@LongJourneyVC Residency building solar, hydro, geothermal powered data centers, planting trees, restoring corals ex @sfcompute, @notionhq, @emcollective
San Francisco
Joined March 2011
- Tweets4.5K
- Following4.8K
- Followers2.2K
- Likes16.3K
Andy Rapista retweeted
I interviewed @neilmovva for @substratemag on latency-insensitive inference, alternative chips, and why massive datacenters face diseconomies of scale.
"We've always predicted that async inference is going to eat the world. That could easily be 95% or more of inference demand."
Neil's been at the frontier of inference for years, from NVIDIA to co-founding @sailresearchco. I learned a lot from our conversation - you can listen below.
Andy Rapista retweeted
The AI infrastructure buildout is about to enter an ENTIRELY different phase in 2027.
SemiAnalysis just released a new report forecasting 51.2 GW of new data center capacity in 2027, compared to 25 GW in 2026 and just 15.6 GW in 2025. We're talking about more than DOUBLING annual capacity additions in a single year. To put that into perspective, SemiAnalysis estimates that the industry could add roughly as much data center capacity in 2027 alone as it had built cumulatively through 2024, excluding China. And here's where things get REALLY interesting. While some investors, including Brad Gerstner, have questioned whether the industry can actually deliver this much capacity, SemiAnalysis points to satellite imagery showing more than 24 GW of data centers already in advanced stages of construction, with those facilities expected to be delivered in 2027. On top of that, the firm reports roughly $1 TRILLION in binding, non cancellable data center leases signed in 2026 so far.
Now, think about what this means for the ENTIRE AI ecosystem. Every additional gigawatt of data center capacity creates enormous demand for GPUs, HBM memory, networking equipment, cooling systems, electrical infrastructure, and POWER. And remember, adding data center capacity doesn't automatically mean every facility will be fully equipped with GPUs on day one. In fact, SemiAnalysis is already raising the possibility that as power and construction constraints ease, the next major bottleneck could once again shift towards semiconductors and memory. And here's what I think a LOT of investors are missing. More compute capacity doesn't necessarily mean demand suddenly gets saturated. As inference becomes cheaper, AI adoption can accelerate across autonomous agents, enterprise applications, robotics, and eventually physical AI. Lower costs can unlock entirely new use cases that simply weren't economical before, creating another wave of demand for compute.
If SemiAnalysis is even remotely close to being right, we're about to witness one of the largest infrastructure expansion cycles in history. And I believe the companies supplying the compute, memory, networking, and power behind this expansion could be some of the biggest beneficiaries over the next several years. If you want to see how I’m investing around this next phase of the AI infrastructure buildout, check out my full portfolio below: link.milkroad.com/tx5qip
Microsoft just walked away from one of the biggest AI data center projects in the United Kingdom.
According to the Telegraph, $MSFT terminated its deal for Nscale's planned site in Loughton, Essex, which ministers had called the country's largest AI data center.
The problem is POWER. Grid officials reportedly told Nscale they do not expect enough electricity for the site until the 2030s.
Nscale says its relationship with Microsoft remains strong and the grid issue has no impact on its other contracts.
The timing is rough. Nscale filed for a United States IPO last month, with Microsoft and Anthropic expected to be its biggest customers.
Demand for compute is not the problem but getting it plugged in is.
Andy Rapista retweeted
Most children (and adults) can identify 100+ logos and nearly 0 native plant species
This is not healthy for society
Native wildlife roots us in our place and we should treat it with such reverence
If we do not root ourselves in the land it is very easy to lose our ourselves
Andy Rapista retweeted
💯 Will has the most clarity here than anyone I follow on this app
Related thoughts:
1. People underestimate how much credit is existential and core to being able to scale *anything* in this new age
2. Regarding the latter, startups will need to move and become sophisticated/creative faster here than ever before as this is where incumbents have a non-trivial advantage
3. Every innovative thing we see (esp in latter weeks with muse wonder, grok bots, etc) is downstream of compounding advantages of capital and physical resource allocation. Will only accelerate as physics and bio become the new frontier
4. Exponentially more than ever before, tech will need to understand The Will Of The People, especially since the things we want to construct/need to build to accelerate as a society (data centers, factories, nuclear reactors, etc) are often viewed by the average American as physical manifestations of a future where they are left behind
Andy Rapista retweeted
New from me in the @FT: The US needs to find a way to build data centers that works for our workers and communities. BAAJA is the beginning of a model (pass it!) but there's so much more we can do to use this growth engine to help Americans. The piece makes a few points:
1 - Data centers pose challenges (emissions, energy use, noise, etc) some of which are highly concentrated, but also economic and strategic benefits that are more diffuse. Finding a way to build them in America allows policymakers to mitigate the challenges and maximize/share the benefits. Anti-building approaches that push construction overseas (the most likely outcome if we don't build) give up our ability to do so. We should want jurisdiction over them.
2 - If you want to take clean energy tech down the cost curve, you need a reason to deploy it--i.e., new demand. Use policy to push data centers to go clean (they can afford it). It would be like the IRA on easy mode.
3 - If you want to reindustrialize and create American manufacturing jobs, use the fact that data centers are massive industrial customers and already driving an American manufacturing boom to kick it into overdrive. The FT piece cites new research I published with @SIEPR showing data centers already support 300K permanent manufacturing jobs. We should want more!
The data center jobs debate that says the jobs are temporary misses upstream supply chain employment; the challenge is these jobs aren't in the communities where data centers are built. Use policy (tax credits, tariffs, permitting rules) to incentivize more domestic manufacturing and share these benefits.
4 - If you want to bring China to the table on AI safety, we need leverage. An overwhelming compute advantage helps with that, but we need to deploy compute. I'd rather race on compute than model capabilities. As I say in the piece, trying to stop data center construction as a safety measure would be like trying to solve traffic fatalities by shutting down car factories--vehicle and highway safety rules are obviously superior, do that instead!
ft.com/content/a1b2bdd8-212c…
Can confirm @camillericketts @zebriez are legends 🙏
We're building this but for frontier energy @DanielleFong
Stay tuned ☀️
For the past two years, some folks in Texas have been creating a new hardware city out in the middle of nowhere.
It's called Proto-Town, and it's both weird and wonderful. There are startups and people living in trailers right next to their machines.
They've never let the media visit. But we managed to get our cameras in there. Come see
m.youtube.com/watch?v=qIob2-…
Andy Rapista retweeted
Every company that matters will become a compute company
Andy Rapista retweeted
cyan banister, gp @LongJourneyVC ($450M fund), turned 50k to 500M by investing early in uber
she was also early on space x, anduril, crusoe, deepmind, and uber
her secret? the biggest opportunities often start where others see something weird
Andy Rapista retweeted
Jensen reveals he first met the Founder of TSMC by cold calling their number off Yellow Pages
"I heard about this company in Taiwan, TSMC and UMC at the time. They were of course quite poor companies themselves, and they were barely making it."
"TSMC was probably about $300 million, something like that. It's a tiny little company. I heard about Morris Chang. And so I tried to call their US office."
"Back in the old days there were yellow pages. I looked it up, TSMC. There's a number there. I called their number and it's a recording. I left my name and my number and said a few things about NVIDIA: we're a new startup company building computer graphics chips. I called several times. Nobody called me back."
"And so I finally wrote a letter to Morris Chang. Didn't hear anything for a long time, and then one day it was 6 o'clock or something, this person called and said 'this is Morris Chang from TSMC.'"
"And meanwhile, everybody outside my office was playing video games and they're all laughing and screaming. 'Hey shut up! It's Morris!' I was so happy that the founder would actually call us back, because hey, maybe we won't have to die."
"And so we talked and he invited me to Taiwan. I came back to Taiwan, and that was the start of everything. Not just that partnership, but just the whole industry."
Andy Rapista retweeted
neoclouds that figure out how to be power-first (not just gpu-first) will do much better than the ones that stay nvidia resellers
this will help them get customers and give them freedom (both in not being locked into just nvidia, and finally not depending on a landlord for their data center)
they're following the playbook of the ones who figured this out first, @CrusoeAI and @IREN_Ltd
now we see other neoclouds trying to do the same:
- @CoreWeave tried to buy @Core_Scientific for $9B, partly to get more power and sites (core scientific’s shareholders voted no)
- @nebiusai is building its own data centers in missouri, pennsylvania, and finland instead of renting space
it's the same reason inference companies want to own their gpus instead of renting them. for neoclouds, owning power matters just as much
- it will get them customers faster. AI companies need more and more compute and will do deals with whoever has compute. a neocloud that owns power won't need to wait
- it gives them the option to swap in new gpus, especially if they start using more than one kind of chip (which is already happening and will happen even more)
- no landlord! if you own the site, you’re not paying part of what you make to a landlord
Andy Rapista retweeted
I’m thrilled by the enthusiasm for nuclear energy in the U.S.
President Trump’s goal of expanding American nuclear energy is being celebrated across party lines and generations—especially among young Americans.
This is the American nuclear renaissance in action.
GOAT @WilliamBryk @jeffzwang !!! Congrats
Exa just passed $100m in annualized revenue.
We've grown almost exactly 10x per year for 3 years. We now handle 2 billion requests a month from 200k+ teams. And it’s still so early:
We predict agents will search 1000x more than humans by 2030.
Agents don't search like humans. They write long, complex queries and want all the world's data, not just the web. We've spent 5 years building a search engine designed for them.
We now serve personal agents, AI labs, coding agents, GTM agents, really any agent. Exa gives these agents access to 100s of billions of web documents and tons of proprietary data.
Our customers are increasingly noticing that information is the bottleneck to agent performance. While the world seems close to perfect intelligence with models like Fable and Astra, we are very far from perfect information.
To our small but legendary team, thank you for getting us here, from engineering to research to GTM to ops.
To our customers, it is an honor to serve you. We have some exciting releases coming soon and can't wait to share them 🫡
Andy Rapista retweeted
Energy capacity = economic capacity, in one chart
There are no low electricity, rich countries
"Energy Capacity = Economic Capacity"
Former Intel CEO Pat Gelsinger on why America's power grid is the ceiling on growth:
"We went through 10 to 15 years where essentially I was taking coal offline at the rate I was adding renewables, and essentially the nation was flatlined in terms of overall energy capacity."
"So essentially our economic capacity as a nation, flat for fifteen years."
"Even with that, we've gone to maybe increasing our national energy capacity 4% per year. Going from essentially zero to one."
"Nuclear as base load is a fabulous thing that we want to go build more of. Unfortunately all of our renewables have deep dependencies in China."
"Energy capacity somewhat dampens how euphoric we can get on AI. Why build a new data center and buy the million GPUs if I can't power them?"
@PGelsinger @RaghuRaghuram @appenz
Compute trading LFG
@sfcompute and @jumptrading are co-hosting an AIE NYC side event focused on markets, models, and martinis. If you're a quant, compute trader, or AI/ML researcher in town, come join us for steak and cocktails.
luma.com/w08n0yov
Andy Rapista retweeted
Come meet PR tech queen, @lulumeservey, and get her secrets on how to build comms and strategy for a ripping tech company. RSVP in comments.
Andy Rapista retweeted
Solar and batteries are not only becoming the biggest energy sources, they are also becoming hugely lucrative markets.
We will see a consolidation in solar and batteries that will be similar to how Nat Gas consolidated in the late 2010s.
Many new big energy companies will be birthed over the next several years.
The US grid plans to add 86 gigawatts this year. Here's where it's coming from:
- Solar: 43.4 GW
- Batteries: 24.3 GW
- Wind: 11.8 GW
- Gas: 6.3 GW
- New nuclear: 0 GW
The AI labs are planning for hundreds of gigawatts. Solar is the one source already produced at that scale, and most of it is made in China.
"Scaling firm power that isn't gas is much more complicated. Uprates and restarts can squeeze a little more from the existing nuclear fleet, but the real upside is new reactors."
a16z's @rmcentush on how America powers the next wave of growth: a16z.news/p/why-texas-is-mak…
I have been dreaming with @cyantist about turning San Francisco into a playground of experiences that create joy.
Meow Wolf meets Disneyland, with the soul of summer camp.
We call it JoyCon. The first experiment is in January. Limited tickets 🚀.
luma.com/joycon
