Software Isn't Dead, Aluminum Is Fuel, and AI Wants Your Spreadsheet Job

LaurenOkay, so get this. My feed is just wall to wall with people declaring software dead. Like dead, dead. Coffin, eulogy, the whole thing.

DerekEvery few months, somebody discovers AI can write code, and suddenly nobody needs an app ever again.

LaurenRight. Plot twist, though, I don't think that's what's actually happening.

DerekOh, you're already picking a fight with the Internet. I love it.

LaurenI'm Lauren, and this is Tech Insider Weekly, where my co-host, Derek, and I try to figure out what's real before Twitter finishes arguing about it.

DerekAnd there's a name in the software world that's making this whole SaaS-is-dead thing way messier than the doomers want it to be.

LaurenOoh, wait for it. Is this the Linear thing?

DerekYou'll see. Numbers are involved, and they're not the numbers people expect.

LaurenLove that for us. Okay, but that's not even the wildest thing on deck today.

DerekNo, because right after that, we're going full physical world: aluminum, power grids, the actual electricity bill behind all this AI hype.

LaurenWhich, dude, nobody talks about enough. Building the brain is one thing. Feeding it is another.

DerekAnd then get this, we're going somewhere a little more personal. What all this automation actually means if you're twenty-two and hoping to get hired into finance.

LaurenYeah, that one's not just a fun stat segment.

DerekNo, it's not.

LaurenBut we'll close it out, zoomed all the way out, the money pouring into robots and physical AI and just how fast that's moving.

DerekFour stories, one thread.

LaurenSo let's not waste the cold open arguing about vibes. Derek, you said Linear has a number that breaks the whole software's over theory.

DerekI did say that.

LaurenThen let's see if it actually holds up. Okay, so get this. Linear, the project management tool half the startup world lives in, just wrapped an employee tender offer.

DerekAnd?

LaurenTwo and a half billion dollar valuation. Business Insider ran the headline practically laughing at itself. Everyone's saying software is dead, and this SaaS company just doubled its price tag.

DerekWait, doubled?

LaurenDoubled.

DerekOkay, that's not a company limping along. That's a company printing money while the obituary's being written.

LaurenRight. So how does a tool for writing tickets and tracking sprints get more valuable in the year everyone decided AI writes the software itself?

DerekBecause it's not really a ticket tool anymore. There's a recap floating around of an interview with Linear CEO, Saarinen, where he says the company went with their calling agent native.

LaurenMeaning?

DerekMeaning agents aren't guests in the product. They're first class users. Same login, same permissions humans get. And he's naming names. OpenAI and Coinbase are apparently running agents directly inside Linear to manage their own workflows.

LaurenSo the AI isn't replacing the software, it's just clocking in.

DerekBasically, it moved into the building instead of burning it down.

LaurenOkay, but is that just a great story from one founder, or does it actually hold up when you zoom out?

DerekThat's the fair question because it's not all sunshine. There was a real bloodbath earlier this year.

LaurenHow real?

DerekA Forbes contributor, Lutz Finger, put a number on it. Roughly two hundred and eighty-five billion dollars wiped off SaaS valuations in a February sell-off. People started calling it the SaaSpocalypse.

LaurenTwo hundred and eighty-five billion? That's not a correction, that's a gut punch.

DerekIt is. But his actual argument isn't that SaaS is finished, it's that the interface is what's changing. He points to Intuit, HubSpot, Salesforce, and Shopify all rolling out something called MCP. Basically a standard way for AI agents to plug straight into their software instead of a human clicking through menus.

LaurenSo the software doesn't disappear. It just stops caring whether a person or a bot is driving it.

DerekExactly the Linear story, honestly. Same move, different logo.

LaurenOkay, but I want to sit in the uncomfortable chair for a second. Isn't Linear just the pretty exception? Dev tools, beloved by engineers, tiny team, easy to retrofit for agents doesn't mean Workday is fine.

DerekSure. And that's the real fight underneath this. Is Linear proof the whole death of software thing was lazy shorthand? Or is it the one company sharp enough to dodge a bullet that's still coming for everybody else?

LaurenI lean toward lazy shorthand. Software isn't dying, it's just getting a new kind of user, and the companies figuring that out first are the ones getting the two point five billion headlines.

DerekI'll meet you halfway. I think it's true for infrastructure heavy tools like this. I'm way less sure it's true for the fifty-seat internal app nobody's rebuilding around agents.

LaurenFair. Different question for a different company.

DerekAnd speaking of infrastructure, all of this, every agent logging into Linear, every model Coinbase is running, none of it happens without power.

LaurenOh, here we go.

DerekBecause the valuation story is fun, but there's a much bigger, much dirtier problem underneath the whole AI boom, and it's not code. It's electricity.

LaurenPlot twist: the sexiest software story of the year runs on the least sexy thing on Earth.

DerekYep. Let's talk about what's actually keeping the lights on.

LaurenOkay, so power, because none of this software fight matters if there's no juice to run the servers.

DerekWait, are we about to talk about burning metal? We are about to talk about burning metal.

LaurenI love this podcast

DerekMIT Technology Review had a piece dated October 23rd on a Boston startup called Found Energy, and they're apparently about to run their biggest real-world test yet of turning aluminum scrap into a zero carbon fuel.

LaurenHold on. Aluminum, like soda cans?

DerekLike soda cans, like scrap from cars, whatever's lying around. You react it, and it puts out heat without carbon.

LaurenThat sounds like something a guy would explain to you at a party, and you'd nod and then Google it later.

DerekFair. But the outside analyst quoted in that same report was careful. They said this really functions as energy storage, not a primary energy source. So it's not a power plant replacing a power plant.

LaurenRight. It's more like a battery you set on fire.

DerekKind of, yeah. And the founder's own projection is that the reactor could eventually get to about half a megawatt.

LaurenWhich is? Give me scale. Is that a house? A neighborhood?

DerekIt's small scale industrial. Think one facility, not a grid.

LaurenOkay, so it's not solving the AI power crunch by itself.

DerekRight. And that's the honest trade-off with early stage energy tech like this. It's not replacing grid power. It's more like a piece you'd bolt on to something that already exists.

LaurenSo less the future of powering AI and more one more tool in the shed next to solar, next to batteries, next to whatever else people are throwing at the problem.

DerekNo, but here's the part that got me. This isn't the only company doing this. There's another one, AlumaPower, and they've built what they call a galvanic generator.

LaurenA galvanic generator. That's a great name. That's a name a super villain has.

DerekYeah. It converts scrap aluminum into backup power specifically for data centers. And the pitch is that it's dispatchable, meaning you can turn it on when you need it, not just whenever the wind blows.

LaurenSo it's competing with diesel generators.

DerekThat's exactly the pitch. Cleaner than diesel, and there's a recycling loop where the aluminum byproduct gets reused instead of thrown out.

LaurenIs this doing the same basic thing as Found Energy actually, burning the aluminum, or is it a different process?

DerekDifferent mechanism, actually. The reporting on AlumaPower describes it as an electrochemical process, not combustion, so it's less set it on fire and more controlled reaction. But the goal is the same either way. Turn scrap metal into electricity you can call on when you need it.

LaurenOkay, that part I actually like. Diesel generators sitting behind data centers is such an ugly little secret of this whole AI gloom.

DerekIt is. Everyone wants to talk about GPUs, and nobody wants to talk about the generator shed.

LaurenThe generator shed. That should be a segment.

DerekMaybe next season.

LaurenIt does make me want to hold on to some skepticism, though. We've covered plenty of this changes everything energy startups that quietly disappeared a year later.

DerekFair. A projection of half a megawatt is a projection, not a delivered product. Worth remembering before anyone gets too excited.

LaurenI do want to say none of this is breaking news this week. This is background. This is the researchers in a lab and startups with a pilot stage.

DerekTotally. It's scene setting. Nobody's Aluminum reactor is running your ChatGPT query tonight.

LaurenBut it tells you something about how weird the search for power has gotten. People are looking at literal scrap metal because the grid can't keep up.

DerekWhich, when you zoom out, is kind of the whole story of this AI moment. The software side gets the headlines, but somebody has to keep the lights on underneath it.

LaurenSpeaking of the software side eating jobs... Actually, wait, I want to swing this totally different direction.

DerekGo for it.

LaurenBecause we've been talking about AI wanting electricity. I've got a story about AI wanting somebody's actual job. Specific job, specific desk.

DerekOh, I know where this is going.

LaurenYou don't. Just wait for it. Okay, so get this. There's a startup called Rogo Technologies, and it built a whole platform to do the parts of banking nobody wants to do at 2:00 in the morning.

DerekThe spreadsheets.

LaurenThe spreadsheets, the slide decks, the deck nobody reads past page three.

DerekEvery junior banker's origin story.

LaurenRight? A report on Rogo lays it out. This thing is built specifically to automate the tasks that used to be the junior banker's whole job.

DerekWait, not adjacent to the job. The actual job.

LaurenThe actual job. Building models, formatting decks, the grunt work that's technically analysis but is really just assembling other people's thinking into a PDF.

DerekSo if you're twenty-four and you just got an offer from an investment bank-

LaurenYou might be training your own replacement without knowing it. That's the uncomfortable part.

DerekOkay, but hold on. Where is this actually operating? Because that changes how scared I get.

LaurenThat same report says Rogo is making Singapore its hub for the whole Asia-Pacific region.

DerekNot a satellite office, a hub.

LaurenA hub. And they're not being shy about it. The plan on the table is a regional headcount of a hundred fifty to two hundred people by the end of 2027.

DerekOkay, so walk me through that number. That's what? Roughly the size of a mid-tier trading floor?

LaurenPretty close, yeah, except none of those two hundred people are junior bankers pulling comps at midnight.

DerekThey're engineers building the thing that replaces the guy pulling comps at midnight.

LaurenExactly the swap.

DerekThat's dark. I love it.

LaurenRight? And people are clearly betting on it because the same coverage puts Rogo's valuation at two billion dollars off an April round led by Kleiner Perkins.

DerekTwo billion for a tool that makes decks.

LaurenFor a tool that makes decks nobody had to stay up all night making.

DerekKleiner's not a small name either. That's a serious firm putting serious money behind junior bankers. We've automated your Tuesday.

LaurenRight. And I want to be careful here because none of what we just said is banks are laying off analysts. That's not in the reporting.

DerekNo, this is a startup hiring aggressively for itself. Two hundred people in Singapore is growth, not a pink slip anywhere.

LaurenTotally different chart.

DerekBut you can't unsee the shape of it. If the entry point into finance used to be do the grunt work for three years so you learn the business, and the grunt work just got a two billion dollar price tag on being automated—

LaurenThen what's the entry point in five years? That's the part that's genuinely just us wondering out loud. Nobody's reported the answer

DerekIt's not a fact, it's a question mark hanging over an entire career ladder.

LaurenWhich, fair warning, is exactly the kind of thing that keeps twenty-two-year-olds up at night scrolling LinkedIn.

DerekGuilty as a demographic.

LaurenBut you know what doesn't have that problem? Robots.

DerekOh, switching gears here.

LaurenSwitching gears hard, because while everyone's debating whether AI eats the entry level finance job, there's a whole other corner of this where the money isn't creeping in, it's sprinting.

DerekPhysical AI, robotics.

LaurenRobotics, warehouses, actual machines that move actual boxes, and the valuations in that world are moving so fast that comparing them to a normal funding round barely makes sense anymore.

DerekFaster than a two billion dollar Series Whatever for a spreadsheet bot?

LaurenFaster, and there's one company in particular that tripled its own valuation in a matter of months, which is the kind of number that makes the finance automation story look almost quaint.

DerekOkay, now you have to tell me who.

LaurenPatience. That's next. Okay, so Generalist, the company I teased, just tripled its valuation in under a year.

DerekWait, tripled? Not doubled, not up a bit, tripled?

LaurenTripled. Robotics, not software. Different world, same fire hose of money.

DerekGive me the shape of it.

LaurenTechCrunch reported Generalist landed a new two hundred million dollar funding extension led by 8VC, and that extension puts the company at a three billion dollar valuation.

DerekThree billion for robots?

LaurenFor robots. And TechCrunch reported the people building it are nobodies. Generalist was founded in twenty twenty-four by two former Google DeepMind researchers, Pete Florence and Andy Zeng, plus an engineer named Andrew Barry who came out of Boston Dynamics.

DerekSo this isn't three guys in a garage with a Roomba.

LaurenNo. This is DeepMind brains plus actual industrial robotics experience under one roof.

DerekThat resume probably landed the money, too. DeepMind pedigree plus real hardware experience out of Boston Dynamics is a rare combination for investors to find in one company.

LaurenRight. It's not just a good pitch deck. It's people who've actually shipped physical products before, and that matters a lot when someone's handing over hundreds of millions to a company that's barely two years old.

DerekOkay, but two years old and already at three billion? What were they doing before this round? Just sitting quietly?

LaurenThat's the part that got me. TechCrunch reported the two hundred million dollar extension lands just months after Generalist had already hit a two billion dollar valuation.

DerekWait, wait, months?

LaurenMonths.

DerekSo somebody wrote a two billion dollar check, and before the ink's dry, somebody else says, "Actually, make it three."

LaurenThat's the pace right now. It's not a slow climb. It's a re-rating every few months.

DerekAt that speed, you could miss a whole valuation while you're on vacation.

LaurenGenuinely might.

DerekWhat is it about physical robots specifically that's got investors moving that fast? Because software rounds don't reprice like that. I think it's the belief that whoever gets a robot to reliably fold laundry or stock a warehouse shelf first basically owns a category nobody's cracked yet. Compare that to SaaS, where the category's already crowded. Here, there's no incumbent to dethrone.

LaurenSo it's not hype for hype's sake. It's a land grab on an empty map.

DerekPretty much. An 8VC leading the extension tells you it's not just excitement money. That's a firm making a repeat bet.

LaurenOkay, but let's push on that a little. A repeat bet from one firm is a signal, sure, but it's not the same as proof the robots actually work the way the company's claiming in an actual warehouse.

DerekTotally fair. None of this tells us the tech performs yet. What it tells us is investors think the payoff is big enough to fund ahead of that proof.

LaurenOkay, so zoom out with me for a second. We started this episode arguing whether AI kills software or just eats its interface.

DerekRight.

LaurenThen we're out in the parking lot talking about burning aluminum to keep data centers running.

DerekWhich still can't get over that.

LaurenThen we're worrying about whether a twenty-two-year-old can even get a junior banking job anymore because a platform's doing the spreadsheet work.

DerekAnd now robots getting repriced every few months.

LaurenIt's the same story wearing four different costumes. Money isn't trickling into AI. It's flooding every layer at once: the code, the power plant, the entry level job, and now the robot arm.

DerekAnd none of those layers are waiting for the others to sort themselves out first.

LaurenWhich is honestly the thing to sit with. Nobodys waiting for a clean answer on jobs or valuations before writing the next check.

DerekNobodys waiting, period.

LaurenSo thats where we leave it. Software, power, careers, and robots all getting rewritten on overlapping clocks, and it is genuinely wild to watch happen in real time.

DerekNobodys waiting, period.

LaurenSo thats where we leave it. Software, power, careers, and robots all getting rewritten on overlapping clocks, and it is genuinely wild to watch happen in real time. So four segments, four sectors, one thread. Nobody actually knows where AI stops being a tool and starts being the whole business.

DerekYeah, and can we just say it plainly? The Rogo story, that's a startup building automation software and hiring more people to do it.

LaurenNot a bank cutting a single job.

DerekRight. Zero confirmed layoffs. Any worry about fewer entry level seats down the road is us guessing out loud, not something Rogo or any bank has said.

LaurenWhich honestly is the whole episode in miniature. Software company says it's thriving because of agents. Power companies are literally burning metal to keep the lights on for this stuff. And robotics money is moving so fast that ink isn't dry on one valuation before the next one lands.

DerekNobody's writing the headline yet. We're just watching the checks get bigger.

LaurenWhich is kind of the fun part of doing this show every week, honestly.

DerekSure, if your idea of fun is mild existential dread with good production value.

LaurenExactly. Okay, if you like this one, subscribe, leave us a review, and if you've got a tip or you just wanna yell at us about something we got wrong, our contact info's in the show notes.

DerekWe read all of it, mostly.

LaurenMostly.

DerekNext Wednesday, we're back with a new one.

LaurenWait, are you gonna tell me what it's about, or is this the part where you go mysterious on me?

DerekMysterious. You'll just have to show up.

LaurenRude. Alright. See you all next Wednesday.