Two items from “News Items” from @EllisItems this AM (https://substack.news-items.com):
“8. A rout in global chipmakers deepened on Tuesday, as fears over the durability of the AI boom intensified ahead of results from some of Silicon Valley’s biggest companies this week. South Korea’s Kospi led declines in Asia, falling more than 10 per cent and prompting a short halt in trading, after investors dumped shares in the country’s two leading memory-chip makers. Shares in SK Hynix fell as much as 10 per cent, while its larger rival, Samsung Electronics, dropped more than 12 per cent. The two companies have tumbled 41 per cent and 34 per cent, respectively, in July so far. In Tokyo, the Nikkei 225 fell 4.4 per cent, with memory-chip maker Kioxia plunging more than 18 per cent. This month’s sell-off has slashed Kioxia’s share price in half. (Source: http://ft.com)
9. A closely watched gauge of risk in holding the debt of companies at the centre of the AI boom is rising rapidly, underscoring growing jitters over Big Tech’s vast spending on data centres, chips and computer memory. Prices for credit default swaps, popular tools to bet against corporate debt, tied to Oracle, SpaceX, Alphabet, Amazon, Meta, Broadcom and Nvidia have risen to record highs in recent days, according to LSEG data. The sharp moves echo a sell-off in debt issued by so-called hyperscalers, which are piling hundreds of billions of dollars into developing vast data centres and sophisticated AI models. It comes as investors have grown increasingly worried about the deluge of debt sold by these companies. (Source: https://t.co/i1iKd1P8Ml)”
85·AShort
n
news17h agonews
Anthropic CEO Dario Amodei breaks silence on open-weight AI after Nvidia, Microsoft, Meta, OpenAI, and Google back open AI models
75·ALong
n
news17h agonews
Anthropic CEO Dario Amodei breaks silence on open-weight AI after Nvidia, Microsoft, Meta, OpenAI, and Google back open AI models
75·ALong
m
meme17h agomeme
BLACKROCK'S $12.3 BILLION AI DATA CENTER BOND DEAL FOR META RALLIED BEFORE PRICING, SIGNALING STRONGER-THAN-EXPECTED INVESTOR DEMAND.
THE BONDS TIGHTENED IN THE GRAY MARKET DESPITE INITIAL WEAK DEMAND, HELPED BY HIGHER-THAN-USUAL YIELDS ATTRACTING BUYERS.
THE REBOUND CONTRASTS SHARPLY WITH RECENT AI-RELATED BOND MISFIRES, SUGGESTING INVESTORS REMAIN SELECTIVE RATHER THAN ABANDONING THE SECTOR.
Just some TLDR news:
- $CXMT IPO tomorrow if you like Chinese memory.
- Samsung Electronics reportedly struggling to secure large FC-BGA substrates. Ibiden (4062) reportedly requested LTA guarantees, Samsung Electro-Mechanics sought prepayments.
- Samsung + $AVGO sign memory + foundry AI framework through 2030, expected to exceed $200b
- $SOI expects FY2027 silicon photonics revenue to double compared to the previous year, surpassing Morgan Stanley's 60% growth projection. Photonics thesis go brrr.
- SK Group + $NVDA sign $500B+ partnership to build out AI DCs and HBM4 memory.
- SKC Absolics glass core delay to 2027 from reports. Targeting final reliability testing EOY, mass production next year. So if you're curious, this does push back some ramps from $LPK and others (hence drop on ER).
- $QCOM price hikes by double digits for smartphone processors, due to upstream supplier hike pricing.
- $META expected to issue $12B in project-level/SPV financing for El Paso, Texas AI DC expansion
- Naver announced a $10 billion investment from $NVDA and Brookfield to construct a 1GW-scale AI Factory. Near term plan is 200MW by 2028. (Nvidia $1B investment, Brookfield nonbinding $9B)
- 64GB DDR5 server modules rises 146% versus end-June contract pricing
- $INTC brought forward 14A process mass production by a year, risk production H2 2027 and volume production in 2028, vs. 2029 HVM expectations.
- Samsung Electro-Mechanics wins $200m MLCC order (existing bottleneck).
- $AMD (the Bandana bottleneck), announces Helios is in full production with shipments Q3 2026. Including an up to 2GW MI455X GPU deployment with Anthropic and a 6GW infrastructure rollout with OpenAI. Gave new >50% CAGR TAM to $220B by 2030 from $26b in 2025 for CPU market.
Rolls out optical interconnects for Mi500 in 2027. From channel checks, AMD is heading down to the CPO route (seems likely to use Ayar).
- $ORCL wins $7B Department of War enterprise software contract
- JX metal doubles semi target capacity at its KR subsidary with a 4B yen investment, with operations to begin H2 2027, amid surging demand from major customers Samsung Electronics and SK Hynix
- Tungsten hexafluoride spot prices surged 2.1-2.5x Y/Y following Japan's Kanto Denka and Chuo Gas announcing permanent production halts. Fluorinated liquid supply faces a vacuum as 3M plans to exit PFAS production.
- From the four optical chipmaker earnings, Yuanjie/Eoptolink/TFC Optical/Dongshan Precision: no major order cuts, 1.6T shipments expected to accelerate into 2027. Optical chip suppliers expected to capture outsized margins from shortages.
- Unitree Robotics Targets 30,000 Humanoid Robot Production Capacity by 2026. Read through on humanoid TAM scaling like $CCXI and others.
- Energy storage lithium batteries orders increase first half orders by 2,900% apparently in China. Not as familiar with EVE Energy and other battery makers.
75·ALong
m
meme7/24meme
Premarket movers:
Tesla leads gains among Mag 7 stocks after the electric vehicle maker plunged about 15% on Thursday (Tesla +1.3%, Microsoft +1.1%, Alphabet +0.7%, Meta +0.7%, Amazon +0.5%, Apple +0.2%, Nvidia -0.4%)
Amkor Technology (AMKR) rallies 11% after the company announced a $1.5 billion multi-year binding agreement with Nvidia to develop advanced semiconductor packaging and test technologies for next-generation AI and accelerated computing platforms.
Intel (INTC) gains 4% (well below the kneejerk surge 12% higher) after the chipmaker’s third-quarter forecast was much stronger than analysts’ expectations. The results highlighted both the durability of AI-related demand, as well as the success of Intel’s turnaround.
MaxLinear (MXL) slides 11% after the semiconductor device company reported second-quarter results that were only modestly ahead of expectations. While its third-quarter revenue forecast was stronger than expected, its view for adjusted gross margin was largely in line at the midpoint of the range.
Oracle (ORCL) is up 2.6% after the software company said it had been awarded a 10-year IDIQ contract by the US Department of Defense under its Enterprise Software Initiative. The contract is valued at $3.31 billion for the first five years and up to $6.99 billion if options are exercised.
80·ALong
m
meme7/22meme
quote: $GOOGL now has 950M monthly Gemini users and processes 22B API tokens/min.
Compared to 750M back in February.
Absurd growth adding 200 million more users in a few months. Kinda explains why they ran out of compute, then cut allocations to $META and others.
Very unlikely AI capex will slow down given. | $GOOGL reported earnings today:
EPS: $9.11 actual vs $2.91 expected - retail reactions to EPS blowout is kinda just accounting noise, since $SPCX, Anthropic, and others were likely large contributors.
Revenue: $119.7B actual vs. $116.98B expected
Google Cloud Revenue: $24.77B +82% Y/Y, vs. 63-65% expected ($22.2-$22.6B) - this is a very large beat and most material part so far.
Cloud Operating Income: ~$8.81B, implying 35.6% margin vs. 32.9% Q1.
Capex: $44.92B vs ~$44.15B (basically in line).
So initial read through is Google Cloud accelerating growth with expanding margins is genuinely bullish for AI demand.
Their former $180-$190B capex guidance is already extremely large, and as long as we get around these numbers in the earnings call + forward projections...
Should be good to go for $LITE and the other optics/networking trade. (eg. last earnings, they said DC and networking would be ~40% of capex spend).
Hyperscaler earnings transcripts are probably the most important thing to pay attention to with the AI capex trade.
And that should be in 3 minutes.
80·ALong
t
twitter7/21meme
quote: havent seen one person from OAI or Ant address Jon's argument here.
the point is simple: the USG does not owe either of the large labs a business model. if the economics of selling tokens don't work due to distillation/cheap clones/Chinese AI magick, the American enterprise and consumer will be A-OK. they will benefit from hyperdeflation in the cost of digital cognition just like everyone else. the hyperscalers will be fine. it's just OAI and Ant that won't be – in their current forms at least. if they are willing to adapt, they can develop new business models.
so what if the token merchants don't do well? the neoclouds will be fine. the internet companies will be fine. the consumer gets cheaper queries. the enterprise will still incorporate AI.
the only world in which this isn't fine, is if you hold a quasi-religious belief that we're on the cusp of a kind of AI rapture in which one of the labs Logs On And Wins Forever, namely hits RSI and we enter some kind of sublime post economic society run by GEOTUS Dario. so to accept that Ant's business model might be suboptimal or impaired by China's commoditization is to accept the unacceptable; namely that someone other than the anointed might kick off the runaway feedback loop and that they, instead might log on and win forever.
this appears to explain the discrepancy in reaction to Deepseek Moment v254 Kimi edition. everyone has bag bias, of course. but leaving that aside, most people think it's pretty much ok if Ant and OAI suffer margin compression due to Chinese distillation / industrial sabotage via open weight models. the American economy is not reliant on those two firms. they could blink out of existence and we would pretty much be ok. the AI capex supercycle will still produce tokens, closed weight or not. American firms will consume those tokens. OAI and Ant would probably still scratch a living, due to the latent preference of some token consumers to buy domestic and face off against a known entity.
this is only unacceptable if you think AI is strongly path dependent; that is, if it really matters who the market leader is when AI reaches a breakout level of capability. this is true both in the good case (superintelligence, singularity, etc) and the bad case (this is the essence of safetyism). but if this sounds more like wishcasting than forecasting, you probably don't mind the labs being pressured economically.
now you can clearly tell which side I'm on. I think AI is a fantastic technology which is hyperdeflating the cost of cognition and will fundamentally reshape society but there are real reasons why it wont diffuse as fast as the AGI people think it well. I would prefer an American firm achieve RSI relative to a Chinese one but I think either outcome would be suboptimal; better that we don't end up with a closed oligopoly composed of Ant/OAI. China by crushing the margins of the labs is doing everyone a favor by eliminating their pricing power and empowering the buyers of AI, namely, everyone.
objections:
-but you can't celebrate America losing to China!
- in my opinion this is a minor victory for China but not necessarily an enduring one. USA still has the chip, datacenter, and neocloud advantage, not to mention, it still has the best frontier models. Chinese labs releasing open weight models have no business model of their own. so even if they hurt the US labs, they have nothing to show for it. it's profoundly unlike their successful dumping campaigns with solar panels, batteries, drones, etc where they eventually built big domestic industries. (if China kills American AI with open weight models, we can even the score the moment they try and release a proprietary model). even if open weights win, the USA can still leverage AI extremely well and potentally retain the aggregate compute advantage. yes, the US would be more assured of victory if OAI or Ant won forever, but I don't know if I want to live in that world.
- no one will ever train a model again
- this is where I think the concern is unwarranted. let's say distillation really is a golden bullet and kills big training runs. that doesn't advantage either China or the US. that's a stalemate. not to mention, the trend seems to be less focusing less on massive pretraining budgets and more on finetuning for specific genres of tasks, thinking machines style. and lastly I find it hard to believe that training runs will stop altogether. the labs can probably develop anti-distillation techniques. you could adopt a whitelist style permission for everyone using your model. different consortia could be put together to share in the cost of training a model, if it is seen as too expensive for an individual firm.
- the AI buildout is path dependent and OAI/Ant are now load bearing GDP infrastructure
- it would be a significant setback for investors if they had to cancel their IPOs and suffered big markdowns, and some neoclouds with lab based RPOs would suffer for a while, but everyone would be fine, really. does Microsoft need OAI or Ant? does Meta? does Google? ordinary Americans have ~no exposure to either OAI or Ant. would the world want any less compute if it turns out to be another order of magnitude cheaper? certainly not. as we all know at this point, consumption would go up. I don't think the economy is so dependent on the labs that it couldn't handle their margins compressing. | Look I think I've figured this whole thing out. Follow along as I try to steelman, and tell me where I'm wrong.
OpenAI guys on the TL believe that if they can't sell metered inference tokens at a sufficient markup, then they will not have enough of a business to fund the next big training run.
They are surely correct about this.
They believe that if people release powerful open models, this will probably fatally impact their ability to sell inference tokens at enough of a markup to fund the next big training run.
They are surely correct about this, too.
They also think that if they cannot fund the next big training run (again, by selling inference tokens at a markup), then NOBODY will be able to fund the next big training run because it means there's no money in it.
This last bit seems to me & many others to be not just wrong, but totally bananas in a "guy, have seen the actual software industry and how it works in real life?!" kind of way.
There are a lot of ways to monetize software out there in the world. Insofar as inference can add new capabilities to software, there will be lots of ways to monetize it.
In other words, if you're telling me, "we can't have a business selling inference if X or Y thing keeps happening," then my only response is, "ok well that sucks for you... sounds like that's a terrible business."
But if you're telling me that "selling metered inference tokens is a terrible business" is tantamount to "nobody will fund big training runs that are upstream of more effective & economically valuable inference tokens", then I think you are extremely wrong and should get out more and learn about other parts of the software ecosystem.
Workplace automation is huge and will be even bigger in the future as models get better. You can sell workplace automation very profitably in lots of different packages (depending on the workplace and the type of automation). Like, I'm sorry that you really really want to be in the metered inference token business and not the workplace automation business, but them's the breaks. The market wants what the market wants. We all need to live in reality and not beg for Uncle Sam to save us all from open source -- because that was already tried and it didn't work.