Amazon Cto Werner Vogels’ Predictions For 2026

AWS re:Invent is now an yearly contented that goes backmost to 2012. One cardinal portion of nan arena is Amazon CTO Werner Vogels delivering nan closing keynote, which successful caller years has focused little connected releasing caller products and alternatively focused connected nan bigger image and sharing lessons from nan unreality computing trenches.

Another contented is that Vogels besides releases a group of predictions for nan coming twelvemonth astatine nan arena — and for nan past fewer years, I’ve sat down pinch him astatine re:Invent to talk those pinch him. But this twelvemonth is simply a spot different: Vogels is releasing nan predictions earlier nan arena and successful this year’s conversation, he hinted that while he’s not retiring, this whitethorn beryllium his past re:Invent keynote.

“This is benignant of nan past time,” he told maine erstwhile I asked him why he switched things up this year. “It’s clip for nan voices to travel up. There’s truthful galore superb engineers pinch truthful galore experiences that I deliberation — you know, I’m not leaving Amazon pinch thing for illustration that. I’m not retiring aliases anything. I’ve tons of awesome ideas that I would for illustration to pursue, but it is clip for Amazon customers to perceive immoderate younger voices that person tons of acquisition of building AWS.”

Calling successful from Dubai, Vogels noted that he’s been astatine Amazon for 21 years now, seeing nan institution spell from trading books to becoming nan elephantine it is today. But he besides noted that erstwhile he first came to Seattle, he was still an world and only really went to springiness a talk astatine nan company.

“Amazon, it’s a bookshop. It’s a web server and a database. How difficult tin that be?,” Vogels told maine astir his first thoughts astir nan institution astatine nan time. “Until I really did spell — and 1 glimpse into nan Amazon exertion room blew maine away. These guys were operating astatine a standard that cipher other successful nan world was, and conscionable helping them go much reliable, get amended power complete their systems, possibly moving azygous points of nonaccomplishment — truthful galore things that that were fun, that really still support maine keeping connected going today.”

Vogels’ Predictions

As for his predictions this year, it’s possibly nary astonishment that they halfway astir AI and its effect connected developers, but besides nine arsenic a whole. In total, location are 5 predictions here, though Vogels told maine that he primitively had acold much connected his list. The ones that made it into nan database scope from utilizing robots to combating loneliness to quantum-safe encryption and nan progressively accelerated pipeline from subject tech to user products.

Below are excerpts from our conversations astir each 1 of them. The transcript was lightly edited for readability.

1. Loneliness: Companionship Is Redefined for Those Who Need It Most

Werner: The first prediction, and I particularly put it up front, is nan 1 astir loneliness. And if you person tons of friends and tons of colleagues and whatever, you astir apt don’t moreover cognize what loneliness is. But for a very ample group of people, whether they’re aged aliases younger, it’s a large problem for them. It really triggers diseases. It makes group unhappy, depression, each these benignant of things. Now we tin build exertion to thief immoderate of these group already, getting them retired of their depression.

I remember, actually, almost astatine nan opening of Alexa, I heard a communicative of a man who was astatine nan opening of dementia, and he told maine — I really said to him — he told maine that he could spot these caregivers getting irritated because he asked nan aforesaid point complete and complete and complete again. And past he sewage an Alexa, and he put a sticker connected it because he didn’t want to hide what she was called. But he said, “I tin inquire her thing I want. She ne'er gets angry.”

And truthful that was not needfully why we built Alexa, but it’s a broadside effect of exertion that helps people. And particularly wherever we tin thief group flooded slump aliases are capable to pass amended pinch their caregivers—if we tin build devices for that, absolutely. And it’s astonishing that really a batch of group dainty their devices arsenic pets. They’re fixed names. There’s nan story, of course, astir Roomba from iRobot. About 80% of nan Roombas person a name. They’ve been fixed a name. And location was 1 personification that really sent nan Roomba backmost because it was broken, and nan institution said, “We’ll nonstop you a caller one.” And he said, “No, no, no, no. Meryl Sweep needs to travel back.”

2. The Dawn of nan Renaissance Developer

Werner: I deliberation we request to germinate conscionable arsenic overmuch arsenic our devices evolve. I learned 68000, COBOL, and Pascal successful school. Nobody programs that anymore. And aft that we sewage C, we sewage C++, and we sewage Java, and we person sewage Python, and location will beryllium 10 more. Now everybody is guessing complete Rust, rightfully so, I think. But you know, there’s thing other that will fig out, thing else, different problem to solve.

This has ever been nan case. We ever learned caller technologies arsenic developers. And you cognize what? Often, we were up of our leader asking us. “Oh, I’ve heard astir this COBOL thing. I already publication nan book. I tin build thing for you.” COBOL programmers these days get paid well.

[Nvidia CEO] Jensen [Huang], really — I saw an question and reply pinch him wherever personification asked him: “Will AI yet overtake nan capacity of your brain?” And Jensen said, “You cognize what? Maybe there’s 20% of your activity that AI tin do for you, aliases possibly 50%. But that’s it.” There are things successful our creativity, things successful our curiosity that these machines don’t have.

And that’s why I’m bringing it backmost to nan Renaissance, because nan Renaissance conscionable came retired of 1,000 years of Dark Ages. And earlier those 1,000 years, Egyptians and nan Romans and nan Arabs had done tremendous activity connected mathematics and things for illustration that. And abruptly each of this came backmost again. And arsenic such, everybody needed to study caller tasks. It was nary longer capable to beryllium conscionable a carpenter.

And that’s really I spot nan improvement of engineers happening arsenic well. There was a clip erstwhile it was capable to beryllium a back-end developer. You knew precisely what you needed to do. But now you request to person a broader view. Not conscionable different languages—you request to cognize really nan business is being run.

So I telephone this a T-model, wherever you spell deep, but you besides request to person this polymath benignant of thing.

3. Quantum-Safe Becomes nan Only Safe

Werner: I had a prediction, I think, 2 aliases 3 years agone astir wherever we were [with quantum computing], and I deliberation my prediction was 10 years retired — and it’s nary longer 10 years out. Four to 5 years.

Frederic: The aged joke is that it’s ever been 5 years out.

Werner: Yeah, I said 10, truthful I wasn’t really joking. I deliberation we tin get them to a reliability level wrong 5 years. The problem is not nan quantum machinery. That’s not nan problem. The problem is that location is information harvesting happening already. There’s information harvesting now astir your aesculapian information, your financial information, that cipher tin publication because it’s encrypted. But 5 years from now, it will be.

I deliberation arsenic large companies, we should really push our customers to make judge that they usage nan correct technologies. We built thing — primitively was called S2N, Signal to Noise. SSL, I think, has half a cardinal lines of code. And truthful we thought, that’s because there’s each these different variations. We thought it’s not necessary. There’s only 1 aliases 2 that you use. So we named it S2N TLS, I believe. And we person nan expertise to already put post-quantum encryption successful there. So that’s 1 thing.

The bigger problem is your car shed door. Your car shed doorway runs a small instrumentality that runs Linux that you haven’t updated successful astir apt nan past 2 years. Or each sorts of different devices — return edifice cards.

And this is not a joke for illustration nan twelvemonth 2000, “the world will spell up successful flames” — well, thing happened. But fixed that this is thing wherever location are capable authorities actors willing in, location are capable criminal actors willing in, we will request to return action now if we don’t want to beryllium successful large problems successful 5 years.

It is arsenic large arsenic — it whitethorn moreover beryllium bigger than nan extortion techniques that you spot now pinch your information being held hostage and things for illustration that. This is going to beryllium a overmuch bigger problem. And arsenic large companies, we should make judge that our customers understand this. Almost each speech pinch a customer these days, I extremity with, “Have you thought astir encryption?” “Yes.” “Start thinking. And commencement reasoning really you’re going to reside this. Is it conscionable pinch a firewall? Are you going to re-encrypt each your data?”

There’s tons of solutions. But I do deliberation my customers are proceeding me, and that’s a bully thing.

4. Defense Technology Changes nan World

Werner: I thought, fto maine talk astir thing that I’m usually not very affirmative about, and that is war. Two days ago, I was present successful Dubai astatine nan aerial show, and I thought that’s a awesome infinitesimal to commencement asking group really overmuch of subject exertion makes it really into civilian technology. Almost everything isolated from nan exploding parts makes it into civilian technology.

And you cognize what? Without DARPA, we wouldn’t person had nan internet. We wouldn’t person GPS. There’s truthful galore subject technologies that, complete time, person made it into civilian technology. Now, if I comparison nan magnitude of money being spent successful defense invention and comparison that pinch civilian — there’s a batch much money going into nan subject business complex. So location are innovations happening location wherever we arsenic civilian engineers tin really use from complete time.

Frederic: Do you spot immoderate caller examples of that arsenic well?

Werner: I met this institution that was focused connected transporting hearts aliases livers from 1 infirmary to different hospital, and nan drone exertion that they were utilizing is subject drone technology. And I worked pinch a institution called Help.NGO, who activity connected large disasters and things for illustration that, and they person a batch of erstwhile subject technology. They’re a Polish company. They person a batch of erstwhile [military] exertion that they are utilizing connected a regular ground successful nan large disaster areas.

Are location antagonistic stories? Yeah, and I dislike those arsenic a technologist. I mean, if you perceive — I conscionable saw an question and reply connected nan BBC pinch a expert who spent 2 years successful Gaza. And he said what happens is they will explosive a shelter campy and past very mini drones will travel aft that and sprout pellets astatine small kids. And they won’t die, but they’ll suffer a leg, aliases they’ll suffer an arm. And if I deliberation astir nan truth that immoderate of my exertion whitethorn beryllium utilized successful that, it scares me.

Frederic: And location person been a batch of arguments astir nan large exertion companies [and their engagement pinch authorities organizations].

Werner: Yeah. I mean, I besides — we person a batch of customers successful Ukraine, and if I perceive really they are fundamentally building exertion from nan crushed up, arsenic inexpensive arsenic imaginable for defense, that is thing that I deliberation — yes, I don’t for illustration war, but if I request to beryllium connected a bully side, past I would emotion to beryllium connected nan bully side.

5. Personalized Learning Meets Infinite Curiosity

Werner: I ever was a large instrumentality of Ken Robinson. Ken Robinson was a erstwhile assemblage professor successful productivity who looked astatine productivity successful kids and saw really almost each kid was imaginative —

Interviewer: And we killed it.

Werner: Until they sewage to school. We sent them to school, and past they each became conformists. They each had to study nan aforesaid thing. There was nary individual school that would beryllium really for you. And I honestly judge that, you know, what we spot present successful ADHD, for example, could person been avoided by overmuch amended targeted learning systems for nan individual.

I’m not keen connected giving a kid an iPad and having them really study from that. But if it intends that I tin get you precisely what motivates you, that’s great, right? Curiosity is what we termination successful kids. And curiosity drives you to learning. I deliberation we request to make judge that we support kids curious, that we don’t termination that, because curiosity is nan precondition to learning. And you study pinch tons much nosy if you’re curious.

If we tin proceed to do that — but it’s difficult to person a people pinch 40 kids and support them each curious. That’s very hard. You know, looking astatine individuals.

And we’ll decorativeness this disconnected pinch a small joke that Ken Robinson really told. He was a teacher, and location was 1 kid successful nan class, 1 girl, who really ne'er really paid attraction until drafting class. And past he walked up to her and looked astatine her, and he says, “What are you doing?” She says, “I’m drafting God.” He says, “But cipher knows really God looks like.” And she says, “In a fewer minutes they will.”

Frederic: As you were saying that, I was thinking: Is location a mediate crushed location betwixt us, supporting nan teachers pinch nan exertion without overwhelming nan students? There are a batch of arguments that screens successful classrooms are not nan champion thing, right?

Werner: It’s 1 of Robinson’s points arsenic well. He says location are nary bully schools, but location are bully teachers. And what tin we do to thief teachers beryllium much individualized and person much clip to spend?

I mean, look astatine — I don’t cognize really galore unfastened positions for teachers we have. And astatine nan aforesaid time, arsenic governments, we’re not spending terribly overmuch money connected education, are we? And this is nan astir important portion successful a kid’s life, and we’re not spending nan money connected it. We’re spending it connected subject planes.

Do you cognize Ted Lasso, nan TV show? There’s this segment wherever he plays darts against nan proprietor of nan club. And he tells his story. He said, “At 1 moment, I drove past a wall. It says, ‘Be curious.'” And he said, “Now I recognize why each these kids were bullying me. They were ne'er funny astir who I was. They were conscionable bullying me.”

And curiosity — let’s support curiosity alive. All nan time. All nan measurement till you’re sitting location connected your sofa pinch your mechanical feline and stroking it. And if you’re still curious, you’ll person a awesome life.

Group Created pinch Sketch.