Baker quantifies Intel's manufacturing lead as 15-20% advantage that Taiwan Semi overtook in 2018.
“And in 2018, and it's important that the Taiwan Semi seven nanometer node is equivalent to the Intel 10 nanometer node.”
Baker says Nasdaq returns in 2021 were dominated by Google, Microsoft, Nvidia, and Tesla, requiring 60% concentration.
“The market return, particularly for the Nasdaq, was really dominated by a few stocks, Google, Microsoft, Nvidia, Tesla.”
Baker explains xAI achieved coherence in a 100,000 GPU cluster with Hopper without next-generation networking technologies.
“And because of that, they're able to get enough density that they could effectively make a 100,000 GPU cluster coherent even with Hopper without next generation networking technologies from NVIDIA, Broadcom, and others.”
Baker reveals high bandwidth memory is a bigger cost component for NVIDIA than Taiwan Semi, with only two capable manufacturers.
“High bandwidth memory is a bigger part of NVIDIA's cogs on GPUs than Taiwan Semi is. And today, there's two companies that can make it, Hynix and Micron.”
Baker explains Google pays Broadcom a 50 to 55 percent gross margin for back-end chip design and Taiwan Semi management.
“And the managing Taiwan Semi is like stamping out that house like Lennar or, you know, D. R. Horton. And for doing those two latter parts, Broadcom earns a 50 to 55% gross margin.”
Baker predicts xAI will release the first Blackwell-trained model in early 2026 because Elon builds fastest.
“One, we will see the first models trained on Blackwell in early twenty twenty six. I think the first Blackwell model will come from xAI.”
Baker explains Blackwell racks weigh 3,000 pounds and consume 130 kilowatts, equivalent to 130 American homes.
“The rack goes from weighing round numbers, a thousand pounds to 3,000 pounds. Goes from round numbers, 30 kilowatts, which is 30 American homes to 130 kilowatts, which is 130 American homes.”
Baker predicts Blackwell models will be amazing because pre-training scaling laws remain intact.
“We know that scaling laws for pre training are intact, And this means the Blackwell models are going to be amazing.”
Baker details Blackwell transition: racks triple in weight and quadruple in power from 30 to 130 kilowatts.
“Goes from round numbers, 30 kilowatts, which is 30 American homes to 130 kilowatts, which is 130 American homes.”
“And the reason for that is it was by far the most complex product transition we've ever gone through in technology. Going from hopper to Blackwell, first you go from air cooled to liquid cooled.”
Baker explains Blackwell's complexity created deployment challenges that reasoning models had to bridge.
“Had reasoning not come along, there would have been no AI progress from mid twenty twenty four through essentially”
Baker argues Google has a temporary pre-training advantage as the lowest cost producer of tokens using advanced TPUs.
“So I think Google for sure has this temporary advantage right now from a pre training perspective. I think it's also important that they've been the lowest cost producer of tokens.”
Baker details Blackwell's complexity: racks tripled in weight and quadrupled in power consumption versus Hopper.
“Going from hopper to Blackwell, first you go from air cooled to liquid cooled. The rack goes from weighing round numbers, a thousand pounds to 3,000 pounds.”
Baker states only NVIDIA and Amazon Trainium have functioning switched scale-up networks for inference today.
“And the only two functioning switched scale up networks in the world today are the ones that power NVIDIA GPUs and Amazon's Trainiums.”
Baker believes NVIDIA could sell $11.5 trillion worth of chips next year if TSMC expanded capacity sufficiently.
“And if they doubled or tripled capacity, like, NVIDIA could probably sell $11,520,000,000,000 worth of chips next year. I really believe that.”
Baker says Jensen Huang visits Taiwan Semi quarterly wanting capacity doubled or tripled but they only expand 5%.
“And so they're just simply not expanding capacity as fast as Jensen wants. Jensen goes there every three months and, you know, maybe they expand, you know, 5%. He wants them to double or triple.”
Baker believes NVIDIA could sell $11.52 trillion in chips next year if TSMC tripled capacity but bubble risks prevent it.
“And if they doubled or tripled capacity, like, NVIDIA could probably sell $11,520,000,000,000 worth of chips next year. I really believe that. But the other side of that might be very painful for everyone.”
Baker predicts Trainium will dominate 2026 like TPUs did in 2025, with Trainium 3 ramping in second half.
“Tranium is going to be to 2026, especially in the second half of this year when Tranium three really ramps, as TPUs were to twenty twenty five.”
Baker believes NVIDIA could sell $1.5-2 trillion in chips if TSMC expanded capacity, but constraints prevent bubble.
“NVIDIA could probably sell $11,520,000,000,000 worth of chips next year. I really believe that. But the other side of that might be very painful for everyone.”
Baker argues Trainium is most underestimated because frontier mixture-of-expert models require switched scale-up networks for inference.
“And so Trainium is for sure the most underestimated, not only because of those design choices, but because the all of these frontier models are what are called mixture of expert models.”
Baker predicts orbital compute will solve power shortages in five to seven years but wafer shortages will persist longer.
“We're gonna address the watt shortages with orbital compute for sure in the next five to seven years. But the wafer shortage, I think, is going to persist for a long time.”
Baker reveals Jensen Huang has never had a contract with TSMC, operating entirely on handshakes and fairness.
“I do think it's wild that Jensen has never had a contract with Taiwan Semi. They do business on what seems fair in handshakes.”
Baker argues NVIDIA could sell $2-3 trillion in GPUs if TSMC expanded capacity, but TSMC's restraint prevents a bubble.
“If Taiwan Semi did what Jensen wanted, I think Nvidia could sell $2,000,000,000,000 of GPUs in '26 or '27, maybe 2,500,000,000,000, maybe 3,000,000,000,000. But there is a limit where consumers would consume so much.”
Baker warns VCs that Jensen sees Taiwan Semi processes before any startup claiming special access.
“If you're a VC and you're investing a semiconductor company that is telling you they are going to have an advantage because of a Taiwan Semi process that they have special access to, I promise you the Jensen saw that process when it was a twinkle in Taiwanese eyes.”
Baker argues Taiwan Semi's capacity constraint prevents AI bubble; NVIDIA could sell $2-3T GPUs otherwise.
“If Taiwan Semi did what Jensen wanted, I think Nvidia could sell $2,000,000,000,000 of GPUs in '26 or '27, maybe 2,500,000,000,000, maybe 3,000,000,000,000.”
Baker predicts GPUs will have 10-15 year useful lives due to prefill-inference disaggregation, extending older chips' value.
“The disaggregation of inference means that I think these GPUs are going to have ten or fifteen year lives.”
Baker argues GPU useful lives will extend to 10-15 years due to inference disaggregation, contradicting AI skeptics.
“The disaggregation of inference means that I think these GPUs are going to have ten or fifteen year lives. The AI skeptics are like, oh, these companies are all cooking their books.”
Baker argues GPU useful lives will extend to 10-15 years due to prefill/decode disaggregation, not 1-2 years.
“The useful life of GPU is only a year or two. The useful life of CPU is only four years because the rapid technological change.”
Baker says Jensen Huang was obviously exceptional when they first met in early 2000s.
“I was a pretty young man. It was obvious to me he was exceptional, like for sure one of the most exceptional people I'd ever met.”
Baker identifies Jensen as one of the two or three most exceptional people he's ever met in his life.
“But like now, for sure, Jensen is, you know, one of the two or three most exceptional people I've ever met in my life.”
Baker argues game theory suggests memory LTAs are durable and Nvidia may have significant cloud revenue shares
“Game theory suggests memory LTAs are likely to be durable. And yes, Nvidia may have a significant cloud business via revenue shares. https://t.co/GFN1kDQmSk”
Baker argues game theory suggests memory LTAs are durable and Nvidia may have significant cloud revenue shares
“Game theory suggests memory LTAs are likely to be durable. And yes, Nvidia may have a significant cloud business via revenue shares. https://t.co/GFN1kDQmSk”
Baker cites market data showing AI chip stocks lost over one trillion dollars in July 2026 with semiconductor index down over twenty percent.
“ Note: Independent market data corroborates this framing. CNBC reported that AI chip stocks shed more than $1 trillion in the July 2026 selloff, with the Philadelphia Semiconductor Index (SOX) losing more than 20% from its June peak after a roughly 130% rally over the prior twelve months. Separately, Yahoo Finance reported that NVIDIA's valuation hit a multiyear low, with its forward P/E falling to levels last seen in 2019, even as the company continued to post record revenue. The market's verdict could not be clearer. ”
Baker suggests memory companies should copy NVIDIA's credit wrapper model to participate in GPU revenue streams as credit markets revolt.
“Which is? I would be going to the buyers of GPUs, tradiums, and whoever and say, I'll participate in the NVIDIA credit wrapper.”
Baker notes NVIDIA trades at its lowest forward PE in ten years.
“NVIDIA is actually, as we record this, at its lowest forward PE of the last ten years.”
Baker says no one predicted old GPU prices would go vertical, everyone expected gradual declines.
“I don't think anyone in '24 or '25 thought that the prices of old GPUs would be going vertical. Everybody thought, hey. We're gonna be smart. We're gonna sign these long term contracts.”
Baker reports GPU rental prices doubled from mid-$2 to nearly $4 per hour over seven months for identical clusters.
“And they had rented a cluster of several thousand black wells, and we'll just call it somewhere in the mid $2 per GPU hour.”
Baker reports inference cloud expects to pay 100% more for Blackwells when contracts expire, showing hyperscalers are underearning.
“They went on a podcast, and they essentially said, we are planning to pay 100% more for Blackwell's when our contract expires. And that just means that essentially all the hyperscalers are under earning.”
Baker identifies four companies that matter at scale in long-term supply agreements for AI chips.
“Let's just think about the game theory of breaking an LTA. So there's four companies that matter at scale. There's Amazon with their tradiums.”
Baker argues only four companies matter at scale for AI chips: Amazon, Google, AMD, and NVIDIA.
“So there's four companies that matter at scale. There's Amazon with their tradiums. There's Google with their TPUs. There's AMD, and then there's NVIDIA who's, like, much bigger than everybody else combined.”
Baker suggests memory companies should copy NVIDIA's credit wrapper model despite business instability risks.
“I would be going to the buyers of GPUs, Traniums, and whoever and say, I'll participate in the NVIDIA credit wrapper. Now their business is just inherently less stable and predictable,”
Baker suggests memory companies should copy NVIDIA's credit wrapper model while they have cash and credit markets are tight.
“I'm just making this up. But, like, do something. Like, you can because you have money now and credit markets are revolting.”
“Delighted to announce @JensenHuang as the first confirmed speaker for the @TomorrowXSummit and even happier that I will be interviewing him. We will have Chicken Fried Steak at the summit. Sawyer’s is my favorite place in Austin but we will of course have Denny’s. https://t.co/4QKBMGh4TU”
Baker says Jensen was number one on Glassdoor best CEOs, roughly half Nvidia employees worth over twenty-five million, 75% millionaires
“Jensen was #1 on Glassdoor’s list of the best CEOs to work for. Still the same guy he was when I met him 27 years ago. Maybe slightly calmer. The ultimate testament is that roughly half of the employees at Nvidia are worth over $25 million, 75% are millionaires and essentially https://t.co/iS2zgYE0jI”
Baker predicts every accelerator company will do some variant of 3D DRAM or zHBM discussed at HotChips
“Every accelerator company, and I mean every accelerator company, is likely going to do some variant of 3D DRAM or the “zHBM” which were both discussed at HotChips today. The larger companies will probably run (are running) parallel programs, but these two memory variants”
Baker says GPU price increase means anyone spending on Blackwell and Rubin before January 2027 gains competitive advantage
“The most obvious implication of GPU price increase is that anyone who is spending a lot on Blackwell and Rubin before January 31, 2027 is in a good relative position vs. competitors. Super curious to know if GPU prices are going up more than rack prices.”
Baker says real comparison will be Jalapeño versus Rubin Ultra and Feynmann given current shipping timelines
“Should note the real comparison will be with Rubin Ultra given Rubin shipping today so kinda have to wait. And Feynmann coming fast. But still impressive debut from Jalapeño.”
Baker credits Jalapeño as first good ASIC outside TPU and Trainium but says disaggregated GPU plus SRAM will outperform
“Impressive that Jalapeño outperforms the comparable TPU and is in the mix with Rubin. Credit where credit is due - first good ASIC outside of TPU/Trainium. However, will likely significantly underperform a disaggregated GPU/Trainum plus SRAM accelerator setup. Especially with”
Baker notes Jalapeño participated in third-party benchmarks before TPUs did, suggesting competitive confidence
“Also pretty funny that Jalapeño participated in a third party benchmark *before* TPUs! Tells you a lot!”
Baker suggests NVIDIA accelerators have residual value that distinguishes them from first-party ASICs, responding to concerns about financing OpenAI hardware
“@dylan522p @TShirtnJeans2 @maxkan Some accelerators (the green ones) are not like the other accelerators.”
Baker suggests NVIDIA accelerators have residual value that distinguishes them from first-party ASICs, responding to concerns about financing OpenAI hardware
“@dylan522p @TShirtnJeans2 @maxkan Some accelerators (the green ones) are not like the other accelerators.”