Quotesby GlyphCompute quotes
85 quotes from 34 people
- If the top U.S. labs have to give up on what is supposed to be really important, they will look very “catchable.” The policy creates a big incentive for some other parts of the world to invest a lot more in compute.
- A world with sufficient compute is one where intelligence is a commodity, and in commodity markets margin comes from a superior cost structure, which I would expect the leading model providers to have, in part because they have a meaningful lead in scaling and can apply superior AI to their infrastructure.
- I have my doubts that the labs can afford to spend a constant portion of this compute on internal R&D as the total volume goes up, especially with plans to IPO, as they face increased scrutiny on basic economics.
+1more from this blog
- If you have five hours, you’re probably going to do a lot better. The AI models are pretty similar.
- AI has reached its inflection point. It’s doing useful work. Its tokens are productive and profitable. Now, compute is revenue...
+1more from this filing
- More broadly, AI is driving a significant expansion in demand for compute across all of our markets, and our leadership portfolio and growing customer visibility position us exceptionally well to capture this expanding opportunity and deliver substantial revenue and earnings growth in the years ahead.
- I want to emphasize the key conclusion here: as AI models become smarter, they’ll better monetize the same amount of compute. If a true human-level software engineer that could run on an H100 equivalent, at current market rates for software engineers, that H100 should rent for over $250k a year.
+2more from this essay
- If by 2028 we’ve automated software engineering and the price of compute is 15x higher than it is right now, then it’s going to be much more difficult for you, with no revenue, to compete for compute against the frontier labs.
- If you’re spending most of your compute on inference, you’re kind of declaring that AI progress has stalled, because it’s not worth investing more in training, and your business is basically that of a cloud provider.
- Every quote from Dwarkesh Patel
- AI is driving unprecedented demand for compute, and as we continue to execute, Intel is well-positioned to capture sustainable growth across our CPU franchise, ASICs, advanced packaging and vast wafer foundry network...
- The narrow problem with that is that fiber construction was far ahead of demand, where AI compute construction is far behind demand, though, as above, we don’t know what that supply/demand balance will look like in the future.
+1more from this essay
- AI is entering a new era, and the countries that build the infrastructure for this transformation will shape the future of technology, industry and society...
- We delivered an outstanding first quarter, driven by accelerating demand for AI infrastructure, with Data Center now the primary driver of our revenue and earnings growth...
- Ultimately, our customers don’t want to buy computers, they want to build AI factories...
- While people will undoubtedly get more efficiency out of the compute they have, this artificial compute constraint may mean no one lab is able to break significantly ahead until 2028 at the soonest - re-enforcing an oligopoly market for LLMs.
- And at this point we have line of sight to much better models that are coming this year, and the amount of pain within OpenAI that we’ve had to decide how to allocate compute goes up, not down, over time.
+1more from this interview
- We’re in the earliest stages of thinking about data centers in space.
- We’re not investing approximately $200 billion in capex in 2026 on a hunch.
- I believe computing demand has increased by 1 million times over the last few years.
- I’ve come to a different conclusion: I don’t think we’re in a bubble (which, paradoxically, maybe is the truest evidence we are).
+1more from this blog
- This is an important step for Meta as we diversify our compute. I expect AMD to be an important partner for many years to come.
+1more from this filing
- If my revenue is not $1 trillion dollars, if it’s even $800 billion, there’s no force on earth, there’s no hedge on earth that could stop me from going bankrupt if I buy that much compute.
- With such strong demand for our existing offerings and seminal opportunities like AI, chips, robotics, and low earth orbit satellites, we expect to invest about $200 billion in capital expenditures across Amazon in 2026, and anticipate strong long-term return on invested capital.
- In 36 months, but probably closer to 30 months, the most economically compelling place to put AI will be space.
+1more from this podcast
- As we scale, we’re getting dramatically more efficient: we were able to lower Gemini serving unit costs by 78% over 2025 through model optimizations, efficiency and utilization improvements.
+1more from this talk
- Complete technological self-containment is unrealistic for any country, because the AI stack is incredibly complex. But strategic autonomy alongside rapid AI adoption is achievable, and it is a necessity for independent nations.
- Our conviction in the essential role of CPUs in the AI era continues to grow...
- For AI to have societal permission it must have real world eval impact. The choices we make about where we apply our scarce energy, compute, and talent resources will matter.
- My prediction, especially coming off of the successful Gemini 3 launch and hype around TPUs, is that the TSMC constraint could become material in 2026.
+1more from this essay
- But now that compute is big, compute is now very big, in some sense we are back to the age of research.
+2more from this podcast
- In fact, we’ve gotten so good at scaling AI with GPUs that power not compute is the primary limitation.
- Blackwell sales are off the charts, and cloud GPUs are sold out...
- The product that is the bottleneck is just a bunch of powered-up shells, right?
- You kind of need to build an infrastructure that’s capable of supporting multiple families and lineages of models. Otherwise the capital you put in, which is optimized for one model architecture, means you’re one tweak away, some MoE -like breakthrough that happens, and your entire network topology goes out of the window.
- Customers are hungry for AI and the compute, storage and networking we provide to deploy intelligence at scale.
- This partnership brings the best of AMD and OpenAI together to create a true win-win enabling the world’s most ambitious AI buildout and advancing the entire AI ecosystem.
- And just last month, we announced the world’s most powerful AI datacenter, Fairwater in southeastern Wisconsin, which will deliver 10x the performance of the world’s fastest supercomputer today.
- In which case, it will be another instance of the bitter lesson, that the things that used human knowledge were eventually superseded by things that just trained from experience and computation.
- Our vision is simple: we want to create a factory that can produce a gigawatt of new AI infrastructure every week.
+1more from this blog
- The AI race is on, and Blackwell is the platform at its center.
- We are seeing robust demand across our computing and AI product portfolio and are well positioned to deliver significant growth in the second half of the year...
- As datacenter production gets automated, the cost of intelligence should eventually converge to near the cost of electricity.
- Global demand for NVIDIA’s AI infrastructure is incredibly strong. AI inference token generation has surged tenfold in just one year, and as AI agents become mainstream, the demand for AI computing will accelerate.
- AI is now infrastructure, and this infrastructure, just like the internet, just like electricity, needs factories...
+1more from this talk
- But, inference will also get meaningfully more efficient in the next couple of years with improvements in model distillation, prompt caching, computing infrastructure, and model architectures. Reducing the cost per unit in AI will unleash AI being used as expansively as customers desire, and also lead to more overall AI spending.
+1more from this letter
- While model training still accounts for a large amount of the total AI spend, inference (which are the predictions or outputs of the models) will represent the overwhelming majority of future AI cost because customers train their models periodically, but produce inferences constantly in large-scale AI applications.
- Every quote from Andy Jassy
- The only thing that's going to happen with all the compute builds is the prices are going to come down.
+1more from this podcast
- And to safeguard America's advantage, the Trump Administration will ensure that the most powerful AI systems are built in the U.S. with American designed and manufactured chips.
- So what used to cost four dollars per million tokens now costs just 13 cents, and I expect this trend to continue. The result: intelligence is more available and accessible than ever before.
- The intelligence of an AI model roughly equals the log of the resources used to train and run it.
- The cost of AI, like the cost of mainframe computing to X.25 connectivity (the early pre-TCP/IP protocol for connecting computers over phone lines), literally forces the market to develop an alternative that scales without the massive direct capital.
- In FY 2025, Microsoft is on track to invest approximately $80 billion to build out AI-enabled datacenters to train AI models and deploy AI and cloud-based applications around the world.
- Making AI that is smarter than almost all humans at almost all things will require millions of chips, tens of billions of dollars (at least), and is most likely to happen in 2026-2027.
- Beyond that, the next frontier will involve designing data centers as flexible load centers, capable of adjusting their consumption based on real-time grid needs.
- AI’s $200B question is now AI’s $600B question.
+1more from this essay
- GPU computing is increasingly turning into a commodity, metered per hour. Unlike the CPU cloud, which became an oligopoly, new entrants building dedicated AI clouds continue to flood the market. Without a monopoly or oligopoly, high fixed cost + low marginal cost businesses almost always see prices competed down to marginal cost (e.g., airlines).
- Every quote from David Cahn
- Do we really want the infrastructure for the Manhattan Project to be controlled by some capricious Middle Eastern dictatorship?
+2more from this essay
- The trillion-dollar, 100GW cluster alone would require ~20% of current US electricity generation in 6 years; together with large inference capacity, demand will be multiples higher.
- White-collar workers are paid tens of trillions of dollars in wages annually worldwide; a drop-in remote worker that automates even a fraction of white-collar/cognitive jobs (imagine, say, a truly automated AI coder) would pay for the trillion-dollar cluster.
- Every quote from Leopold Aschenbrenner
- Over the past year, the talk of the town has shifted from $10 billion compute clusters to $100 billion clusters to trillion-dollar clusters.
- Our company has a one-year rhythm. Our basic philosophy is very simple: build the entire data center scale, disaggregate and sell to you parts on a one-year rhythm, and push everything to technology limits...
- If something is moving a million times faster every 10 years, what should you do?
- I actually think before we hit that, you're going to run into energy constraints. I don't think anyone's built a gigawatt single training cluster yet.
+1more from this podcast
- In terms of your question of large models, I think we’ve got to push scaling as hard as we can and that’s what we’re doing here.
- Accelerated computing and generative AI have hit the tipping point. Demand is surging worldwide across companies, industries and nations...
- But what’s interesting is that just scaling things up, and I put just in quotes because it’s actually hard to scale things up, but just scaling things up also appears to pay huge dividends.
