Weekly Dose of Optimism #153

Trump AI, Hyperion, Bedrock Robotics, IVF Threesome, GABA Labs, Reindustrialize, Floundering

Hi friends 👋 ,

Happy Friday and welcome back to our 153rd Weekly Dose of Optimism. Pretty electric week for our little corner of the internet: Windsurf drama, Grok models, Epstein files, and Coldplay concerts. If you’re chronically online and enjoy some good gossip, this was a busy week for you. Luckily, there’s some pretty important stuff work happening outside of the X-sphere. We’re here to cover that stuff.

Let’s get to it.

---

Today’s Weekly Dose is brought to you by… Stripe Startups

Packy here. I have been a big Stripe fan for a long time, and wrote about the company all the way back in August 2020, so I’m proud to welcome them as our newest sponsor and tell you about Stripe Startups.

Stripe’s mission is to increase the GDP of the internet, and it’s working. Last year, 1.3% of global GDP flowed through Stripe, over $1.4 trillion and growing. 78% of the Forbes AI 50 use Stripe’s financial infrastructure to monetize faster, experiment with pricing, and grow revenue.

As I wrote in my original Deep Dive, Stripe’s strategy from the earliest days has been to make it as easy as possible for startups to begin accepting payments, and then grow with them as they get really big. With Stripe Startups, they’re making it even easier.

Stripe Startups is the company’s new program designed to support early-stage, venture-backed businesses as they build, iterate, and scale. Founders enrolled in Stripe Startups get access to credits to offset Stripe fees, expert insights, and a focused community of other founders building on Stripe.

You’re going to use Stripe anyway. Get the white-glove experience. Apply to Stripe Startups today.

---

(1) Trump Touts Billions in Investments to Create AI Hub in Pennsylvania

Amrith Ramkumar and Brian Schwartz for WSJ

Google said it would put $25 billion into data centers needed to train AI models and related infrastructure in Pennsylvania and the surrounding region over the next two years. Private-equity firm Blackstone promised another $25 billion. AI startup CoreWeave announced a $6 billion investment.

Power companies FirstEnergy and Constellation Energy are part of a group pouring billions more into increasing electricity generation in the area, according to the White House.

The Trump Administration announced a bigly investment from the private sector into AI infrastructure, with the majority of the investment into (notable swing state) Pennsylvania. At the event organized by PA Senator David McCormick (no relation, despite being a McCormick from PA), executives from Google, Anthropic, Blackstone, and Exxon pledge their allegiance and wallets to the initiative. It appears to be a combination of data center build outs energy infrastructure to feed those data centers, with Google and Blackstone footing most of the bill.

This fresh $56 billion in energy and $36 billion in data center investments comes on the heels of Amazon’s $20 billion commitment to the Keystone State.

We love to see a good public<>private partnership like this, and Dave McCormick is particularly well-suited to make it happen having split his career in politics and as the President/CEO of Bridgewater, one of the world’s largest hedge funds.

We particularly love to see this partnership happening in Pennsylvania, our home state, where projects like this will generate thousands of jobs. It’s also a big win for our boy Josh “Rebuild I-95 in 12 Days” Shapiro. I guess there are some pretty serious perks of being the swing state. Go Birds.

(2) Mark Zuckerberg says Meta is building a 5GW AI data center

Maxwell Zeff for TechCrunch

Meta is currently building out a data center, called Hyperion, which the company expects to supply its new AI lab with five gigawatts (GW) of computational power, CEO Mark Zuckerberg said in a Monday post on Threads.

Not to be outdone, Zuck is planning a massive data center build out, the scale of which has never been seen before. The new wave of superclusters, including one called Hyperion, could reach 5 gigawatts or about 5x the size of the largest existing data centers. In a Threads post, Zuck posted an animation showing that the data center will be almost the size of Manhattan.

Zuck and team are moving quickly here, even planning to build some of the centers under temporary tents in order to meet surging demand for AI compute. Just this year alone, Meta is planning to spend approximately $70B in capex in support of its AI initiatives. $70B!

The overall strategy is coming together. Attract and assemble a world-class team of AI researchers and operators, with the help of $100M+ comp packages. Invest extremely heavily in AI infrastructure and compute. In the short term, monetize its AI investment through its existing family of cash-gushing apps and in the long term pull ahead in foundational model work, fuse AI into its new hardware products (Glasses, VR, etc), and use its massive footprint to spin up its own cloud computing business.

The strategy, if it works, will be good for Meta. The strategy, even if it doesn’t work, will be good for everyone else, pulling forward data center buildouts and pushing forward AI research. Call him GigaZuck.

(3) Ex-Waymo engineers launch Bedrock Robotics with $80M to automate construction

Kirsten Korosec for TechCrunch

Bedrock Robotics, an autonomous vehicle technology startup founded by veterans of Waymo and Segment, has been operating quietly for more than a year. Now, it’s breaking cover with an $80 million funding round from investors Eclipse and 8VC. Bedrock Robotics is focused on developing a self-driving kit that can be retrofitted to construction and other worksite vehicles, according to the company.

A team of former Waymo and Google X engineers launched Bedrock Robotics, a company focused on automating physical labor in construction, with a cool $80 million in Series A funding. Bedrock is building robotic systems designed to take on repetitive, labor-intensive tasks like drilling, lifting, and positioning materials and its starting with rebar installation. Every infrastructure project needs rebar, and it’s super time intensive and repetitive. Perfect first use case.

Unlike traditional factory robots, Bedrock is not designed to execute against tightly controlled instructions, but rather to dynamically operate in a messy construction environment. This is the new thing in robotics. You don’t just want a robot that does a task, you want a robot that understands the task and then does it.

We hope we’re building a lot in this country over the coming decades. Data centers, houses, roads, etc. A lot of that work can and will be done by humans. But a lot of that work should and will be done by robots. It’d be a lot safer, cheaper, and faster.

IT’S TIME FOR ROBOTS TO BUILD.

(4) ‘Landmark’ study: three-person IVF leads to eight healthy children

Ewen Callaway in Nature

The procedure has been dubbed three-person in vitro fertilization (IVF), because the resulting children carry nuclear DNA from a biological mother and father, alongside mitochondrial DNA from a separate egg donor.

In a scientific twist on the classic ménage à trois, UK researchers reported the first long-term results from a regulated “three-person IVF” technique that combines DNA from three individuals to prevent inherited mitochondrial disease. The procedure involves transferring nuclear DNA from a fertilized egg with faulty mitochondria into a donor egg with healthy ones, effectively bypassing the mother’s disease-causing mitochondrial DNA.

The latest data, published in the New England Journal of Medicine, shows that eight children born through this technique are all developing normally, with only minor unrelated health issues in a few cases. This marks the first peer-reviewed evidence that the method can produce not just live births, but healthy toddlers. The UK remains the only country with explicit regulations allowing mitochondrial donation, but these results will likely change that. For families with a history of mitochondrial disorders, this represents a nice step towards healthier reproduction, with a little help from a third DNA donor.

(5) The Quest for a Hangover-Free Buzz

Kristen V. Brown For Bloomberg

At its core, the idea is to amplify the effects of gamma-aminobutyric acid, a neurotransmitter that, in lay terms, carries chemical messages to your nerve cells telling your brain to chill out. The molecule aims to bind to a GABA receptor, increasing the flow of negative ions that inhibit neuronal activity. By enhancing the brain’s GABA function, the molecule is meant to deliver a version of alcohol’s pleasant, relaxing high without the downsides.

Trust tree: I’m a lil hungover writing this. Had a friendly steak dinner last night with the boys that included 2 beers and 2 glasses of a bold, full bodied red. Nothing crazy in the moment, but I am certainly feeling a bit groggier than I’d prefer this AM. It was a tradeoff I made though: I got to feel a little looser at dinner with some friends and in return I feel a bit crappier this morning. It’s a tradeoff many of us confront on a near-weekly basis. It’s kind of part of life.

But perhaps this isn’t a necessary tradeoff. GABA Labs is developing a synthetic molecule called Alcarelle designed to mimic the relaxing, social effects of alcohol without the hangovers and health risks. The molecule works by selectively activating a specific GABA receptor in the brain linked to sociability, aiming to produce the effects of a glass of wine (calm, connected, mildly euphoric…the feeling I was aiming for at dinner last night) without impairing motor skills, triggering addiction, or setting you up for a slow next morning.

The company is still working towards regulatory approval, but plans to sell/license its technology to beverage makers. It won’t be commercially available until at least 2027. So in the meantime, get out there this weekend and enjoy a few good ole fashioned pops, if that’s your thing.

Bonus: Notes from Reindustrialize

Telerobotic Palmer Luckey; photo by Ashlee Vance

Packy here, reporting live from Detroit, where I spent the past couple of days at Reindustrialize. I don’t like conferences, and I love this one. Hats off to the organizers: Aaron Slodov, Austin Bishop, Falon Donahue, Gregory Bernstein, Kate Moon, and Mike Slagh. The density of people building incredible companies is through the roof.

I’ll be writing more about a bunch of the companies that presented in the months to come, but wanted to share some quick thoughts here.

First off, Detroit really impressed me. If you haven’t been in a while, it’s worth a trip. The city is a symbol: once great, then left to rot and left for dead, now revitalized. Detroit Mayor Mike Duggan addressed the conference, and said that what made the turnaround possible was that the city didn’t try to become the next Silicon Valley or the next bio hub. California and Boston are really good at that. Instead, it returned to its manufacturing roots and went all hands on deck to attract manufacturers back to the city. Its unemployment rate dropped from 20% to 8%. People are moving back into Detroit. And downtown is awesome; my hotel overlooks Comerica Park, everything is within walking distance, the people are Midwest nice, and it feels fresh, clean, and vibrant. It’s possible to turn things around. Detroit is winning by differentiating.

That is, of course, a metaphor. A few years ago, the question was: should we Reindustrialize? Now, the question is: can we?

After a couple of days here, my answer is yes. But it won’t look like China, and it shouldn’t. We’re not going to be manufacturing iPhones. Instead, we’re going to Reindustrialize by doing what we do best: letting thousands of companies innovate and compete, and create millions of good jobs and new capabilities in the process.

There’s no silver bullet, and it’s hard to imagine a task as big as reindustrialization top-down. Bottoms-up, though, things look good. Hadrian announced a fresh $260 million in funding to build new factories and create thousands of jobs. Nuclear companies - both fission and fusion - are making rapid progress and will need a lot of smart people to scale against aggressive timelines. Companies at the conference are building flying cars and supersonic planes and drones and all manner of new, fast things that will need a lot of skilled hands and big brains to make fly. Some are even building new cities.

Many of these companies will fail. Some will succeed, enormously, and their success will mean new jobs, new supply chains, and new capabilities. From the front lines, it’s easy to feel the momentum. The government wants to make this happen. The talent is flocking to hard things. The capital allocators are allocating.

There’s a lot of work left to be done. Let’s get to work.

Bonus Bonus: Flounder Mode with Brie Wolfson

Packy again. I loved this conversation with Brie Wolfson. It’s about her wonderful profile on Kevin Kelly, Flounder Mode, and it’s also about what it’s like to work on Flounder Mode.

Isaac Asimov attempted to predict the professions of future in his classic essay, Whatever You Wish. He wrote:

But to most people the field of choice might be far less cosmic. It might be stamp collecting, pottery, ornamental painting, cooking, dramatics, or whatever. Every field will be an elective, and the only guide will be "whatever you wish."

Each person, guided by teaching machines sophisticated enough to offer a wide sampling of human activities, can then choose what he or she can best and most willingly do.

Is the individual person wise enough to know what he or she can best do? —Why not? Who else can know? And what can a person do best except that which he or she wants to do most?

Won't people choose to do nothing? Sleep their lives away?

If that's what they want, why not?—Except that I have a feeling they won't. Doing nothing is hard work, and, it seems to me, would be indulged in only by those who have never had the opportunity to evolve out of themselves something more interesting and, therefore, easier to do.

In a properly automated and educated world, then, machines may prove to be the true humanizing influence. It may be that machines will do the work that makes life possible and that human beings will do all the other things that make life pleasant and worthwhile.

Reflecting on the conversation, I realized that this is how people like Kevin Kelly and Brie Wolfson work today. They do “whatever they wish,” which doesn’t mean that they do nothing. They do “that which he or she wants to do most,” “all the other things that make life pleasant and worthwhile.”

If this is the way that more and more of us will work, learning how Kevin and Brie do it seems like a lovely way to spend a weekend hour.

---

Have a great weekend y’all.

Thanks to Stripe Startups for sponsoring. We’ll be back in your inbox next week.

Thanks for reading,

Packy + Dan