficient CPU for
[47:19] the orchestration of tasks. I think the
[47:22] industry now understand the importance
[47:23] of CPU for orchestration and where the
[47:26] orchestrator is going to run. You need
[47:29] very power efficient and high compute
[47:32] density of NPU and GPU to be able to run
[47:35] local models.
[47:38] One other thing that I'm going to do
[47:40] today is to show clearly how AI is going
[47:42] to be distributed across cloud and
[47:44] device. I almost feel that it's going to
[47:47] be so obvious that we're going to start
[47:48] talking about it. We're going to stop
[47:50] talking about cloud and edge because
[47:51] it's all going to be one system. And for
[47:53] that, you're going to need uh the
[47:55] computing power and those devices to do
[47:58] that on the NPU and then the GPU.
[48:04] You need to have sensor data. Context is
[48:08] super important. Without context, you
[48:10] can not have the agent to be useful and
[48:13] pro and for you to be useful and be
[48:17] proactive.
[48:18] All of those things are going to be
[48:20] necessary as we think the next
[48:22] generation of devices for the agent AI
[48:26] and an upgrade is truly coming.
[48:31] So, and it's not the same thing for the
[48:36] for every device. It is not. We know
[48:39] this at Qualcomm in personal devices.
[48:42] That means all day intelligence, very
[48:46] focused on sensors in context that
[48:48] matter to you. Incredibly power
[48:51] efficient because you're going to carry
[48:54] you are mobile by definition. You're
[48:56] going to carry with you and you have
[48:58] high speeded connectivity in cars and
[49:00] robots. That means sustaining processing
[49:04] power on the most demanded condition and
[49:08] also independent of connectivity and
[49:11] cloud. when you have to and in the data
[49:14] center inference will require incredible
[49:18] amount of scale and I'll walk you to it
[49:21] when we think about the demand for
[49:22] tokens
[49:24] and power efficienc
[49:28] I'm going to tell you something
[49:31] for companies that have the benefit of
[49:34] being working in the mobile industry the
[49:36] mobile industry for decades you had a
[49:40] different growth rate of computing power
[49:43] and availability of energy. So you have
[49:46] to come up with solutions that allow you
[49:49] to get all the computing power that you
[49:51] need with the available energy because
[49:54] the battery has to last all day and is
[49:57] limited in its size. That's also going
[50:00] to be true when you think about what's
[50:02] going to happen on the data center. But
[50:04] we'll tell you more about that at the
[50:05] end of the month.
[50:08] And that in itself is the Qualcomm
[50:13] advantage
[50:14] because
[50:16] we have been working with all of all of
[50:20] those devices at the edge. All of it and
[50:22] especially the new Qualcomm that had
[50:24] incredibly diversified.
[50:26] We now have an ability to build systems.
[50:30] There's sub 2 millwatts from an earbud
[50:34] with micro power Wi-Fi that connect to
[50:37] an agent for personal AI audio device
[50:41] all the way to uh you know kilowatts
[50:44] what's going to happen in the data
[50:45] center and I think that is the
[50:48] incredible opportunity that we see for
[50:51] us and for our partners when all of
[50:54] those devices are going to change for
[50:57] this future of agent AI time and for
[51:00] that you need dedicated computing
[51:04] solution that scale all the way down to
[51:08] a submill device uh to a 2,00 kilowatt
[51:13] and that's an incredible opportunity I
[51:16] think that we have in front of us
[51:19] leading performance per watt across
[51:20] smartphones PC automotive comput and
[51:23] robots and extending that to the data
[51:25] center I think that's our mission and
[51:28] that is uh the incredible opportunity.
[51:32] The last thing I'm going to tell you as
[51:35] we keep moving in the presentation,
[51:37] it is
[51:40] there is excitement about what's
[51:43] happening especially on on the PC side
[51:47] that with all of those agent AI. So as
[51:52] we go through the presentation, I will
[51:54] come back to this and I will come back
[51:57] to the phones because I believe that
[52:01] what's going to happen on the phone is
[52:03] going to be a very good proxy of what
[52:06] could happen on PC and other devices.
[52:08] But hold that thought.
[52:10] So, so I'm going to break this down for
[52:14] you across the different categories of
[52:16] devices and I want to show you how those
[52:19] experiences are evolving. I want to make
[52:21] it less abstract and more concrete for
[52:24] you because this is happening right now.
[52:26] So, I want to start with the devices
[52:28] that really powered by Snapdragon
[52:30] platforms. You will see agents now being
[52:34] deployed and starting to get scale
[52:36] across phones, PC, and this new category
[52:39] of personal AI devices. I would like you
[52:41] to watch a video. Let's run the video.
[53:01] Move.
[53:15] I'm having friends over tonight. My
[53:17] smart home devices start getting ready
[53:19] on their own.
[53:21] >> Cork, search for the same item.
[53:24] >> Found the same suit for you.
[53:28] Ask Cloud Co-work to create a daily
[53:29] briefing for you. It can pull from
[53:31] things like your email, your calendar,
[53:32] your local files. A powerful CPU like a
[53:34] Snapdragon one just means all this will
[53:37] run faster, more privately, and more
[53:39] efficiently.
[53:40] >> I walk in, got
[53:43] pull up like honey, I'm home, always
[53:47] tricks up my sleeve. It feels
[53:52] go crazy. I have the setup running on my
[53:54] laptop powered by Snapdragon because its
[53:56] powerful CPU lets me run multi- aent
[53:58] workflows in parallel.
[53:59] >> It has a dedicated AI engine. I don't
[54:00] throttle on any long coding sessions.
[54:02] You just tell it what to do and it does
[54:04] it.
[54:06] >> Can I just get one item from box 3?
[54:19] If you know you know
[54:23] you know you know what I came here to do
[54:26] it.
[54:35] So this is incredibly exciting and it's
[54:38] really already happening and you can see
[54:42] the agent isn't tied to the device. It
[54:46] actually moves with the user and is
[54:48] there with the user regardless of the
[54:51] device that you have. So that is one way
[54:54] for you start to understand this change
[54:56] when the center of the ecosystem
[54:59] especially mobile is no longer the
[55:00] smartphone. It's about the agent and the
[55:03] agent's going to be across all of those
[55:05] devices.
[55:06] This is incredible and as I said it is
[55:10] easy to see now why we say that this the
[55:12] year agents and adoption is really
[55:16] accelerating and you started to see
[55:18] beginning to run on commercial devices.
[55:21] Orchestrators like Open Claw and Hermes
[55:24] are now running on Snapdragon. Aentic
[55:27] assistants like claw desktop are running
[55:30] native natively on Snapdragon PCs and
[55:34] you see this thing only accelerating and
[55:37] how the users are going to use more and
[55:39] more in their machines and the machines
[55:41] going to be running on their own doing
[55:43] things and cloud platforms for example
[55:47] like perplexity computer are now
[55:49] building orchestration layers that
[55:51] connect to the device. Google is
[55:54] bringing Agentic AI directly into
[55:57] Android with Gemini intelligence.
[55:59] Microsoft's doing the same and partners
[56:02] even like Humane are developing full
[56:04] Aentic operating system running on
[56:06] Snapdragon. It's it's truly incredible
[56:09] and I'm going to use now this moment
[56:11] before I keep going to presentation to
[56:13] tell you about the point I'm going to
[56:14] make about phones because actually an
[56:16] important point to understand there's
[56:18] six billion phones and how uh user
[56:21] experience going to be defined
[56:23] those orchestrators are a significant
[56:26] milestone in the transition of AI to
[56:29] agents and I remember for example when
[56:32] orchestrators like open claw became
[56:34] available lot of lot lot of consumers
[56:37] were buying uh computers. They're buying
[56:40] Mac minis and other computers and they
[56:42] run uh those agents and the computer
[56:44] were doing tasks and you can message to
[56:46] the computer.
[56:48] But this is not unique to those early
[56:51] adopters. This is how computers going to
[56:53] be utilized by a lot of people. If you
[56:56] have a phone, you don't have the ability
[56:59] to buy a computer, connect it to a
[57:01] better, put it in a backpack and carry
[57:03] it with you. You're going to carry one
[57:04] device and those things are going to
[57:06] happen in the phone. So those devices
[57:08] are going to have two personalities and
[57:10] I'll come back to this. One personality
[57:12] is you going to operate the device as a
[57:14] human. That's how the device of today
[57:17] has been designed uh to to operate but
[57:20] the agent is going to operate the device
[57:22] as well and that's how devices are going
[57:25] to be in the future. And that's going to
[57:27] happen on the same device and the same
[57:30] device doing both like phone is also
[57:32] going to influence how it's going to
[57:34] happen on PC with laptops and those
[57:37] personal AI devices are going to get
[57:39] scale because they're very natural for
[57:41] Asian glasses for example as close to
[57:44] your eyes to your mouth to your ears
[57:46] they became very natural and that is
[57:49] this incredible transition we're going
[57:52] to see in compute
[57:54] but it's not just unique to personal
[57:57] computing. This is also happening as we
[58:01] think about the transition that we're
[58:04] going to have to the physical world. The
[58:08] physical world is one of the most
[58:09] exciting areas in our industry right
[58:11] now.
[58:13] Cars, robots, industrial systems are
[58:18] going to be also
[58:20] significant
[58:22] uh end points for a gent AI and in those
[58:26] systems you have other requirements.
[58:28] Latency safety are measure in
[58:31] milliseconds in millimeters every single
[58:34] what matters. So I want to show you a
[58:37] demo next and I'll walk you through it.
[58:39] uh for example how things are going to
[58:42] work in a car um when you think about
[58:46] the car
[58:49] there will be two layers of
[58:50] intelligence.
[58:52] So in the cockpit, AI will personalize
[58:55] your experience. And that is the same
[58:58] thing that we just talked about. The the
[59:00] agent is your agent regardless of the
[59:03] device. When you're behind the wheel,
[59:05] you're going to take that experience
[59:07] with you is going to interact with with
[59:09] the agent. It doesn't matter if you're
[59:12] going to be on the car, on a personal AI
[59:14] device, on the phone, on your PC,
[59:16] because the agent is at the center. all
[59:18] the devices around it including the car
[59:21] but also the car has other tasks besides
[59:25] interacting with you is interacting with
[59:27] you within the context of the car and on
[59:30] the road you also have physical AI that
[59:32] is perceiving it is planning it is
[59:35] acting using cameras radar sensors and
[59:39] maps for your navigation
[59:42] so that's the interesting thing you have
[59:45] two different layers of intelligence You
[59:48] have the interaction with you in the
[59:50] cockpit which is personal to you when
[59:53] you are within the context of the car,
[59:54] but you also have the physical AI that
[59:57] drives the car for you. And this is
[01:00:00] designed to operate as one integrated
[01:00:04] system. Watch this video of this in
[01:00:09] action which will understand how the
[01:00:11] system work as one.
[01:00:18] Can you get me the information from the
[01:00:20] back of the truck?
[01:00:21] >> I've got the information you have
[01:00:23] requested. What would you like to do
[01:00:25] next?
[01:00:26] >> What time do they close today?
[01:00:28] >> Coastal Flowers closes at 5:00 p.m.
[01:00:30] Monday through Saturday and at 2:00 p.m.
[01:00:32] on Sundays.
[01:00:34] >> What kind of flowers do they sell?
[01:00:36] >> Coastal Flowers specializes in nature
[01:00:38] inspired designs using seasonal blooms.
[01:00:43] So this is this is very interesting
[01:00:45] because what you can see the cameras of
[01:00:48] the car the radars and all the sensors
[01:00:51] they're being on this layer of
[01:00:52] intelligence to drive the car now that
[01:00:54] also input and context for the agent for
[01:00:58] your experience and we're just the
[01:01:00] beginning of this and what we actually
[01:01:02] see within Qualcomm this whole
[01:01:04] transition of the softwaredefined
[01:01:07] vehicle is now augmented when the
[01:01:09] vehicle evolving to the AI AI defined
[01:01:12] vehicle. Now I talk about personal
[01:01:16] computing cars. I want to talk about the
[01:01:20] next big category of physical AI which
[01:01:24] is robotics.
[01:01:27] The entire robotics industry right now
[01:01:29] is being transformed and robotics is
[01:01:32] very interesting because it pushes the
[01:01:34] limit of computing technology and
[01:01:37] incorporate aspects of what I show you
[01:01:39] to you. Incorporate many aspects and
[01:01:42] technology of what we see in in personal
[01:01:46] computing for sensing perception
[01:01:50] high integration of sensors. The robot
[01:01:53] like your glasses will see what you see,
[01:01:56] read what you read, understand the
[01:01:58] context. At the same time, you have all
[01:02:00] of those elements of technology
[01:02:03] that comes from the automotive industry.
[01:02:06] You need precision. You need industrial
[01:02:09] grade. You need you need safety, you
[01:02:11] need redundancy. So that is exciting
[01:02:15] exciting element of technology
[01:02:17] transition which has the best of both
[01:02:20] worlds. And to succeed in robotics, you
[01:02:24] need to understand how to design this as
[01:02:27] a hierarchical compute system.
[01:02:30] First of all, you need to understand
[01:02:32] that a robot has three layers of
[01:02:35] computing. There's instant execution.
[01:02:38] someone like us, you're going to pick up
[01:02:41] something and you haven't didn't get a
[01:02:43] good grip. Uh you don't think it, you
[01:02:45] just go again and you get it again.
[01:02:47] That's instant execution. Same thing how
[01:02:50] you keep your balance or everything that
[01:02:52] you need to do without thinking like us.
[01:02:55] Then there's action and then there's
[01:02:58] action and grounding and then you have
[01:03:00] reasoning. You have to build a
[01:03:02] hierarchical compute system. You also
[01:03:04] have to distribute intelligence. It's
[01:03:06] not just about building the brain of the
[01:03:08] robot. You need to have a central
[01:03:10] computer. You have to have motion
[01:03:13] control. You have to have the ability to
[01:03:17] do actuation things that are going to be
[01:03:19] very important which is perception,
[01:03:22] dexterity. So bring the best of
[01:03:27] mobility, personal computing, uh what
[01:03:30] you see in some of the automotive and
[01:03:32] navigation. It is a great technology
[01:03:35] platform.
[01:03:36] What we're building is a comprehensive
[01:03:38] platform across different form factors
[01:03:41] from AMR, industrial arms, four-legged
[01:03:43] robots, humanoids, drones, and I
[01:03:46] actually think industrial opportunity is
[01:03:48] happening right now. You have to build
[01:03:51] hardware, you have to build software,
[01:03:52] you have to build an AI operations,
[01:03:55] ability to do fleet and data management,
[01:03:57] uh, and everything that our partners
[01:03:59] will need to go from prototype to
[01:04:01] production. And the robot is another
[01:04:06] great example when battery life matters,
[01:04:10] when highlevel integration of of com
[01:04:13] computing and sensors. If you want the
[01:04:16] robot to get scale and the right price
[01:04:19] points, you need to bring from consumer
[01:04:20] electronics the capabilities to do high
[01:04:23] level integration and that's an exciting
[01:04:25] opportunity for us and our partners as
[01:04:27] well.
[01:04:29] And when you think about it, uh this is
[01:04:32] actually developing very fast. Those are
[01:04:34] two examples. We've been working with
[01:04:36] companies like Vinmotion, Neura Robotics
[01:04:39] and Stogard. And soon we'll share what
[01:04:43] we're doing with figure AI as well.
[01:04:46] But beyond robotics, physical AI has an
[01:04:50] incredible opportunity
[01:04:53] in industrial.
[01:04:55] I've been saying this industrial is not
[01:04:59] demand bound is solution bound. It's
[01:05:02] about maturity of those system and
[01:05:04] technology. The demand is enormous in
[01:05:08] enterprise and industrial environments.
[01:05:10] For example, vision AI powered cameras
[01:05:14] can monitor security compliance. They
[01:05:17] can monitor traffic flow on a smart
[01:05:19] city. They trigger actions, autonomies,
[01:05:21] and there's many more applications. If
[01:05:24] you want to understand what's also going
[01:05:25] to happen in the enterprise, you're
[01:05:26] going to see new form factors of things
[01:05:28] that an enterprise uh you know uses
[01:05:31] every day from helmets to safety glass
[01:05:34] to badges to everything. And the
[01:05:37] opportunity is really incredible when
[01:05:39] you think about for example the computer
[01:05:42] vision. You can see I think the role of
[01:05:45] AI and Agent AI changing everything.
[01:05:50] Computer vision can run locally on
[01:05:51] devices. They can run on the cloud. They
[01:05:54] dynamically can access more powerful
[01:05:56] models on prem or on the cloud when
[01:05:59] needed. They'll be able to see,
[01:06:02] understand, decide, take action, turning
[01:06:07] perception into real time uh
[01:06:10] intelligence. And you can see that the
[01:06:12] applications are are truly incredible.
[01:06:15] And we're just at the beginning of this.
[01:06:18] the the input from vision uh and other
[01:06:23] sensors connected with Agentic AI will
[01:06:26] fundamentally change how many of those
[01:06:30] systems and enterprises are going to
[01:06:32] operate and we see this happening for
[01:06:34] example at retail at warehousing
[01:06:38] in general I think office management and
[01:06:42] building we see this in the oil and gas
[01:06:44] industry we see this with energy and of
[01:06:46] course we can see this in smart cities.
[01:06:49] This is an incredible opportunity uh uh
[01:06:52] for a Gent AI.
[01:06:56] Now you have seen a Gent AI transforming
[01:07:01] personal and computing and those new
[01:07:04] classes of personal devices you see
[01:07:07] cars, robots, industrial. But there's
[01:07:10] another layer coming to all of this that
[01:07:13] will change the equation. And I'm
[01:07:15] personally very excited about this.
[01:07:18] is the next generation of wireless and I
[01:07:21] am going to explain to you today why
[01:07:24] this is actually very relevant to the
[01:07:26] conversation we just had. At Mobile
[01:07:28] Congress would describe 6G as the first
[01:07:31] generation of wireless design for the
[01:07:33] age of AI and 6G has three pillars.
[01:07:37] connectivity is distributed computing
[01:07:41] and is sensing
[01:07:44] and distributed computing and sensing
[01:07:47] are new things that didn't existed
[01:07:49] before in the telecom sector. I'm going
[01:07:52] to break each of those pillars for you.
[01:07:55] So, let me start with connectivity.
[01:07:58] My simple description to you of the
[01:08:00] connectivity change in 6G is based on
[01:08:04] what we just talked about at the
[01:08:06] beginning of this presentation. If you
[01:08:08] have smart glasses, the see what you
[01:08:11] see. So the the connectivity needs to
[01:08:15] enable a very fast uplink.
[01:08:20] So,
[01:08:21] I know it's going to sound uh
[01:08:23] interesting what I'm going to say next,
[01:08:25] but 6G is going to make all of us into
[01:08:28] walking cameras in this world because
[01:08:30] you're going to have the ability to get
[01:08:33] high definition video that you see and
[01:08:36] be able to send that to the system into
[01:08:38] the cloud. It is designed for continuous
[01:08:42] contest exchange between devices, mach
[01:08:45] machines, agents, car and with an a
[01:08:48] native air interface, the network
[01:08:50] understands and adapted radio
[01:08:51] environment. You're going to see I won't
[01:08:53] spend a lot of time talking about it but
[01:08:55] you see a lot of different techniques
[01:08:57] autonomous networks prediction of radio
[01:09:00] signals so you can have high speed and
[01:09:02] wake signal conditions significant
[01:09:05] improvements in speed in the density but
[01:09:10] bottom line is ability to connect
[01:09:12] everything and enable even that use case
[01:09:17] of see what I see that's the
[01:09:19] connectivity part the second part is
[01:09:22] computing and that's going to become
[01:09:25] clear when I get to the pillar number
[01:09:27] three which is sensing but that is the
[01:09:30] biggest change in the infrastructure of
[01:09:33] telecommunications in wireless since the
[01:09:36] beginning of mobility the whole network
[01:09:39] in itself as an AI network with
[01:09:41] distributing AI computing and inference
[01:09:44] at the radiobased stations to the
[01:09:46] central office and the data center the
[01:09:51] network. It's part of a very large data
[01:09:54] center. The central office looks like an
[01:09:57] on-prem data center and then you have
[01:10:00] computing in real time at the radiobased
[01:10:04] stations
[01:10:06] and that is going to be important not
[01:10:07] only for the connectivity aspects I just
[01:10:09] spo spoke about autonomous network
[01:10:12] autonomous allocation of resources uh
[01:10:15] prediction of of of u RF signals but
[01:10:20] it's because of sensing that you need
[01:10:22] this capability and sensing is the
[01:10:26] biggest change in this sector. That's
[01:10:28] what's going to make the telecom telecom
[01:10:31] sector really a completely different
[01:10:34] company. What sensing will do is we'll
[01:10:38] look at the RF signals as physical AI
[01:10:42] input for models that are being trained
[01:10:44] on their performance. what it's going to
[01:10:47] do. Each one of those radio connections
[01:10:49] are going to be like a radar that you
[01:10:51] see used on a car for autonomous driving
[01:10:54] and you're going to get hundreds of
[01:10:56] millions of those connections uh in real
[01:10:59] time. You triangulate all of this data
[01:11:01] and what you do it will create a digital
[01:11:04] twin not only of the neighborhood of the
[01:11:06] city but the whole country. And what
[01:11:09] you're going to do with this
[01:11:11] drone detection is an example. You
[01:11:14] detect everything that moves and flies.
[01:11:18] You manage the entire lowaltitude aerial
[01:11:22] economy. You're going to detect on every
[01:11:24] road, every car, every bicycle, every
[01:11:27] truck, every pedestrian. You can
[01:11:29] actually identify those objects and
[01:11:32] annotate on that image. You bring this
[01:11:34] whole conversation we just had about
[01:11:36] computer and vision intelligence as an
[01:11:38] example into a large digital twin. And
[01:11:41] that is going to change everything. And
[01:11:43] what's the connection with this
[01:11:45] conversation we have of Agentic AI? This
[01:11:48] is all realtime context for agents that
[01:11:52] is going to be running on the end points
[01:11:56] uh across all of those devices.
[01:11:59] This an exciting I think future for
[01:12:03] everything that is happening on the edge
[01:12:06] because AI is the new form of computing
[01:12:08] and is going to be how computers are
[01:12:10] going to run everywhere. So now I'm
[01:12:13] going to get to the last part of my
[01:12:15] presentation
[01:12:17] and this is the second part of what I
[01:12:19] said it was my objective today to
[01:12:22] explain to all of you how the devices at
[01:12:24] the edge is going to change of AI. The
[01:12:27] second part is trying to get a very very
[01:12:31] simple understanding of how cloud and
[01:12:35] edge are going to work together.
[01:12:38] It's now very obvious but I want you to
[01:12:41] I want you to get the same conclusion
[01:12:43] that we got how this word is going to
[01:12:46] evolve. So agents are now how tokens
[01:12:53] how demand for tokens is created and
[01:12:56] it's how AI is going to get a scale and
[01:12:59] is defining both the architecture and
[01:13:01] the economics of AI. So let's go back
[01:13:04] and recap.
[01:13:07] Today most of the software, the OS, the
[01:13:11] app store, they have been designed for a
[01:13:15] human to operate. So we operate and we
[01:13:18] do things. We we start an app or a
[01:13:21] program or go to the web. We do
[01:13:24] different things. It's been designed for
[01:13:26] us.
[01:13:27] Now with agents, everything changes.
[01:13:31] The agents are autonomous. They interact
[01:13:34] with software much faster. They do it
[01:13:37] across multiple services at once. That's
[01:13:40] why the device is going to have to be
[01:13:41] upgraded. And every workflow generate
[01:13:45] tokens at machine speed, not human
[01:13:48] speed.
[01:13:49] Just as an example,
[01:13:51] when you think about an application or
[01:13:55] even the SAS industry, you think about
[01:13:58] software for the SAS industry. Those
[01:14:00] clients to date they're being designed
[01:14:03] for a human to operate the client. Not
[01:14:06] anymore. You have to design your
[01:14:08] application and your SAS for a human and
[01:14:10] an agent to interact and the agents are
[01:14:12] going to interact faster.
[01:14:16] That is going to fundamentally change
[01:14:19] the demand for tokens. So as an example,
[01:14:23] you can see those transitions right um
[01:14:26] how quickly token consumption is
[01:14:28] growing.
[01:14:30] So lot of different estimates but it
[01:14:32] current looks like this. So level one
[01:14:35] when we started conversational
[01:14:39] you go in and you a single turn you
[01:14:41] prompt you get a response that's about
[01:14:44] you can talk about 10,000 tokens uh on a
[01:14:47] single turn prompt response for level
[01:14:50] one conversational
[01:14:51] level two reasoning. Then you now have
[01:14:54] multi-turn
[01:14:56] uh and you can see an order of magnitude
[01:14:59] increase about 100,000 tokens per task.
[01:15:03] And then when you think about level
[01:15:04] three a gentech AI autonomous multistep
[01:15:10] multiple to toe calls it's about 1
[01:15:13] million to tokens per task and is
[01:15:15] growing. you see the order of magnitude
[01:15:18] increase. That's about 100x increase in
[01:15:22] the number of tokens in two generations.
[01:15:25] Now let me give you the same data
[01:15:27] different perspective. So this year 2026
[01:15:33] it's estimated that the global token
[01:15:35] demand within 10 seconds is about 31.7
[01:15:40] billion tokens.
[01:15:42] So the projection
[01:15:46] uh for 2030 for the same 10 seconds is
[01:15:50] 1.27 trillion tokens. It's a 40x
[01:15:54] increase.
[01:15:56] And that's because agents are going to
[01:15:58] generate a lot of tokens. So if you
[01:16:01] think about the latest estimate, and I'm
[01:16:04] sure that's going to change, but the
[01:16:06] latest estimate for total token demand
[01:16:09] globally in 2030 is in the quintilians.
[01:16:12] I don't even know how to say the number.
[01:16:13] I'll say is a four gazillion uh tokens.
[01:16:18] Uh the the token demand for 2030.
[01:16:22] And that is the new currency of AI.
[01:16:26] Tokens are the currency of AI. And let
[01:16:28] now share you how this is going to work.
[01:16:31] And there's going to be a lot of people
[01:16:32] there's going to argue there's no need
[01:16:34] to argue this is exactly how it's going
[01:16:36] to work. Uh it's it's we have seen this
[01:16:40] before. I resistance is futile. Okay. So
[01:16:44] let me give you some example how this is
[01:16:46] going to work. So let's talk about
[01:16:48] coding. Coding workloads are some of the
[01:16:50] most token in uh intensive tasks today.
[01:16:55] Using a real session running cloud code,
[01:16:59] an orchestrator is now intelligent
[01:17:02] routing the workload, keeping certain
[01:17:05] tasks on the compute available on the
[01:17:07] device and sending what's necessary to
[01:17:08] the cloud. When you apply the
[01:17:11] distributed Agent AI, making use of all
[01:17:14] the compute across this compute
[01:17:17] continuum, you can see you save about
[01:17:20] 1.4 million tokens, 60% lower cost for
[01:17:23] the same result. Now that's coding. Let
[01:17:25] me give you a completely different
[01:17:27] example.
[01:17:28] Um
[01:17:30] so the next demo I'm going to show you
[01:17:35] is a task create a Snapdragon web page.
[01:17:40] Uh there's a lot of specifications on a
[01:17:43] prompt about this web page needs to be
[01:17:45] created. On the left you can see that it
[01:17:48] runs on the cloud uh and and generates
[01:17:52] tokens in the cloud. On the right is the
[01:17:55] same result but you have intelligent
[01:17:58] distributed intelligent orchestrator
[01:18:01] uh defines this routing of the task
[01:18:05] about the compute that is available on
[01:18:07] the device and the computer available on
[01:18:09] the cloud. result is you get exactly the
[01:18:11] same result 30% fewer tokens 4x lower
[01:18:16] cost and that is the power of
[01:18:20] distributed AI and this is how this is
[01:18:22] going to play out I like to provide this
[01:18:25] example um it's interesting because
[01:18:27] we've been seeing this conversation
[01:18:29] about cloud and edge and they're both
[01:18:31] incredibly important and now you've
[01:18:34] started to see how this is actually
[01:18:36] going to play out it's not an abstract
[01:18:38] thing exactly
[01:18:40] that the conversation is never about
[01:18:42] this thing that runs on the cloud can
[01:18:44] also run on the edge. No, what needs to
[01:18:47] run on the cloud is going to run on the
[01:18:48] cloud. When needs to run on the edge is
[01:18:49] going to run on the edge. It's
[01:18:50] different. I like to do this analogy.
[01:18:53] It's not perfect, but it's good for us
[01:18:55] to understand it. I'm sure each and
[01:18:57] every one of you here has probably over
[01:18:59] 200 apps organized in folders in your
[01:19:02] smartphone. And if I ask you anybody to
[01:19:04] go to the exercise of say for each app
[01:19:07] what runs on the computer on your
[01:19:09] smartphone what runs on the computer in
[01:19:11] the cloud it's an exercise in futility
[01:19:13] and and you know you know the importance
[01:19:16] of the cloud because if you put your
[01:19:17] phone in airplane mode it's useless but
[01:19:21] you need a lot of computing and this is
[01:19:23] how it's going to work. AI agent AI is
[01:19:27] agents running the computer for you and
[01:19:30] as the workloads started to shift the
[01:19:33] things that you do yourself and the
[01:19:35] thing that the agents going to do is
[01:19:37] going to make use of all the computing
[01:19:40] is available and it's going to go
[01:19:41] exactly how works today the computing uh
[01:19:44] is going to be utilized everything will
[01:19:46] be running AI and it's going to happen
[01:19:49] across the compute continue and I think
[01:19:51] that's the exciting part for Qualcomm
[01:19:55] because that's going to create demand
[01:19:57] for AI computing everywhere and and
[01:20:01] every computing engine becomes
[01:20:04] interesting and and necessary for this
[01:20:08] agentic future and you're going to
[01:20:10] naturally process the workloads where
[01:20:13] it's more efficient and and that is the
[01:20:17] opportunity because we're uniquely
[01:20:19] positioned in across all of those
[01:20:22] different devices and it's never one
[01:20:24] sizefits-all. You need to have the right
[01:20:27] platform for AI for every device. As we
[01:20:30] talk about it, they are different. They
[01:20:33] have uh different purpose. It's about
[01:20:35] maximum intelligence and maximum
[01:20:37] efficiency everywhere. So, as I get to
[01:20:40] the end of the presentation, uh I have
[01:20:43] one last thing uh that I want to tell
[01:20:46] you. So, let's watch the video.
[01:21:13] So today at Computex, we're announcing
[01:21:17] the new product brand for Qualcomm data
[01:21:21] center products. Um,
[01:21:24] we're already working with hyperscalers
[01:21:26] and global partners on real world
[01:21:29] deployments. We're incredibly excited of
[01:21:31] this new chapter of the Qualcomm
[01:21:33] diversification.
[01:21:34] And now with Dragonfly, our portfolio
[01:21:38] spans every single tier of the compute
[01:21:41] continuum from the smallest wearables
[01:21:44] that will connect to agents to data
[01:21:47] centers at a very high performance.
[01:21:50] Uh we'll have a lot to share on our
[01:21:53] investor day in just a few weeks uh June
[01:21:57] 24. I wish I can tell you all about our
[01:21:59] road map but that is coming.
[01:22:02] And in summary, I'm just going to tell
[01:22:04] you uh we're incredibly excited about
[01:22:08] the future of technology
[01:22:11] across every single device. Uh it's this
[01:22:15] is going to get transformed.
[01:22:17] We show the agents are not coming into
[01:22:21] the future. They're already here. It's
[01:22:23] changing a lot of the compute. It's
[01:22:25] going to generate a lot of demand for
[01:22:27] new classes of devices and computing.
[01:22:31] And this upgrade cycle can be one of the
[01:22:33] largest that the industry has sent and
[01:22:37] together we're building the products and
[01:22:39] the technology that will define this new
[01:22:41] era. Thank you so much for listening to
[01:22:44] my presentation. Like I said, is a
[01:22:47] pleasure and an honor to be here.
[01:22:49] Hopefully, this was helpful for you and
[01:22:52] I can't wait to build this future
[01:22:53] together with our partners across the
[01:22:56] globe and our partners in Taiwan. Thank
[01:22:58] you very much. Enjoy compete tax.