Episode #97 - Three Breakthroughs: The AI Arms Race

Frequently Asked Questions

FAQ

• Where is the transcript for this webinar?

  Speaker 1:
  Some big shifts are happening in tech and science. AI investments are scaling to unbelievable levels. Social media is breaking apart in unexpected ways, and cancer research is making real strides toward personalized treatments. Welcome back to the Tech Optimist, where we break down the breakthroughs shaping the future. Today we’re covering an AI arms race, the unbundling of social media, and a major breakthrough in cancer vaccines. You heard me right. This episode is packed with insights you won’t want to miss. So let’s get started real quick—we’ve got an ad and a disclaimer. We’ll be right back.

  Speaker 2:
  Do you have a venture capital portfolio of cutting-edge startups? Without one, you could be missing out on enormous value creation and a more diversified personal portfolio. Alumni Ventures, ranked a top 20 VC firm by CB Insights, is the leading VC firm for individual investors. Believe in investing in innovation. Visit av.vc/foundation to get started.

  Speaker 3:
  As a reminder, the Tech Optimist podcast is for informational purposes only. It’s not personalized advice and it’s not an offer to buy or sell securities. For additional important details, please see the text description accompanying this episode.

  Speaker 1:
  And we’re back. Now, let’s dive into today’s episode where we’re unpacking some of the biggest breakthroughs that I mentioned earlier—from the AI arms race to a game-changing medical discovery. This is one episode you’re not going to want to miss. Let’s get started.

  Speaker 4:
  Hey, and welcome to this week’s Three Breakthroughs. I’m here again with Lucas. I’m Mike Collins, the founder and CEO of Alumni Ventures. We get together every week and talk about what’s going on in tech and science and innovation and venture capital and whatever the hell else we want to talk about.

  Speaker 1:
  This seems to be my cue in this type of episode. Before we dive deeper into the conversation, I’m going to set the stage a bit more for what’s ahead because today’s breakthroughs aren’t just important—they’re reshaping the future in real time.
  First up, we’re talking about the AI arms race, and things are escalating fast. Google, Microsoft, and Amazon are investing billions into compute power. I’m going to explain what that means a little bit later, but they’re building out massive AI supercomputers to fuel the next generation of models. But is this a necessary step forward, or are we racing toward an AI monopoly where only the biggest players can compete? And what happens when AI advancement is limited by who can afford the most GPUs? We’re going to break that down today.
  Next, social media is evolving as users seek more tailored experiences. We’re seeing a shift from all-in-one platforms to smaller, more specialized communities. So what does this mean for the future of online interaction, and how will platforms adapt to this changing landscape?
  And then finally, we’re going to shift from technology to medicine, where one of the most promising breakthroughs in cancer research is unfolding. A new cancer vaccine trial has delivered stunning results, with kidney cancer patients showing zero recurrence of the disease after three years. It’s an incredible step forward in personalized immunotherapy, and if this approach scales, it could change how we treat cancer forever. The world is shifting, industries are evolving, and these breakthroughs are leading the charge. Stick with us—we’re about to break it all down.

  Speaker 4:
  So the big splash a few weeks ago was DeepSeek, and everybody freaking out. We talked a little bit about it, but we’re now a couple of weeks into it. And from the people writing big checks in the know, it seems to be a yawn, and we are going to continue to spend enormous amounts of money on computers and data centers and energy. And if you listen to the earnings reports, they all believe that they are compute-constrained and not backing off at all. Whether that’s Meta, whether that’s Google, it just seems that they’re plowing ahead. What’s your take on it, Lu?

  Speaker 5:
  Yeah, I mean, not backing off is putting it lightly. Last week was Google’s earnings announcement, and they announced plans to increase CapEx to $75 billion in 2025. And that’s off of $50-something, I think, in 2024. Analysts were expecting them to increase to like $58–60 billion, but they just blew that out of the water. So big tech is going all in, writing checks so big they may as well be government defense budgets. It’s Amazon, Google, Microsoft, Meta—pouring tens of billions of dollars into AI infrastructure.

  Speaker 1:
  Alright, let’s take a second to talk about how massive $75 billion really is, because when you throw around numbers that big, it’s easy to lose perspective. So let’s break it down.
  If you had $75 billion in hundred-dollar bills, the stack would reach about 50 miles high—that’s nearly twice the height of the stratosphere. And if you spent $1 million every single day, it would take you over 205 years to run out of money.
  Now, let’s put that into some real-world comparisons. Netflix, the streaming powerhouse, made about $39 billion in total revenue last year. So $75 billion is nearly double Netflix’s entire business. It’s also 40% of the entire global gaming industry.
  Or how about this? When the US sent astronauts to the moon, the entire Apollo program cost $25 billion in the 1960s—adjusted for today’s dollars, that’s about $150 billion—meaning $75 billion is half the cost of putting humans on the moon multiple times.
  And for one more example to round this off: in the world of tech, companies like Meta spent $33 billion last year on AI infrastructure, meaning that $75 billion could more than double Meta’s AI investment. OpenAI’s next big model, GPT-5, is rumored to cost between $1 and $10 billion to develop, so $75 billion could fund multiple generations of the most advanced AI systems we’ve ever seen.
  Bottom line: $75 billion isn’t just a big number. It’s a game-changing amount of money that could shift entire industries, fund moon missions, or shape the future of AI for years to come.

  Speaker 5:
  And I think their big bet is even in the face of what DeepSeek brought to the market a couple of weeks ago, they believe that dominance in compute today is just going to mean greater control in this sector in the future.

  Speaker 4:
  And I’m in the camp too that these people are not run by speculative, go-from-the-gut people. These are run by the smartest. They’re seeing ROI and ROIC. So I tend to trust them that this is not just game theory—“I’ve got to spend it because that guy’s spending it.” I believe they’re looking at usage. I believe they’re looking at their business model and what they can charge, and they’re basically saying, “This is being used, this is being paid for. This is an investment that is going to pay back for our company and our shareholders.”
  So I believe that is true—that this is not mass hallucination and a bubble. I believe people are using this stuff every day and are going to be using it more and more, and they’re investing where there is demand. So I could be wrong, but I tend to fall on that end of the spectrum—that these people are seeing things, seeing numbers, sharing it, and they literally are, “We cannot make this stuff fast enough.” Even our portfolio company, just this week in the Middle East, announced a big deal—a billion and a half dollars for Groq.

  Speaker 4:
  Yeah. And again, I think that this is not just the big tech investing here. I think you’re going to increasingly see countries investing in this infrastructure.
  I think you’re going to see small and medium-sized businesses looking to control their own models, customize their own models, train their own models. I think all of the above—yes, yes, yes. And we’ve talked about this too, which is that there are going to be a lot of picks and shovels. There are going to be a lot of energy constraints. I think the interesting debate is: where is there venture money to be allocated? Where in the stack do you want to play? Do you want to play it close to the metal? Do you want to play way up the stack at the application layer? I think those are all interesting debates, but I think big tech has spoken, which is: we’re not blinking. We need more compute. And obviously that’s going to feed into Nvidia and the like for sure. But this whole shock and recalibration—I’m not seeing it.
  And to plug another podcast, I think there was a really good discussion over on BG2. I’d encourage people who want to dig into DeepSeek and this analysis to check it out. I thought their analysis was phenomenal. So again, from what I hear about viewership and numbers and stuff—we’re trying to add value mostly to our customers and our portfolio company CEOs—go check them out. Great discussion.

  Speaker 5:
  Great discussion. And I think in terms of the two-week fallout from DeepSeek—sure, it’s not going to cause big tech to spend any less—but what the BG2 guys were talking about, I think, was really fascinating. DeepSeek’s rise as an open-source AI powerhouse is something that the rest of the industry is reckoning with and rethinking. For years, OpenAI, Google—they were able to justify their closed models by arguing AI is too powerful, too dangerous, too expensive to give away for free.

  Speaker 4:
  That genie’s out of the bottle for sure. In fact, I think you raise a really good point, Lucas, about Sam Altman in an interview just recently saying, “Hey, I think we need to rethink that.” So again, I encourage people to check out Sam’s fascinating blog on the three rules of AI where he talked about scaling laws and one-tenth of the cost, but also this opening of: “Hey, I think there’s a place for open source even within our framework of the world.”
  I think you’re absolutely right. Scale is one dimension: how much money, how big, how fast do we build compute? I think energy is the dimension. And I think there’s a fascinating question of where value is going to accrue in this whole open-source discussion. Fascinating.

  Speaker 5:
  Yeah, it is really interesting how Sam’s tone has shifted in the last couple of weeks. It started with “gear up for the lawsuits,” distillation practices—

  Speaker 4:
  “They ripped me off.”

  Speaker 5:
  —and now he’s acknowledged—

  Speaker 4:
  —Ironic from the guy who scraped the internet.

  Speaker 5:
  Right? But he has acknowledged: no, this was impressive. And there was real innovation here. I don’t know if he’s going to drop the lawsuits ultimately—we’ll see what he does—but he definitely indicated an understanding that the R1 model from DeepSeek really did require some impressive innovation. And he also indicated a potential openness to implementing a more open plan—an open plan for OpenAI—which I think, initially when the company was founded, was supposed to be a pretty open model and has been closed really to date.

  Speaker 4:
  And again, I think this discussion too about what does “open” really mean? A lot of people throw that around like it’s black and white. I think one of the things that I imagine happened is Zuck lost his mind, because he really positioned: “Hey, we’ve got Google over here trying to finesse what they’re doing while losing their search business, and you’ve got a nonprofit-for-profit closed system ChatGPT over here, and we are going to be the open-source guys—with an asterisk.” But they’re really on the scale of what I’ll call semi-open source.
  Because if you were big enough and reading the fine print—again, Facebook talking a big game but doing a lot of other stuff, a lot of things they’re trying to protect—and then really the renegade entrepreneur who, in this case, happened to be in China says, “No, we’re going to be actually a little further down the scale of open source and rock the world.” So again, from where I sit in the cheap seats, all good competition.
  We’re trying to invest in the elephants’ footsteps here, but I think it’s good. I think it’s good for humanity. I think this illusion of: “Oh, we’re going to keep the genie in the bottle, and we’re going to keep it in the United States, and there’s only a few players that are going to play”—whether that ultimately would have been the right thing to do or not—I think that was never going to happen. Technology, life finds a way. Technology finds a way. Innovation finds a way. Capitalism finds a way. I think all that is true.

  Speaker 2:
  Exceptional value creation comes from solving hard things. Alumni Ventures’ Deep Tech Fund is a portfolio of 20 to 30 ventures run by exceptional teams who are tackling huge opportunities in AI, space, energy, transportation, cybersecurity, and more. These game-changing ventures have strong lead venture investors and practical approaches to creating shareholder value. If you are interested in investing in the future of deep tech, visit av.vc/deeptech to learn more.

  Speaker 5:
  Let’s move on to something that we’re calling the social media unbundling. So for a long time now, social media has been synonymous with a handful of giants in the arena: Meta, X, TikTok, Snap to some degree. But there are cracks that are forming—consumer frustration with algorithmic feeds and privacy concerns, and the sheer dominance of big tech has been fueling this new wave of decentralized, open-source social networks.
  So keeping on the theme here of open source—what that means is, unlike their predecessors, these newer platforms are built on open protocols like ActivityPub or AT Protocol, and it allows users to move between apps without being locked into a single company’s ecosystem.
  Think about how email works, right? You have your Gmail account, but that doesn’t stop you from emailing someone who uses Outlook or Proton Mail or a different client. The service you use is just a preference; it’s not a restriction in any way. And we’re starting to see social media move in that direction, and I think it’s really exciting. We are early investors in BlueSky—that is kind of the open protocol—

  Speaker 4:
  —version of that—

  Speaker 5:
  —version of X. And it allows the apps to kind of communicate with one another just like email providers do. So in the future, social media is not about choosing one platform; it’s about choosing the actual experience that you prefer.
  And one of the reasons that I want to talk about this is there are a rising number of, just like BlueSky has been a rising competitor to X. Mastodon was another rising competitor to X that was big last year. BlueSky has amassed 25 million users—

  Speaker 4:
  30 now, I just heard. Yeah.

  Speaker 5:
  Incredible. And so we’re starting to track something similar on the Instagram side. There’s one called PixelFed that’s gaining some traction on the ActivityPub protocol, and it’s something that we’re looking at really closely here at Alumni Ventures.

  Speaker 1:
  So, as Lucas prepped us earlier, social media is shifting and BlueSky is at the center of it. In just four months, the platform has grown from 10 million users in September of last year to over 27 million by mid-January of this year—a remarkable 174% increase. So what is driving this rapid adoption? A major factor has been the 2024 U.S. presidential election, which led to a significant number of users seeking alternate platforms. BlueSky saw daily increases of over 500,000 users, with some reports showing up to a million new signups per day in late November.
  But this isn’t just about one election—it reflects a broader trend in how people are engaging online. Many users are exploring new platforms that align with their evolving preferences for social media. Studies indicate that BlueSky has gained traction with certain political groups while other platforms continue to maintain their own dedicated audiences.
  Additionally, recent changes in leadership and platform policies across social media have influenced how users interact, contributing to BlueSky’s growing appeal. Its decentralized approach and user-controlled experience have been major factors in attracting new members. Even political figures are recognizing its potential—several elected officials, including U.S. representatives and state leaders, have joined the platform, reinforcing its role in digital communication.
  BlueSky’s growth highlights an ongoing shift in the social media ecosystem where users are increasingly seeking platforms that reflect their communication preferences, content policies, and digital experiences. So what does this mean for the future of online interaction? And will this trend continue? What does that look like?

  Speaker 4:
  No, and again, I think this is really healthy for the ecosystem, right? When you have one dominant platform, competition is good. It leads to innovation, it leads to customer choice, it leads to options. And a lot of these social networks have such powerful network effects that you really have huge lock-in, and that makes for a really good business. But I think consumers benefit with choice.
  And as you know, we’re big fans of BlueSky and their approach and their tech stack. And this idea of—everybody thinks, “Oh, Alumni Ventures was early at BlueSky. That must’ve been done by the Women’s Fund because there’s a woman CEO and founder.” And it was actually done by our blockchain team, and it’s in several of our funds. We made the initial investment because we really just loved this idea of decentralized tech stack, consumer choice, and the ability to move.
  Yeah, we’re just very excited about that as being a good option for people—to put more control and more options in the hands of the user with their information, their preferences, and their control. So it’s good stuff, and I think we’re going to be active with more companies that are innovating along that dimension as well. We’re big believers.

  Speaker 5:
  Yeah, it’ll be interesting to see what new business models come to prominence in this future. Big tech’s dominance in social media has been fueled by advertising revenue, and they thrive because they control the user experience. They track every click, like, and comment, and they can serve these hyper-targeted ads.
  In a more open, decentralized social web, that control erodes completely. And so it’ll be interesting to think about how these social platforms actually make money. Instead of squeezing revenue out of ads, we could see a move toward subscription-based models, creator-driven monetization, and community-funded platforms. Mastodon is sustained by donations, BlueSky is exploring more premium features. But at the end of the day, what’s mostly—

  Speaker 4:
  If you have attention and you have eyeballs—I mean, again, coming back to the venture capital preference—the way I look at it is if you have attention and you have eyeballs, you can figure out a business model. And obviously we saw that with Google, we saw that with Facebook—we live in an attention economy, and good, smart people will figure out a fair way to build value in a good business model. But it all starts with getting people to use your thing, for sure.

  Speaker 5:
  And make no mistake, this makes our jobs hard. We like investing in companies that have high switching costs, and this reduces switching costs immensely. I’m annoyed as—, but better competition is good. And as a consumer, I wake up to notifications that YouTube TV is going to increase my subscription and they’re going to lose all of their Paramount-affiliated channels possibly, and it’s really annoying. So the more that players in the ecosystem need to provide the better experience in order to keep users, I think is a good thing.

  Speaker 4:
  It’s capitalism, right?

  Speaker 1:
  Imagine a world where cancer treatment isn’t one size fits all, but instead uniquely tailored to each patient. That’s what NeoVX, a personalized cancer vaccine, is designed to do. And the latest clinical trial results are nothing short of groundbreaking.
  NeoVX is a custom-built vaccine that trains the body’s own immune system to recognize and eliminate remaining cancer cells after surgery. How cool is that? Using neoantigens extracted from a patient’s tumor, the vaccine creates a highly targeted immune response, reducing the chances of recurrence in advanced kidney cancer patients.
  And here’s where it gets exciting—every single patient in a recent phase one trial remains cancer-free nearly three years after receiving the vaccine. That’s 100% of participants with stage three or four kidney cancer showing a strong and sustained immune response. The vaccine works incredibly fast. T-cell responses spiked within three weeks, increasing by a staggering 166-fold and staying at high levels for up to three years. And while some patients experienced minor side effects like injection site reactions and flu-like symptoms, no severe adverse effects have been reported.
  But this is just the beginning. Larger trials are already underway with researchers exploring how NeoVX could work alongside other immunotherapies to create even more powerful, long-lasting treatments. A vaccine that fights cancer after it’s removed, trains the body to prevent its return, and has zero recurrences in early trials—if these results hold, we could be looking at a game-changing future for kidney cancer treatment and beyond.

  Speaker 4:
  And the last thing on today, right? We’re in a day where we, I believe, confirmed a new person to lead the department that oversees vaccines who have been accused—fairly or unfairly—of being a vaccine skeptic. I just want to spend a minute talking about a really promising new drug discovery in the area of vaccines for cancer.
  And so again, I go to science, and unequivocally, vaccines have done more for humanity than almost any technology of the last 100 years—in lives saved, misery avoided. The data of good being done—the chart for vaccines—is very, very strong. So huge fans of vaccines, and exciting news just down the road at Dana-Farber.
  Again, a very small trial, just a handful of people, but NeoVX, which is dealing with kidney cancer—the vaccine was administered after surgery, and at median follow-up after three years, all patients remain cancer-free, showing that the immune response kind of swooping up all the random little kidney cancer cells floating around had yet to find a foothold, had yet to find traction.

  Speaker 4:
  These patients’ immune systems had been effective in gobbling this stuff up with relatively minor flu-like symptoms that, again, are the equivalent of you and I having a bad night out with airport sushi. So a really good trade-off there. Again, you and I have talked about Lucas, at length, the real good work being done now in life sciences. We’ve talked extensively about the innovation with GLP-1 drugs. And again, data came out just this week about helping with treating alcohol and fatty liver disease. And again, kind of the fourth thing on the topic list that we didn’t get to—which is people using surgery to really effectively combat advanced disease around liver problems.
  And so again, I just want to put a plug in for vaccines today and this really promising thing coming out of Dana-Farber for kidney cancer.

  Speaker 5:
  Yeah, Mike, thank you for sending me the article. I read through the study summary from Dana-Farber and I got goosebumps. This is a big moment. I have someone very close to me in my life who, in the summer of 2023, was diagnosed with late-stage kidney cancer. Thankfully, the immunotherapy he’s on has been very effective, but the prognosis initially was not good, and the therapy took a toll.
  And the idea that, in these patients in this phase one study, at stage three and four—meaning there’s been some degree of metastasis—all now…

  Speaker 4:
  Three and four is “the horse has left the barn,” and it’s very bad news.

  Speaker 5:
  Right. All nine of these patients are cancer-free still three years after treatment.

  Speaker 4:
  I—

  Speaker 5:
  —I mean, this is a huge moment. Cancer vaccines have been talked about for years. They’ve been mostly theoretical or limited to only a couple of types of cancers. And if these results hold up in larger trials, we could be witnessing the beginning of an entirely new category—

  Speaker 4:
  —of hope for later-stage cancer patients. And again, I think under the sheets, we’re seeing the benefit of CRISPR. I think we’re seeing the benefits of AI. I hear from a lot of cancer researchers that they’re using these tools—some of the tools coming out of Google with protein folding. All of this is creating a mesh of opportunity and optimism and hope for people.
  And again, just underscoring that this specific example is just one of many and hopefully the beginning of much, much more. I go back to—we just lost President Carter at age 100, but I think he was diagnosed with brain cancer—I forget the exact age, but it was a while ago. And people were basically saying “hospice, he’s done.” And because of immunotherapy, the ability to take one’s own immune system—which is, however religious you are, it is a beautiful system to keep us alive and keep us healthy.
  And frankly, to fight cancer and fight other diseases and things—it is a wondrous system. And by tweaking it and attuning it and activating it, it is turning out to be the best cure, the best way to address something like the scourge of cancer. So I am just super excited about these cancer vaccines and other related things—whatever you want to call them. It may need a new branding, but just turbocharging our immune system to fight this problem—everybody’s all on board on that, for sure. We don’t have to get political or anything, but if you’re stage three cancer, you’re looking for answers. And if the answer is your own body, super exciting, super promising. Like you say, it gives us all goosebumps because we’ve all been touched by cancer in our families and lives.
  So, well, Lucas, it’s been great. I think your sprint tour of duty—you’ve done great. In a couple of months, we’re on to another young buck. Drew Sek, I think, is going to be kind of the next on the baton, and we’ll sprinkle back to—

  Speaker 5:
  You’re not rid of me yet. I think we’re going to go once more next week, Mike.

  Speaker 4:
  Excellent. I am preparing our millions of listeners and followers for the coming baton passing. So, alright, we’ll do it again next week, Lucas, and have a good night.

  Speaker 5:
  I’ll make sure to bring it home with a good one next week. Thanks, Mike.

  Speaker 4:
  Bring it up. A lot of pressure. Alright. Okay. See you, buddy.

  Speaker 3:
  Thanks again for tuning into The Tech Optimist. If you enjoyed this episode, we’d really appreciate it if you’d give us a rating on whichever podcast app you’re using, and remember to subscribe to keep up with each episode. The Tech Optimist welcomes any questions, comments, or segment suggestions. So please email us at [email protected] with any of those, and be sure to visit our website at av.vc. As always, keep building.