AWS称AI需求持续超过供给并锁定多年容量，强化AI算力链订单能见度

# 美股盘后催化 1/12: AWS和大型云厂商AI需求继续超供给，资本开支与长期容量锁定强化算力链多年订单能见度

- 美股会话: 2026-08-03
- 方向: POSITIVE；类型: 行业数据；规划重大性: 10/10
- 映射模式: customer_demand
- 美股驱动（规划器初判，需独立核验）: 事实：多条材料显示Amazon/AWS增长加速，Amazon市值突破3万亿美元，AWS AI收入run-rate超过250亿美元，AWS管理层称需求显著超过供应、部分容量已锁定至2027-2028年，并预计资本开支继续增加。推断：市场把AI capex重新定价为可见云收入和长期订单，而不是不可回收成本，带动GPU、ASIC、HBM、先进封装、服务器ODM、网络、光模块和电力基础设施。

## 本催化相关美股真实收盘行情

- 英伟达 (NVDA.US): 206.64 +2.93%
- AMD (AMD.US): 484.64 +1.78%
- 博通 (AVGO.US): 392.23 +0.76%
- 微软 (MSFT.US): 487.65 +4.93%
- 谷歌 (GOOGL.US): 373.51 +4.88%
- Meta (META.US): 590.24 +6.02%
- 亚马逊 (AMZN.US): 284.02 +4.58%
- Coherent (COHR.US): 288.14 +9.60%
- Lumentum (LITE.US): 779.89 +9.24%
- Arista (ANET.US): 184.89 +2.52%
- Credo (CRDO.US): 218.35 +5.49%
- Dell (DELL.US): 429.02 +5.83%

## 亚太映射假设（仅候选，必须查证）

- 台积电 (2330.TW) [TW/供应商/positive]: 云厂商GPU、ASIC和自研芯片扩容若延续，将拉动先进制程和CoWoS/先进封装需求。；核验问题: 核实AWS、Microsoft、Google、Meta最新capex中GPU、ASIC、网络和电力设备拆分，以及对应台积电产能占用。
- SK海力士 (000660.KS) [KR/供应商/positive]: AI训练和推理集群扩容通常提升HBM和高端服务器DRAM需求。；核验问题: 核实AWS及其它CSP 2026-2028 GPU/ASIC平台HBM供应商份额和长约价格。
- 广达 (2382.TW) [TW/供应商/positive]: 长期容量承诺可能提高AI服务器和整机柜ODM订单可见度。；核验问题: 核实AWS、Meta、Microsoft等AI服务器订单在台湾ODM间的实际分配。
- 台达电 (2308.TW) [TW/供应商/positive]: AI机柜功率密度提升会带动数据中心电源、UPS、散热和电力管理需求。；核验问题: 核实台达AI数据中心电源、液冷和北美hyperscaler客户订单。

## 必答问题

- AWS AI收入run-rate的口径是否包含训练、推理、托管服务和自研芯片收入？
- 云厂商capex增量最终落在GPU、ASIC、网络、服务器、电力和土地中的比例是多少？
- 亚洲供应商是否已经通过订单、排产或指引确认本轮上修？

## 原始证据全文（7 条）

### [E0997] Bloomberg TV+ | 2026-08-03 16:00 UTC | 盘中
「Bloomberg TV+」直播转录 23:30–00:00
URL: http://39.105.42.197:20307/event/live
【实时直播转录 · Bloomberg TV+】
来源:Bloomberg TV+(bloomberg 直播间 btv_us) · 标题:Bloomberg TV+
时段:2026-08-03 23:30–00:00(Asia/Shanghai),共 193 段
说明:以下为该时段该直播间的连续语音转录(直播 ASR,可能有同音/识别误差)。请从中识别与 AI算力/芯片/服务器/光模块/液冷/数据中心/大模型/机器人·具身智能/自动驾驶 等相关的、**可交易**的公司/产品/订单/产能/需求/合作/高管或官员表态,判断是否构成可交易事件;若通篇无可交易内容,判为不做。识别误差请结合上下文与常识修正,勿据孤立音近词臆断。

--- 转录正文 ---
[23:30] Of companies are doing. And so as companies think about how they take models in something in Amazon bedrock and run those models to get value out of their business.
[23:30] Some of them are automating processes, many of them are running agentic workloads that they build on AWS to further drive efficiencies in their business.
[23:30] Or deliver new customer experiences, we consider all of that work, whether it's agent growth, whether it's inference, and some of that is training.
[23:30] New models, all of that encompasses the AI business for us I think I've asked you this question at various points in time.
[23:30] You became AWS CEO, but is there a percentage split right now training versus inference that you can give me?
[23:30] Yeah, you do and it actually keeps shifting, I would say more and more it keeps shifting more towards inference, I think we still see large training flow.
[23:30] is being used by a number of companies, but as these models get really popular and really powerful, more and more companies are integrated.
[23:30] That inference into their workload, so I don't know the exact percentage today, but it keeps shifting more and more towards inference and we expect that to continue.
[23:31] The economics makes sense where you really want most of that cost and spend being where value is being created for end customers.
[23:31] Amazon's overall Capex number for this year is very big, $220 billion so it's up $20 billion and Andy Jessse explained.
[23:31] That's mostly AI, but it also accounts for higher memory pricing,, Capex higher because the cost environment's higher.
[23:31] But from AWS's perspective, what's the trajectory for next year, you expect that Capex will be bigger still because I.
[23:31] The things that Amazon is quite clear about is that even at $220 billion, it might not be enough to meet current demand.😊
[23:31] One of the things that we're quite excited about is just the potential business for us is just massive, and we see this as a huge opportunity.
[23:31] For us to really invest and help customers take advantage of the AI opportunity. And so we will keep investing. We think that there's a big opportunity.
[23:32] For us and for customers and as Andy kind of called out last week, we have really great insight into what that demand is.
[23:32] A be from customers and so when we invest and we'll keep investing in Capex next year as well, we have great insight into.
[23:32] When that revenue is going to land, and so it's pretty well known for us and as andy mentioned, much of our capacity has already spoke.
[23:32] And through through the end of 27 and even through much of 28, and so as we're investing, we're getting five year commitments from customers, we're getting these long.
[23:32] commitmentAnd so we're out there making investments to keep being able to grow the business and try to meet customer demand, but as you say today demand.
[23:32] still significantly outstripped supply and we're trying to build and invest to keep up with what customers are asking for.
[23:32] And so we expect the capex number will be bigger next year than it is this year us to keep investing as we see opportunities.
[23:32] And some of that we'll see how the market continues to grow, but right now we're excited about our investments and we'll continue to invest.
[23:33] The chip business has also got a lot of momentum, you know trainingium for some time has been a big part of the strategy and we've discussed that.
[23:33] Able to go to Austin, spend a lot of time in Annaperna Las to look at the server design.
[23:33] by asking when you say the chip business has a revenue run rate of $25 billion, what does that mean that is the business of selling the chips to third parties.
[23:33] Or renting out capacity based on the chips Yeah, renting out capacity based on those chips. So today we run all of our own chips inside.
[23:33] Of the AWS cloud and that includes both trainingium chips as well as graviton chips, which are general purpose processors.
[23:33] Graaviton is incredibly popular, we've been building Grviton for many years now, and in fact, almost all of our large customers use Grviton.
[23:33] Some part of their deployment and it's a key part of how help customers deliver extra value by lowering their costs.
[23:34] Is wildly popular and as we've mentioned a couple of times now, we're largely sold out through the end of next year for training capacity.
[23:34] I say that's really really popular we're starting to see both large customers as well as startups really love the cost benefits of running on trainingium.
[23:34] And having that choice in a cloud where you can run on purpose built processors like Traanium that give you really great AI performance.
[23:34] Together with NviDdia chips for when you need some of that processing for GPUs as well, gives customers the right.
[23:34] That they're oftentimes looking for. And so we find that combination to be quite powerful, and our traininganium3 chips are really, really popular.
[23:34] With customers right now we're live on Bloomberg Tele and Bloomberg Radio. This is Bloomberg Tech in conversation with Matt Garman, the CEO of A8.😊
[23:34] Explain the difference in economics between capacity based on trainingium and capacity based on black wealth.😊
[23:34] Ex, like what is the selling point of going with Amazon Silicon Yeah, we think that look.
[23:35] We think customers want choice, and so we offer great offerings for both products, some customers really like to build on NviIDdia GPU.
[23:35] U and those are fantastic and we're one of Nvidia's absolute largest customers in the world. we've been offering NviDdia GPUs in the cloud for over.
[23:35] Decade now, and AWS is the most scalable and secure and stable place to run Nvidia.
[23:35] Serverers anywhere. and so we're very excited about that business. But when you add trainingium, we're able to lower the costs for many workloads.😊
[23:35] Customers like to have that opportunity where they can take these really massively scaled AI workloads or inference alert workloads where we can tune them we.
[23:35] Con that whole stack we control the training chips, the data centers and the inference stack and can really tune that whole stack.
[23:35] It helps customers lower their costs and improve performance, and we're really starting to see that flywheel go where a model provided.
[23:36] Of all types and customers love that they get that value and great performance. And we think that that's a real flywheel that we can.
[23:36] Keep turning and are quite excited about continuing to grow that business Could you make it tangible for me, What is the cost saving?
[23:36] It's workload by workload, so there's not like a specific cost savings, but for workloads where we see this optimization happen, customers can often.
[23:36] Times save 20,30% off of their inference costs when they run that on trainingium interesting would you consider selling the chips?
[23:36] Outright as opposed to renting capacity based on them Yeah, we've mentioned that we might consider that in the future right now we have so much demand.
[23:36] Inside of the AWS cloud that we really love that business, and so that's the place we're focusing right now, but it's an interesting opportunity.
[23:36] It's something that we would definitely consider in the future. The other big development was AWS signing the open weights letter.
[23:36] Commun thatW was it important to you, Matt to sign that and participate in that.
[23:37] Again, a lot of this boils down to customers wanting choice and the big frontier labs have awesome models today, whether it's open AI.
[23:37] Running on bedrock, whether it's anthropic, running on bedrock, customers really love using those models, but they also like being able to customize models.
[23:37] Having a broad ecosystem of open weight models, I think is incredibly important for innovation it' important for our customers.
[23:37] And so you think about things like the Nematron models from NviDdia, or you think about Kimmi3 or some of the Chinese open weight models.
[23:37] Really help customers be able to customize to their own data and really be able to innovate. And so we think that it's important to.
[23:37] Not over legislate there and give that flexibility for customers to use whichever models they find to be the best fit. and that's actually why we really.
[23:37] on having all of those available inside a bedrock so that customers can choose which ones they want to use not to overislate I mean, I had the conversation with.
[23:38] Gensen Wong, the day the letter was posted about why then, what is the rationale, it does seem like the concern is that the USS government.
[23:38] Moly regulates open models. Was that sort of a motivation for you. Yeah, I think that's the concern. And I think you just want to make sure.
[23:38] That we kind of have an even playing field and it doesn't mean that there's none, by the way, I think there should probably be the same level of oversight of both frontiers.
[23:38] Models as well as open weight models, and I think we should have a consistent framework, but kind of making sure that folks realize how important.
[23:38] That is to what customers and our industry is out there building on and that those open weights are a key building block of that and we want to make sure that however legislation,
[23:38] It lands evenly across all of those Matt, a question from our Bloomberg tech audience for you is AWS' attitude.
[23:38] K3The open weight was released July 27th, given the platforms AWS offers what is the plan?
[23:38] Yeah, it's a great model that the team there has done a really nice job and we see really great performance there.
[23:39] And and we'll continue to support in bedrock offering the Kimmi models, I think the interesting thing is also is.
[23:39] You see many of these openweight companies, they're starting to think about how they make money in these models as well, and so they're starting to introduce licensing around.
[23:39] some of these open weights models as when customers want to run them in a cloud environment, and so I anticipate that there's going to be a blending of some of these models.
[23:39] To where these companies won't continue to spend lots of money and then just offer their IP up to the world, there're going to be licensing models.
[23:39] have ways of making money on those as well, and you're starting to see that with some of the open weights models actually having some licensing around them.
[23:39] Keep supporting these in bedrock and working with those companies in order to offer these in the best possible way.
[23:39] That's why I'd like to end the conversation, the business case and economic opportunity for open models from AWS's perspective.
[23:40] Odttlemakers themselves would like to make money. You How does AWSC that going in your favor to make money from why.
[23:40] You use of iPhone models yeah, yeah, I mean we we basically that is exactly what AWS, it's a great platform for companies to come and offer there.
[23:40] IP and their capabilities to the world and so everybody, whether they're startups, whether they're governments, whether they're large enterprises, everybody.
[23:40] Can build on top of aWS and if they have choice and they have access to these different models, it allows companies to be.
[23:40] To come and monetize them and sell when they have value to customers to be able to offer value when they think that they can go at lower prices or better performance.
[23:40] Having that open platform allows everybody to have that competitive chance and AWS is a fantastic partner for all these model providers to get.
[23:40] to the broad set of customers, and so it's a great ecosystem where model providers benefit and customers benefit by having all these in the same place.
[23:40] Matt Garmin, CEO of AWS, Amazon Web Services back on Lubergg Tech, thank you very much.
[23:41] Now coming up, home battery startup base power has a new valuation, $13 billion we're going to talk to CEO Zach Dell about.😊
[23:41] 🎼Its latest funding round and the launch of its US built home powering battery does next, this is Bloomberg Tech.😊
[23:42] Bringing you up to the minute geopolitical news.
[23:42] And wherever it happens i'm tyler kendall in Geneva sitzerland and this is bloomberg.
[23:42] Comes Downing Bloomberg has you covered And Burnham arriving for all the context and clarity you need.
[23:42] 🎼ion at first on Bloomberg in case you missed it on Daybreak.😊
[23:42] We talked going with the E regulator on this, though. Look, as I told you, I interact a lot with Americans and Chinese players.
[23:42] You see Trump fighting for their companies all the time. You see the Chinese fighting for their companies all the time. here in Europe, I feel I'm fighting against the European.
[23:42] 🎼Not that they are fighting for us. They are fighting against us. It's a big mistake because we need to have really big global tech leaders in Europe.
[23:42] 🎼Europeope live every weekday.
[23:43] People are probably going to have one, maybe two, maybe a small handful of agents that they rely on and they.
[23:43] have a personal agent that's focused on things like their health and maintaining their personal relationships and you perhaps they use that same age.
[23:43] In their work lives, especially if they're working in a small business or they're an entrepreneur, or maybe there's worlds where if you work within a larger enterprise.
[23:43] 🎼Larger company than these become bifurcated and separated, not too much unlike email, let's say Yeah, we think that ultimately.
[23:43] It should be, you know, agents will be something that become deeply personal and should be things where.
[23:43] Over time, you find yourself being able to rely on them more and more and more, for more and more of your personal life, more and more of your work life, and that'll be.
[23:43] ces that yout wall society goes through it together week.
[23:44] Bloomer puts the spotlight on a different aspect of global finance, flows, funds and the forces shaping markets on ETF IQ.
[23:44] 🎼The future of money with Bloberg crypto the players behind major transactions on Bloberg deals rates, risk and the cost of capital.
[23:44] On Bloomberg Real Y and personal F retirement and wealthth management on Bloomberg Money, a different topic every weekday.
[23:44] Only on bloomloer television.
[23:44] Rising demand for electricity in the US is fueling investor interest in backing battery startups among them.
[23:44] Base power, which just closed a $1 billion series d round at a $13 billion post money valuation.
[23:44] And co founder Zach Dell joins us now on Blueloomberg Tech, a big round,, and inevitably part of needing.
[23:45] The funds is to scale, to get moving on the operation side, but let's start there like what is it you want to do in the first instance, Zach?
[23:45] Yeah, well, we really want to scale energy infrastructure to meet rising demand for power, so we build a hardware product.
[23:45] That we install on homes and we leverage, it said home battery that we design and manufacture ourselves, and we leverage those batteries to support the grid.
[23:45] Times it high demand. So we're adding capacity to the system to meet this rising demand from AI electrification of heavy industry.
[23:45] Of transportation, we're adding capacity to the system through distributed batteries, the grid is able to use those batteries when demand is high and the homeowners.
[23:45] That are have those batteries installed in their' home are able to leverage those batteries when the grid goes out. Sorry, this is a residential product.
[23:45] At a basic level, how tied is your fate and future to the adoption of solar in this.
[23:46] Not really at all. Solar is a generation technology, much like wind or gas or.
[23:46] And batteries are really more akin to a transmission and distribution technology. So the way to think about it is that poles and wires move energy through.
[23:46] Space and batteries move energy through time. And so the battery is really a utilization tool that is helping increase the capacity.
[23:46] ActorOf the system and bring down costs for everybody. So I ask you that because the energy or electricity.
[23:46] 't the problem here. What is the problem that the base power is trying to solve for.
[23:46] problem on the grid today. So the grid is really built for peak demand, which means that most of the time we're not utilizing.
[23:46] All of the infrastructure, which ends up being very expensive for rate payers or consumers of electricity, so by deploying batteries to the edge of the grid.
[23:46] You're able to increase the utilization, increase the capacity factor, and thereby lower costs for consumers.
[23:47] At home, I have a Tesla Power War 3. It does rely on solar to charge it during the daytime, but let's talk about.
[23:47] That is a comparison or a case study at the technology level, the base core, it is better how is different how.
[23:47] It's reallyally designed to be a grid resource. So it's almost three, maybe over three times the size. It's installed much faster. It's actually.
[23:47] OwnAnd operated by us the company based power and used as a grid resource. So when you go buy a home battery or home generator, it's owned by the homeowner.
[23:47] And it's really only there when the grid goes out, our asset is really a piece of grid infrastructure and so in partnership with.
[23:47] Some of our utility partners, those batteries are used as a grid resource by the utility in times of high demand, like I mentioned earlier.
[23:47] And then in the deregulated markets in Texas and Illinois where we operate, we use those batteries as a wholesale power asset, so we're really operating.
[23:48] Fleet of infrastructure, whereas those companies that sell batteries or generators, they're just trying to sell high margin hardware out front and our view is that batteries are really.
[23:48] Inrastructure asset, and they're better sold and built kind of as such,, and not really as a high margin productach.
[23:48] I feel like this is a Texas story from the manufacturing standpoint. right, this is going to be an American built technology.
[23:48] The challenges in establishing Texas as the base, the supply chain, making this an economically viable.
[23:48] Plan well, Texas is a fantastic place to build companies and also to to establish manufacturing and we see actually.
[23:48] Huge advantages in colocating engineering and manufacturing because the team that designs the product needs to be able to be very close to the team.
[23:48] that builds the product and so by coupling engineering and manufacturing and having that done by one team, one company and one location, you get.
[23:48] A bunch of efficiencies, it is much easier to iterate and you can move a lot faster Should we talk about some of your investors in the round?
[23:49] What was the importance of assembling that group beyond just the checks that came in, what are you hoping to leverage from there?
[23:49] Experiences and their expertise were showing Riit, Co2, Valor, Alttia Sands Capital D1.😊
[23:49] Building the company to compound over many decades, and we want to establish a cap table that is aligned with that kind of long term thinking.
[23:49] SoThese investors who joined our cab table today and then who've been investing with us now for a couple of rounds, they see the opportunity to scale the business across.
[23:49] The US internationally and do that over a long period of time, and they add a lot of interesting insights, perspectives.
[23:49] And they're really our thought partners and our operating partners day to day, so we're super lucky to have a whole host of new investors joining the capital.
[23:49] And then basically, all of our major existing investors reinvesting in the company Earler in the program, we had valoratomics here.
[23:50] Isaiah Taylor on, you know different part of the energy S, but we talked about how is America doing in this domain?😊
[23:50] Right, we talk a lot about China, how quickly China moves, I guess at the nation level and at the federal level.
[23:50] Everything you need to be able to execute in the way you want, get the regulatory support.
[23:50] A lot of energy being applied. no plan intended to the space today, whether that's from the regulators the policy makers of also entrepreneurs.
[23:50] Like our team at Base and the folks you mentioned earlier, some of the other companies entering the space. I mean, when we started base three years ago.
[23:50] It kind of unique to start an energy technology company, and there are not a lot of those out there, and now they're very common, and you see them pop up all the time. and we love that we think that,
[23:50] That the industry needs more great entrepreneurs and teams solving hard problems and helping us scale our capacity to consume electricity in this country.
[23:50] I think we're in a really good spot and we've just got to keep executing relentlessly and building great solutions to some of these hard problems in the energy industry.
[23:51] You're building across the electric stack and I kind of read a lot about you and think you're pitching yourselves as like a next generation utility.
[23:51] Almost and I think we should end the conversation with like, what is your relationship like with the existing utilities with the rest of the grid?😊
[23:51] You're all kind of trying to move together in the same direction. Yeah, I wouldn't think of us really as a next generation utility. I would think of us as.
[23:51] Modern energy technology company that is a provider of choice to the utilities to scale their infrastructure so we work very closely with utilities.
[23:51] With all kinds of utilities, with municipal utilities like Austin Energy, with Coops like Coserve and GVAC, with investor owned utilities like El Paso Electric.
[23:51] And we build them technology solutions so they can scale their infrastructure and meet the moment, meet the rising demand that's coming from AI and all these things.
[23:51] that you just got all the time. there's an incredible inflection happening in energy consumption in America, and it's never been.
[23:52] More important for the utilities to kind of step up and adapt. And we are really helping them do that. And it's our honor to work with you.
[23:52] These are incredible partners and they have a huge responsibility in this coming decade, and we're really excited to partner with them Zach very quickly.
[23:52] Is it like trying to hire right now the talent you need, Look, I think that we have a incredibly mission.
[23:52] Driven culture and company, and there's a lot of people out there, hopefully some of them are watching this show that feels strongly that.
[23:52] That's scaling our capacity consume electricity and really driving towards energy abundance on the path to human process.
[23:52] Vity is one of the most uplifting and exciting missions that you can work on. And so we've had the great fortune of having tons of the best engineers in the world come.😊
[23:52] B and say, hey, this is the mission I've got to be a part of, and so we've had a really great time, frankly, building out an incredible team down in Austin.
[23:52] very mission aligned individuals, and we're excited to continue to grow that team as we continue to scale a business Zach Do.
[23:53] Co foundunder and CEO of Base Power. thank you very much and team. There's a lot of other tech headlines. It's time for talking tech and first up South Korean AI chip design.😊
[23:53] 🎼Dex secureded funding at roughly four times its previous valuation buffing the firm to $2.2 billion.
[23:53] 🎼After the start, which developed semiconductor architecture, tailored for AI tasks, said it secured over 30 mass production contracts.😊
[23:53] 🎼Custom across eight countries plus according to a PwC survey, the0 finance executives86% of ex said they.
[23:53] 🎼AI training was more valuable than an MBA for new hires, 91% of those execs also said they were willing to pay accordingly.
[23:53] 🎼Increasing compensation for AI literate employees, pointing to a growing urgency for firms to adopt AI tech.😊
[23:53] 🎼Spiderman, brand new Day was the biggest debut in Hollywood history with a movie opening to $360 million in the US.😡
[23:54] 🎼In Canada was a smashing success in China too, hauling the country out of its post pandemic, Hollywood slump and contributing to.😡
[23:54] 🎼A global box office debut of $927 million. That does it for this edition of Bloomberg Tech.
[23:54] 🎼These are the earnings that we're looking at this week, earnings continues, we've got palantier, SpaceX first ever earnings report and call as a public company.😊
[23:54] 🎼AMD, a big one in the AI infrastructure space and then later a focus on entertainment the gig economy, Paramount, Disney.
[23:54] 🎼All on Wednesday sorry Disney and Uber on Wednesday, recap everything on the podcast was a pretty fast start to the week.😊
[23:54] 🎼Kn where to find it on the terminal, on all Bloomboow platforms and online on Apple, Spotify and Ihp. Have a great week.😊
[23:54] I bag tech.
[23:56] Bringing you the latest geopolitical news whenever and wherever it happens I'm David Garra and Solaruni in Ololivia and this.
[23:56] 🎼I when news breaks a red head across the Bloberg terminal Bloomberg has you covered Trump's global.
[23:56] 🎼A struck down by the US Supreme Court for all the context and clarity you need. There's going to be now tons of tariff headlines until midterm election.
[23:56] First on Bloomberg.
[23:56] 🎼
[23:56] 🎼M good morning. this is We back to the opening trade. Itberg money. This is the trade. This is Wall Street week.😊
[23:57] 🎼We balance of you're watching Bberg deal We to Bberg this weekend. this is Bberg television and and.
[23:57] 🎼In every aspect of our lives and that revolution is playing out in real time from finance to defense tech, AI to end.
[23:57] 🎼tain from the road to the stars Bomberg is bringing you the stories of companies and people that are pushing tech to new fronts.
[23:57] 🎼And the politics reshaping global tech markets I'm Ed Lodlow live in San Francisco in this.😊
[23:57] It bloomberg every weekday only on bloomberg television get your fixed income bloom.
[23:57] Real yield every thursday at 12 pm Eastern right here on Bomberg, context changes everything.

### [E1143] Bloomberg | 2026-08-03 16:24 UTC | 盘中
Amazon Cloud Chief Says Potential AI Business is ‘Just Massive’
URL: https://www.bloomberg.com/news/articles/2026-08-03/amazon-cloud-chief-says-potential-ai-business-is-just-massive
Amazon.com Inc.’s cloud chief said customers are starting to shift from using the company’s services to train AI models to integrating them into their businesses — fueling demand for a process known as inference.

“We still see large training clusters being used by a number of companies, but as these models get really popular and really powerful, more and more companies are integrating that inference into their workloads,” Amazon Web Services Chief Executive Officer Matt Garman said on the Bloomberg Tech television show on Monday.

Training is the computing-intensive process of assembling an AI model. Inference is the practice of using that model to respond to a user’s prompts.

Amazon, the largest seller of rented computing power and data services, last week said it would spend $220 billion in capital expenditures in 2026, up from a prior forecast of $200 billion. The uptick in spending reflects the rising price of memory and other components required for data centers.

Garman declined to forecast the company’s spending next year, but said the potential artificial intelligence business for the company is “just massive.”

He said Amazon is confident in its projections for demand, and is signing up customers with an eye toward long-term deals that help AWS project use for new data centers.

“As we’re investing, we’re getting five-year commitments from customers,” Garman told Bloomberg Television. “Today, demand still significantly outstrips supply and we’re trying to build and invest to keep up with what customers are asking for.”

The shares have soared in recent days, propelling Amazon’s market value beyond $3 trillion, after the company reported last week that cloud-computing revenue accelerated for a fifth straight quarter.

### [E1385] 瑞银 | 2026-08-03 17:18 UTC | 盘中
Key Takeaway
With the stock up ~7% AH, outperformance was driven by AWS revenue beat of +37% YOY growth, above our, Street, and investor expectations.
North America retail also came in higher vs the Street by 2% largely driven by the pull f
URL: https://open.gangtise.com/research/?id=474991048180502528#/ResearchDetails
Key Takeaway
With the stock up ~7% AH, outperformance was driven by AWS revenue beat of +37% YOY growth, above our, Street, and investor expectations.
North America retail also came in higher vs the Street by 2% largely driven by the pull forward of its US Prime Day into June vs regular July timing - this also likely drove the revenue guidance shortfall for 3Q26 vs Street expectations. We tune into the call for any revisions to Amazon's FY26 CapEx guide of $200B.
2Q26 Results: Topline outperformance driven by AWS and NA retail likely due to US Prime Day in June
Net Revenue: $200.6B (+20%) vs UBSe $196.2B and the Street $196.7B or +17%, guidance $194B-$199B (+16%-19%)
Operating Income (Margins): $27.5B (13.7%) vs UBSe $24.2B (12.3%) and the Street $23.8B (12.1%), guidance $20B-$24B (10%-12%)
AWS: $42.2B (+37%) vs UBSe $40.8B (+32%) and the Street $40.5B (+31%), we believe investors were anticipating ~36%
North America Retail: $116.2B (+16%) vs UBSe $111.9B (+12%) and the Street $113.8B (+14%)
International Retail: $42.2B (+15%) vs UBSe $43.6B (+19%) and the Street $42.6B (+16%)
CapEx: $54.2B vs UBSe $54.6B and the Street $52.6B
3Q26 Guidance: Guidance range falls short likely due to dual Prime Day events in June and July
Net Revenue: $197B-$202B vs UBSe $207B (+15%) and the Street $204B (+13%) and excluding the impact of Prime Day in 2025 and 2026, 3Q26 growth would be 400bps higher. The guide includes ~80bps of unfavorable foreign exchange impact.
Operating Income (Margins): $22.5B-$26.5B (11.4%-13.1%) vs UBSe $25.6B (12.4%) and the Street $25.4B (12.4%)
Other Comments
2Q26 net income includes ~$53.4 billion from its investments into Anthropic
Pending any revisions on $200B CapEx bill in FY26 vs investor expectations for $210B-$215B
First Read: Amazon.com 30 July 2026
Company Description
Amazon.com, Inc. is a customer-centric company providing for three customer sets: consumers, sellers, and enterprises. It operates a range of businesses including e-commerce/retail, AWS/cloud services, TV streaming, digital advertising, and financial services.
Valuation Method and Risk Statement
Risks to our valuation includes: 1) rise in competitive intensity from any number of Amazon's online e-commerce and offline retail competitors, 2) deterioration in consumer sentiment, which could impact transaction volume velocity, 3) higher-than-expected capital intensity for either the e-commerce platform or AWS, which could weigh on our free cash flow projections.
We value AMZN shares on a price to free cash flow basis as the company has always centered discussions with investors around free cash flow generation.

### [E1799] Seeking Alpha | 2026-08-03 19:19 UTC | 盘中
AI infrastructure: Checking in on hyperscaler capex
URL: https://seekingalpha.com/news/4623520-ai-infrastructure-checking-in-on-hyperscaler-capex?utm_source=feed_news_all&utm_medium=referral&feed_item_type=news
Investors Should Know: Big Tech's AI infrastructure buildout has entered a new phase, with investors shifting focus from raw spending levels to whether capital expenditures are generating visible returns. Recent earnings from the four largest cloud operators show a widening gap between those hitting that bar and those still falling short.
Background
Four of the largest cloud and AI infrastructure operators, namely, Microsoft (MSFT), Alphabet (GOOGL), Meta Platforms (META), and Amazon (AMZN), have collectively committed hundreds of billions of dollars to AI data centers, chips, and compute capacity. Their Q2 earnings reports, released over the past couple weeks, brought that spending under intense scrutiny.
The central question driving analyst and investor sentiment is no longer whether to spend, but whether the spending is producing returns fast enough. The market has moved from rewarding aggressive investment to demanding evidence that revenue growth can outpace capital expenditure growth, a threshold only some operators have clearly crossed.
Stocks to Watch
The buildout extends well beyond the hyperscalers themselves. Advanced Micro Devices (AMD) and Nvidia (NVDA) are competing for rack-scale chip dominance inside these data centers. Semiconductor equipment makers, including ASML Holding (ASML), Lam Research (LRCX), and Tokyo Electron (TOELY) sit further up the supply chain, supplying the tools needed to manufacture the chips powering AI workloads. Memory chipmaker Micron (MU) has also been cited as a potential beneficiary as memory's share of total AI infrastructure spending grows.
Adjacent infrastructure players are also seeing a direct impact. SPX Technologies (SPXC) raised its data center capacity target, and Navitas Semiconductor (NVTS) projected AI infrastructure will represent more than a third of its sales by year-end. Power availability has emerged as a separate bottleneck, reshaping energy infrastructure investment alongside compute capacity.
Key Takeaways
Microsoft has become the benchmark for AI infrastructure balance, with Azure revenue surpassing $100 billion annually and the company guiding to positive free cash flow in fiscal 2027, even as capital expenditures rose roughly 70% year-over-year.
Alphabet and Meta face heavier investor skepticism, with both companies reporting negative or near-zero free cash flow while raising their forward capex outlooks.
Morgan Stanley analysts estimate hyperscaler AI infrastructure can generate 25–50% long-term returns on invested capital, with GPU rental businesses potentially producing 60–70% incremental operating margins. However, these returns remain forward-looking.
The buildout is reshaping adjacent sectors, including chip supply chains, semiconductor equipment, memory, and power infrastructure, as bottlenecks shift across the ecosystem.
News Roundup
Recent earnings and announcements from the major cloud operators and their chip partners illustrate how differently each company is navigating the tension between AI investment and financial discipline. The items below span quarterly results, analyst reactions, and strategic moves that together define the current state of AI infrastructure spending.
Microsoft (MSFT) — Azure revenue surpassed $100 billion annually for the first time, with quarterly growth accelerating four points sequentially to 43% year-over-year. BNP Paribas analyst Stefan Slowinski noted Microsoft had reached "an important step... in achieving an equilibrium state where revenue growth can ultimately outpace CapEx growth," while RBC Capital Markets, J.P. Morgan, and Morgan Stanley all reiterated bullish stances. The company generated $19.6 billion in free cash flow despite capital expenditures rising roughly 70% year-over-year to $41 billion, and guided to positive free cash flow in fiscal 2027.
Alphabet (GOOGL) — The company spent $44.9 billion on capital projects in a single quarter and raised its 2026 capex guidance to $195–205 billion, up from $180–190 billion previously, sending shares lower. Alphabet reported negative free cash flow of $5.9 billion for the quarter, even as cloud revenue grew 82% and contracted backlog reached $514 billion. Management acknowledged that leaning on third-party capacity as a bridging strategy creates "modest margin pressure," and the CFO confirmed the company remains supply-constrained even after the guidance increase.
Amazon (AMZN) — Following strong Q2 results, Amazon announced it would boost AI and technology spending by an additional $20 billion, with CEO Andy Jassy describing AWS as "booming" and noting its AI and chips businesses each surpassed $25 billion run rates. Ahead of the print, Wedbush had projected Amazon would raise 2026 capex guidance above $200 billion and 2027 guidance to $266 billion, with analysts watching Trainium chip capacity and free cash flow trends closely.
Meta Platforms (META) — Shares fell after earnings as investors reacted to AI capex expenses surging 55% quarter-over-quarter, free cash flow dropping to $784 million against $31.9 billion in operating cash flow, and the company taking on $24.9 billion in additional debt while suspending buybacks. Meta raised its 2026 capex outlook to a $137.5 billion midpoint, presenting what analysts described as the most polarized sentiment among the four hyperscalers.
AMD (AMD) — An expanded Azure partnership deploying AMD's Helios rack-scale platform, combining Instinct MI455X GPUs, EPYC Venice CPUs, and Pensando networking, gave a boost to AMD shares, positioning the company as Nvidia's most credible rack-scale competitor. At its Advancing AI 2026 event, AMD raised its long-term AI accelerator total addressable market forecast to roughly $1.4 trillion by 2030, alongside a CPU total addressable market raised to over $200 billion.
Nvidia (NVDA) — Nvidia has struck revenue-sharing deals with AI cloud partners to deploy multi-tenant AI factories, positioning the company within the broader $1.5 trillion in projected 2027 global cloud and AI capex that analysts are tracking.
SPX Technologies (SPXC) — The company lifted its data center capacity target to $1.1 billion from $750 million and raised full-year adjusted earnings per share guidance, reflecting direct demand from the AI infrastructure buildout.
Navitas Semiconductor (NVTS) — Navitas projected that AI infrastructure will represent more than a third of its sales by year-end, citing record book-to-bill as evidence of accelerating demand from data center customers.
AI Infrastructure Stocks in Focus
The screener below identifies publicly traded companies classified under Seeking Alpha's AI cloud and high-performance computing themes, as well as the AI chip supply chain, covering the compute capacity buildout and the semiconductor supply chain that supports it. Companies are ranked by relevance to these themes, with the most theme-aligned names appearing first.
The thematic screener draws primarily from the chip supply chain layer of the AI infrastructure ecosystem. Results include semiconductor equipment makers and chip manufacturers whose businesses are directly tied to producing the compute capacity at the center of the hyperscaler buildout. Market caps and Quant Ratings vary across the list, reflecting the range of company sizes operating across this supply chain.
Companies in the screener include Tokyo Electron Limited (TOELY), ASM International NV (ASMIY), SUSS MicroTec SE (SESMF), MKS Instruments (MKSI), Lam Research Corporation (LRCX), QuickLogic Corporation (QUIK), Photronics (PLAB), Atomera Incorporated (ATOM), ASML Holding N.V. (ASML), and Onto Innovation (ONTO).
Explore Further
Stock Screeners — Browse and customize predefined screens for AI infrastructure and technology themes.

### [E2684] AWS Newsroom | 2026-08-03 21:04 UTC | 收盘后
AWS CEO @mattsgarman talks to @EdLudlow about what's behind the fastest AWS growth in nearly 5 years: "We're not just concentrated on one or two large customers — it's actually growth from a really broad set of customers." AI demand across
URL: https://x.com/AWSNewsroom/status/2084384991232868788
AWS CEO @mattsgarman talks to @EdLudlow about what's behind the fastest AWS growth in nearly 5 years: "We're not just concentrated on one or two large customers — it's actually growth from a really broad set of customers." AI demand across every industry. $25+B AI revenue run rate.

### [E1804] Bloomberg | 2026-08-03 19:20 UTC | 盘中
OpenAI, Anthropic, Google to Join White House AI Safety Meeting
URL: https://www.bloomberg.com/news/articles/2026-08-03/openai-anthropic-google-to-join-white-house-ai-safety-meeting
The Trump administration plans to host artificial intelligence companies at the White House on Tuesday to discuss a new US framework for conducting voluntary safety tests of AI models, according to people familiar with the matter.

OpenAI, Anthropic PBC and Alphabet Inc.’s Google are among the AI developers planning to attend, according to the people, who described the meeting on condition of anonymity because it hasn’t yet been announced. The recently completed framework springs from a June executive order by President Donald Trump on AI cybersecurity that outlined an opt-in approach to safety reviews of models and greater efforts to shore up critical computer systems.

The framework has not yet been released publicly, and Trump’s order indicated that some of its benchmarks would remain confidential. Nonetheless, it marks the Trump administration’s latest step toward addressing AI security issues, an initiative that accelerated in April after Anthropic warned that its Mythos model was adept at finding computer vulnerabilities and strictly limited its release.

Spokespeople for the White House and the three companies declined to comment. The Information reported earlier on the Tuesday’s meeting plan.

Read More: Anthropic, OpenAI Cyber Failures Point to US Security Risks

Adding urgency to the effort, OpenAI and Anthropic have disclosed in recent weeks that a handful of their models escaped secure testing environments and hacked third-party organizations.

At the same time, both companies have chafed at what they see as an inconsistent approach to AI safety reviews, and are looking to the new framework to bring more consistency to what had been initially presented as a voluntary program.

In June, weeks after Trump signed the order and signaled a largely hands-off approach to testing, the Commerce Department barred Anthropic from sharing its two most powerful models with foreign nationals, citing concerns that safety guardrails could be circumvented. The company temporarily disabled access to its Fable 5 model and its selectively shared Mythos 5 model, eventually making changes to address the administration’s objections.

Read More: US Considers Creating Finra-Like Watchdog to Vet Top AI Models

Partly in response to the industry’s complaints of an ad-hoc approach to model reviews, Treasury Secretary Scott Bessent proposed creating an independent regulatory agency that would give companies a significant say over safety reviews. The contours of that plan align with a proposal from Google DeepMind’s leader Demis Hassabis, who was in Washington last month to meet with policymakers.

Bessent’s model would set up a system similar to the Financial Industry Regulatory Authority, but apply to the artificial intelligence industry, Bloomberg News has reported. It remains unclear if that plan would be part of the AI cybersecurity order implementation or a separate effort.

Tuesday’s discussions will take place less than two months before Trump meets with his Chinese counterpart Xi Jinping for a high-profile summit in Washington where the competition for AI leadership between the world’s two largest economies will take center stage.

Read More: Alibaba Adds to China AI Breakthroughs With New Qwen Model

US officials are now grappling with how to respond to a wave of new AI models from Chinese firms that rival top offerings from American labs. Last month’s release of the new Kimi K3 model by Chinese startup Moonshot AI undermined investor confidence in the durability of the US lead in AI while raising questions about the hundreds of billions of dollars in data center investments. Kimi, which costs a fraction of the price of top US offerings, nears their performance in some domains.

That challenge was highlighted by recent moves by other Chinese developers. Over the weekend, DeepSeek expanded access to its latest model and on Monday, Alibaba Group Holding Ltd. unveiled its own new system, Qwen3.8-Max, which the company said is on par with Anthropic.

The slew of releases has fomented tech industry debate over how the US should respond to open-weight AI models from China that cost far less than the best US products while approaching their performance in some domains.

Silicon Valley leaders including Nvidia Corp. Chief Executive Officer Jensen Huang have insisted that open-weight models, which allow users to download and customize the technology, are good for the long-term development of AI and help bolster its security. In public statements, with dozens of companies signing in support, they’ve urged the US to refrain from restricting open models.

Read More: Moonshot’s Kimi Uses 20,000 Nvidia Chip Cluster From Alibaba

Other key tech industry figures, including Anthropic CEO Dario Amodei, have suggested that the gains made by Chinese AI models violate US rules, concerns echoed by Bessent and other top US officials.

Last month, White House Director of Science and Technology Policy Michael Kratsios accused Moonshot of illegally acquiring Nvidia’s advanced Blackwell chips, the most powerful AI accelerators on the market. It’s unclear how many of these processors the US government believes Moonshot has. Bloomberg has reported that Moonshot has a computing power agreement with Alibaba for the use of about 20,000 Nvidia chips, forming a substantial chunk of the overall capacity Moonshot uses for its Kimi models.

Kratsios also said Moonshot improperly extracted data from US models through a process known as distillation, a style of reverse engineering that US AI companies have warned officials about for most of this year. Using this technique, researchers feed the outputs of one model into a new version, a process that turns out to be an efficient way to recreate some of the parent model’s capabilities.

### [E1874] 华尔街见闻快讯 | 2026-08-03 19:50 UTC | 盘中
美国AI债务融资史无前例！城堡证券：2028年将超5000亿美元
URL: https://wallstreetcn.com/articles/3778602
城堡证券预测，到2028年，为支持AI芯片采购，美国科技公司将在公开和私募市场累计发行逾5000亿美元债务，仅2028年单年芯片制造商发债规模就将突破2500亿美元。这一融资体量将重塑整个投资级信用市场格局，投资者或不得不压缩在科技、媒体及电信板块的配置，以腾出资金承接。

## 研究交付要求

这是一条已经从整场材料语义拆出的独立催化，不要重新把它压回大盘总评。核实美股异动原因和传播时序；用结构化供应链、内部 KB 与公司披露坐实每个亚洲候选；逐腿查真实行情和本地市场 price-in；候选不成立也要逐条给出有证据的 AVOID。