Gavin Baker
Founding Partner and CIO of Atreides Management · Former Fidelity Investments portfolio manager · Former Micron analyst (year 2000) · investor (Atreides Management) · CIO, Atreides Management
“I think capitalism is going to solve the Watts shortage absent big regulatory or political blowback... I think the Watts shortage will probably begin to alleviate 27, 28 and then I think Orbital Compute will really solve that.”
“I do want to reframe Orbital Compute because I think when people hear Data Centers in Space... they picture a Pentagon sized building in space. They're like, well, we can't do that. That's not what it is. A Blackwell rack weighs 3,000 pounds, it's 8ft high, it's 4ft deep, 3ft wide. It's racks in space... You keep it in a sun synchronous orbit. So those solar panels are always in the sun.”
“Starship is going to change the space economy in ways we cannot imagine. And particularly if regulation becomes a constraint to data centers, none of it's going to matter. You're going to sell as much orbital compute as you can make.”
“Capitalism is hard at work on Watts on wafers though. It's just this group of flinty older humans in Taiwan who are the most important humans in Taiwan, whatever they are. The overwhelming fraction of the country's GDP, water usage, electricity usage.”
“If Taiwan Semi did what Jensen wanted, I think Nvidia could sell $2 trillion of GPUs in 26 or 27, maybe 2.5 trillion, maybe 3 trillion. But there is a limit where consumers would consume so much they probably would be in an overbuild. So Taiwan Semi. If we don't get a bubble, we need to throw a party for them because they will have single handedly prevented a bubble.”
“The history of markets is I don't know who but one of intel and Samsung. They're not going to stay disciplined. They will break and then at some level that will force everyone else to break. I think a lot of this may come down to the degree to which Taiwan Semi can maintain a lead over intel and Samsung... there's a Goldilocks zone where they expand enough they make it hard for intel or Samsung to really truly emerge as a at scale second source with something well north of 30% market share. And yet they also keep this fundamental constraint on wafers that helps us avoid a bubble.”
“It's a SpaceX. I believe Tesla's involved as well. Joint venture to build the world's largest fab here in America. I think they're going to be successful. One they have a partnership with intel which is very important because they're getting access to 50 years of institutional knowledge... It's also an advantage that I believe that Terrafab is going to get attention from the A teams, all the Semi cap equipment companies. One big reason Taiwan Semi caught up is ASML and KLA Tin Core and LAM Research and Applied Materials. They wanted them to catch up. They don't like having a monopsony. The A teams were in Taiwan working Intel made some mistakes and presto.”
“Anthropic, they added $11 billion of arrangement. And what is astonishing to me about this is that the SaaS and Cloud Revolution... created between 5 and 10 trillion dollars of value, I would say arguably the three highest profile SaaS companies in the last 10-12 years are Palantir, Snowflake and Databricks. And these three companies employ thousands of people, tens of thousands collectively. They've all spent 10 years building their businesses. And Anthropic added their combined businesses in one month. Nothing like that has ever happened in the history of capitalism.”
“One thing that's really good about the current build out is it's still overwhelmingly funded out of operating cash flows, which is a really important fundamental difference versus the year 2000 has is valuation has is the fact that every GPU is running at 100% utilization when 99% of fiber was unutilized.”
“I would say based on every memory cycle we have had for the last 25 years, this is the time to be selling memory 100%. I was actually the Micron analyst in the year 2000... I'm a veteran of many, many memory cycles and based on history, this is the time to sell. However, there's one cycle where you absolutely do not want to sell and that's the cycle we had in the mid-90s, which is the last true capacity cycle that I would argue we've had in memory. And based on that cycle, we may still be very early.”
“If OpenAI and Anthropic are at, call it $100 billion of ARR now with 80% ish gross margins on inference, like the returns are there. And then if we add in Gemini, we add in Cursor, we add in Xai, we add an open source, it's not hard to see 200, 300, $400 billion of ARR the end of this year at high margin across all of that.”
“I do think what Karpathy is working on, recursive self improvement is really important and unlocking that and continual learning, you know, maybe the two final frontiers for AI... If that comes to pass, [the 10x-per-year improvement rate] might seem conservative... continual learning is the holy grail, where the model learns from experiences the way humans do. And that's something we haven't unlocked yet. And those two combined, I think they might pull the future forward in a very real way.”
“[Anthropic has] a decent lead on everybody else, whether it's three months or six months. Obviously they're probably six 12 months ahead of open source. Maybe they're three, six, nine months ahead of their contemporaries, but they have a lead.”
“Long only mutual funds... They all can, per SEC rules, allocate up to 15%... When a company goes public and lockup expires, it moves out of that bucket. So this is going to be hundreds of billions of dollars of new late stage demand”
“From a pure risk reward perspective, I thought MGM was the best. Your downside is really capped because of the Barry Diller bid... And I do think talent [Talen] is also a very compelling risk reward. I just think everything in AI is going to need to grapple with increasing regulatory risk... the big negative externality for talent is nothing to do with talent.”
“xai's deal with Google for cloud computing generates more operating profit per gigawatt than Anthropic, than Meta, than Google, than OpenAI.”
“We do know from Jensen that Elon brings data centers up faster than anyone. 122 days speed is literally cost because every day you're paying electricians and plumbers that's cost.”
“In 30 days we went from not being an AI hyperscaler to being number four. And we passed a lot of companies, including Oracle.”
“I think we end this year well over 200 billion in inference revenue. Well over.”
“she calculated a 55% ARR on Colossus 1”
“It's actually really bullish for compute and hardware because if the frontier models are capturing less of the margin then you're going to spend more on compute.”
“because space power cooling are effectively free in space.”
“as long as these satellites in space aren't failing at an astronomical rate, the math, maths. By the way, we know GPUs melt and lasers fail.”
“But my understanding is that Cursor and Anthropic have more tokens of proprietary coding data than anyone else.”
“And then the cursor data is being injected into the pre training process, not just reinforcement learning.”
“he called it bitter lesson adjacent that coding may be the fastest path to AGI”
“nobody has run Mythos for a year continuously.”
“memory capacity and bandwidth are foundational to the performance of every AI model. So this is the most important bottleneck.”
“memory is DRAM is probably going to be 30 to 40% of all hyperscaler capex next year. Hundreds of billions of dollars... going straight to dram.”
“for the DRAM you need in these AI servers there are three companies that can make it. It's really hard to do this is as close to magic as science can get.”
“They announced that they have these SCAs, these supply chain agreements that have a floor and a ceiling for prices... this covers essentially 50% of their revenue... And the floor pricing in these new contracts is ahead of prior cycle peaks from a gross margin perspective.”
“CXMT is going public in China. They may be the cure for Apple's ills. They will flood the market with to some degree cheap consumer grade dram. But for the DRAM you need in these AI servers there are three companies that can make it.”
“these stocks still trade are cross sectionally cheap relative to the rest of AI.”
“the disaggregation of inference into pre fill and decode... you can lift H1 hundreds, A1 hundreds out of some old data center, put them in one of these mega megapods... put a Grok or a Cerebrus in front of it and you can get a very competitive solution... We're going to be using GPUs for seven years, 10 years, 12 years. And that's great because it lowers the cost to finance them, which makes this AI revolution more financeable.”
“to stand up a 1 gigawatt data center, $35 billion in semiconductors... and it's $25 billion of power and cooling equipment... clearly inflationary because a lot of that 25 billion is the human labor... when starship is reusable, it's going to cost $5 billion to put a gigawatt of compute into space.”
“what I'm focused on as an investor. How many megawatts can they bring on?”
Gavin Baker
One-line summary: Atreides CIO; coined the 'watts and wafers' framing for AI's two binding physical constraints; explicitly endorses the orbital-compute thesis ('racks in space, not data centers'), TSMC-as-bubble-preventer mechanism (TSMC capacity discipline keeps Intel/Samsung in line), and Terafab as a credible third foundry leg. Memory-cycle veteran sees current cycle as the 1990s analog (true capacity cycle), not a typical cyclical top.
What they're known for
Brief factual context — fill in.
Why they matter to stock-market
Why this person's claims are tracked here — fill in.
Said
Speaker-attributed claims extracted from diarized sources. Each bullet mirrors one entry in quotes: frontmatter — keep them in sync.
-
On ai-capex-to-power-and-materials-cascade, terrestrial-power-flat-to-orbital-dc-arbitrage:
"I think capitalism is going to solve the Watts shortage absent big regulatory or political blowback... I think the Watts shortage will probably begin to alleviate 27, 28 and then I think Orbital Compute will really solve that." — 2026-05-20-podcast-invest-like-the-best-gavin-baker-watts-and-wafers-invest-like-the-best (2026-05-20)
-
On terrestrial-power-flat-to-orbital-dc-arbitrage:
"I do want to reframe Orbital Compute because I think when people hear Data Centers in Space... they picture a Pentagon sized building in space. They're like, well, we can't do that. That's not what it is. A Blackwell rack weighs 3,000 pounds, it's 8ft high, it's 4ft deep, 3ft wide. It's racks in space... You keep it in a sun synchronous orbit. So those solar panels are always in the sun." — 2026-05-20-podcast-invest-like-the-best-gavin-baker-watts-and-wafers-invest-like-the-best (2026-05-20)
-
On terrestrial-power-flat-to-orbital-dc-arbitrage:
"Starship is going to change the space economy in ways we cannot imagine. And particularly if regulation becomes a constraint to data centers, none of it's going to matter. You're going to sell as much orbital compute as you can make." — 2026-05-20-podcast-invest-like-the-best-gavin-baker-watts-and-wafers-invest-like-the-best (2026-05-20)
-
On tsmc-capacity-shortfall-and-pricing-power, tsmc-saturation-to-intel-anchor-stack:
"Capitalism is hard at work on Watts on wafers though. It's just this group of flinty older humans in Taiwan who are the most important humans in Taiwan, whatever they are. The overwhelming fraction of the country's GDP, water usage, electricity usage." — 2026-05-20-podcast-invest-like-the-best-gavin-baker-watts-and-wafers-invest-like-the-best (2026-05-20)
-
On tsmc-capacity-shortfall-and-pricing-power, tsmc-saturation-to-intel-anchor-stack:
"If Taiwan Semi did what Jensen wanted, I think Nvidia could sell $2 trillion of GPUs in 26 or 27, maybe 2.5 trillion, maybe 3 trillion. But there is a limit where consumers would consume so much they probably would be in an overbuild. So Taiwan Semi. If we don't get a bubble, we need to throw a party for them because they will have single handedly prevented a bubble." — 2026-05-20-podcast-invest-like-the-best-gavin-baker-watts-and-wafers-invest-like-the-best (2026-05-20)
-
On tsmc-capacity-shortfall-and-pricing-power, us-fab-capacity-bottleneck, tsmc-saturation-to-intel-anchor-stack:
"The history of markets is I don't know who but one of intel and Samsung. They're not going to stay disciplined. They will break and then at some level that will force everyone else to break. I think a lot of this may come down to the degree to which Taiwan Semi can maintain a lead over intel and Samsung... there's a Goldilocks zone where they expand enough they make it hard for intel or Samsung to really truly emerge as a at scale second source with something well north of 30% market share. And yet they also keep this fundamental constraint on wafers that helps us avoid a bubble." — 2026-05-20-podcast-invest-like-the-best-gavin-baker-watts-and-wafers-invest-like-the-best (2026-05-20)
-
On terafab, us-fab-capacity-bottleneck, picks-and-shovels-leading-edge-fab-buildout:
"It's a SpaceX. I believe Tesla's involved as well. Joint venture to build the world's largest fab here in America. I think they're going to be successful. One they have a partnership with intel which is very important because they're getting access to 50 years of institutional knowledge... It's also an advantage that I believe that Terrafab is going to get attention from the A teams, all the Semi cap equipment companies. One big reason Taiwan Semi caught up is ASML and KLA Tin Core and LAM Research and Applied Materials. They wanted them to catch up. They don't like having a monopsony. The A teams were in Taiwan working Intel made some mistakes and presto." — 2026-05-20-podcast-invest-like-the-best-gavin-baker-watts-and-wafers-invest-like-the-best (2026-05-20)
-
On ai-capex-to-power-and-materials-cascade:
"Anthropic, they added $11 billion of arrangement. And what is astonishing to me about this is that the SaaS and Cloud Revolution... created between 5 and 10 trillion dollars of value, I would say arguably the three highest profile SaaS companies in the last 10-12 years are Palantir, Snowflake and Databricks. And these three companies employ thousands of people, tens of thousands collectively. They've all spent 10 years building their businesses. And Anthropic added their combined businesses in one month. Nothing like that has ever happened in the history of capitalism." — 2026-05-20-podcast-invest-like-the-best-gavin-baker-watts-and-wafers-invest-like-the-best (2026-05-20)
-
On ai-capex-to-power-and-materials-cascade:
"One thing that's really good about the current build out is it's still overwhelmingly funded out of operating cash flows, which is a really important fundamental difference versus the year 2000 has is valuation has is the fact that every GPU is running at 100% utilization when 99% of fiber was unutilized." — 2026-05-20-podcast-invest-like-the-best-gavin-baker-watts-and-wafers-invest-like-the-best (2026-05-20)
-
"I would say based on every memory cycle we have had for the last 25 years, this is the time to be selling memory 100%. I was actually the Micron analyst in the year 2000... I'm a veteran of many, many memory cycles and based on history, this is the time to sell. However, there's one cycle where you absolutely do not want to sell and that's the cycle we had in the mid-90s, which is the last true capacity cycle that I would argue we've had in memory. And based on that cycle, we may still be very early." — 2026-05-20-podcast-the-compound-and-friends-it-s-a-wave-not-a-bubble-nvidia-preview-google-s (2026-05-20)
-
On ai-capex-to-power-and-materials-cascade:
"If OpenAI and Anthropic are at, call it $100 billion of ARR now with 80% ish gross margins on inference, like the returns are there. And then if we add in Gemini, we add in Cursor, we add in Xai, we add an open source, it's not hard to see 200, 300, $400 billion of ARR the end of this year at high margin across all of that." — 2026-05-22-podcast-all-in-podcast-spacex-s-2t-case-nvidia-s-shock-selloff-america (2026-05-22)
-
On ai-capex-to-power-and-materials-cascade:
"I do think what Karpathy is working on, recursive self improvement is really important and unlocking that and continual learning, you know, maybe the two final frontiers for AI... If that comes to pass, [the 10x-per-year improvement rate] might seem conservative... continual learning is the holy grail, where the model learns from experiences the way humans do. And that's something we haven't unlocked yet. And those two combined, I think they might pull the future forward in a very real way." — 2026-05-22-podcast-all-in-podcast-spacex-s-2t-case-nvidia-s-shock-selloff-america (2026-05-22)
-
On (no topic linked):
"[Anthropic has] a decent lead on everybody else, whether it's three months or six months. Obviously they're probably six 12 months ahead of open source. Maybe they're three, six, nine months ahead of their contemporaries, but they have a lead." — 2026-05-22-podcast-all-in-podcast-spacex-s-2t-case-nvidia-s-shock-selloff-america (2026-05-22)
-
On ipo-comeback-public-market-value-capture, mega-issuance-supply-wave:
"Long only mutual funds... They all can, per SEC rules, allocate up to 15%... When a company goes public and lockup expires, it moves out of that bucket. So this is going to be hundreds of billions of dollars of new late stage demand" — 2026-06-07-podcast-all-in-podcast-inside-the-private-stock-market-boom-spacex (2026-06-07)
-
On mgm-diller-bid-floor-japan-dubai, talen-energy-replacement-cost-pjm-scarcity:
"From a pure risk reward perspective, I thought MGM was the best. Your downside is really capped because of the Barry Diller bid... And I do think talent [Talen] is also a very compelling risk reward. I just think everything in AI is going to need to grapple with increasing regulatory risk... the big negative externality for talent is nothing to do with talent." — 2026-06-12-podcast-all-in-podcast-all-in-s-best-ideas-pitch-competition-4-investors (2026-06-12)
-
On operating-profit-per-gigawatt, elon-web-services-to-spacex-hyperscaler-rerate:
"xai's deal with Google for cloud computing generates more operating profit per gigawatt than Anthropic, than Meta, than Google, than OpenAI." — 2026-06-11-podcast-bg2-pod-the-spacex-ipo-fable-5-ai-capex-update-market (2026-06-11)
-
On operating-profit-per-gigawatt:
"We do know from Jensen that Elon brings data centers up faster than anyone. 122 days speed is literally cost because every day you're paying electricians and plumbers that's cost." — 2026-06-11-podcast-bg2-pod-the-spacex-ipo-fable-5-ai-capex-update-market (2026-06-11)
-
On elon-web-services-to-spacex-hyperscaler-rerate:
"In 30 days we went from not being an AI hyperscaler to being number four. And we passed a lot of companies, including Oracle." — 2026-06-11-podcast-bg2-pod-the-spacex-ipo-fable-5-ai-capex-update-market (2026-06-11)
-
On ai-inference-revenue-run-rate-dispute:
"I think we end this year well over 200 billion in inference revenue. Well over." — 2026-06-11-podcast-bg2-pod-the-spacex-ipo-fable-5-ai-capex-update-market (2026-06-11)
-
On ai-inference-revenue-run-rate-dispute, compute-utilization-overhang-as-latent-supply:
"she calculated a 55% ARR on Colossus 1" — 2026-06-11-podcast-bg2-pod-the-spacex-ipo-fable-5-ai-capex-update-market (2026-06-11)
-
On open-source-share-shift-bullish-for-compute:
"It's actually really bullish for compute and hardware because if the frontier models are capturing less of the margin then you're going to spend more on compute." — 2026-06-11-podcast-bg2-pod-the-spacex-ipo-fable-5-ai-capex-update-market (2026-06-11)
-
On rapid-reusability-to-orbital-compute-capex-arbitrage:
"because space power cooling are effectively free in space." — 2026-06-11-podcast-bg2-pod-the-spacex-ipo-fable-5-ai-capex-update-market (2026-06-11)
-
On rapid-reusability-to-orbital-compute-capex-arbitrage:
"as long as these satellites in space aren't failing at an astronomical rate, the math, maths. By the way, we know GPUs melt and lasers fail." — 2026-06-11-podcast-bg2-pod-the-spacex-ipo-fable-5-ai-capex-update-market (2026-06-11)
-
On proprietary-coding-data-as-moat:
"But my understanding is that Cursor and Anthropic have more tokens of proprietary coding data than anyone else." — 2026-06-11-podcast-bg2-pod-the-spacex-ipo-fable-5-ai-capex-update-market (2026-06-11)
-
On proprietary-coding-data-as-moat:
"And then the cursor data is being injected into the pre training process, not just reinforcement learning." — 2026-06-11-podcast-bg2-pod-the-spacex-ipo-fable-5-ai-capex-update-market (2026-06-11)
-
On proprietary-coding-data-as-moat:
"he called it bitter lesson adjacent that coding may be the fastest path to AGI" — 2026-06-11-podcast-bg2-pod-the-spacex-ipo-fable-5-ai-capex-update-market (2026-06-11)
-
On agi-definitions-and-benchmark-saturation:
"nobody has run Mythos for a year continuously." — 2026-06-11-podcast-bg2-pod-the-spacex-ipo-fable-5-ai-capex-update-market (2026-06-11)
-
On hbm-supply-bottleneck, hbm-cowos-as-binding-bottleneck:
"memory capacity and bandwidth are foundational to the performance of every AI model. So this is the most important bottleneck." — 2026-06-26-podcast-all-in-podcast-socialists-sweep-nyc-china-catches-up-in-coding (2026-06-26)
-
On hbm-cowos-as-binding-bottleneck:
"memory is DRAM is probably going to be 30 to 40% of all hyperscaler capex next year. Hundreds of billions of dollars... going straight to dram." — 2026-06-26-podcast-all-in-podcast-socialists-sweep-nyc-china-catches-up-in-coding (2026-06-26)
-
On hbm-cowos-as-binding-bottleneck, hbm-supply-bottleneck:
"for the DRAM you need in these AI servers there are three companies that can make it. It's really hard to do this is as close to magic as science can get." — 2026-06-26-podcast-all-in-podcast-socialists-sweep-nyc-china-catches-up-in-coding (2026-06-26)
-
On hbm-cowos-as-binding-bottleneck:
"They announced that they have these SCAs, these supply chain agreements that have a floor and a ceiling for prices... this covers essentially 50% of their revenue... And the floor pricing in these new contracts is ahead of prior cycle peaks from a gross margin perspective." — 2026-06-26-podcast-all-in-podcast-socialists-sweep-nyc-china-catches-up-in-coding (2026-06-26)
-
On hbm-cowos-as-binding-bottleneck:
"CXMT is going public in China. They may be the cure for Apple's ills. They will flood the market with to some degree cheap consumer grade dram. But for the DRAM you need in these AI servers there are three companies that can make it." — 2026-06-26-podcast-all-in-podcast-socialists-sweep-nyc-china-catches-up-in-coding (2026-06-26)
-
On hbm-cowos-as-binding-bottleneck:
"these stocks still trade are cross sectionally cheap relative to the rest of AI." — 2026-06-26-podcast-all-in-podcast-socialists-sweep-nyc-china-catches-up-in-coding (2026-06-26)
-
On inference-demand-to-wafer-scale-advantage, compute-utilization-overhang-as-latent-supply:
"the disaggregation of inference into pre fill and decode... you can lift H1 hundreds, A1 hundreds out of some old data center, put them in one of these mega megapods... put a Grok or a Cerebrus in front of it and you can get a very competitive solution... We're going to be using GPUs for seven years, 10 years, 12 years. And that's great because it lowers the cost to finance them, which makes this AI revolution more financeable." — 2026-06-26-podcast-all-in-podcast-socialists-sweep-nyc-china-catches-up-in-coding (2026-06-26)
-
On rapid-reusability-to-orbital-compute-capex-arbitrage, terrestrial-power-flat-to-orbital-dc-arbitrage:
"to stand up a 1 gigawatt data center, $35 billion in semiconductors... and it's $25 billion of power and cooling equipment... clearly inflationary because a lot of that 25 billion is the human labor... when starship is reusable, it's going to cost $5 billion to put a gigawatt of compute into space." — 2026-06-26-podcast-all-in-podcast-socialists-sweep-nyc-china-catches-up-in-coding (2026-06-26)
-
On inference-demand-to-wafer-scale-advantage:
"what I'm focused on as an investor. How many megawatts can they bring on?" — 2026-06-26-podcast-all-in-podcast-socialists-sweep-nyc-china-catches-up-in-coding (2026-06-26)
Sources
- 2026-05-20-podcast-invest-like-the-best-gavin-baker-watts-and-wafers-invest-like-the-best
- 2026-05-20-podcast-the-compound-and-friends-it-s-a-wave-not-a-bubble-nvidia-preview-google-s
- 2026-05-22-podcast-all-in-podcast-spacex-s-2t-case-nvidia-s-shock-selloff-america
- 2026-06-07-podcast-all-in-podcast-inside-the-private-stock-market-boom-spacex
- 2026-06-12-podcast-all-in-podcast-all-in-s-best-ideas-pitch-competition-4-investors
- 2026-06-11-podcast-bg2-pod-the-spacex-ipo-fable-5-ai-capex-update-market
- 2026-06-26-podcast-all-in-podcast-socialists-sweep-nyc-china-catches-up-in-coding