The Social Community forThe Most Important Thing

Tesla and #SpaceX 🚀 are deepening ties (https://www.theinformation.com/articles/elon-musk-building-terafab-team-inside-tesla?offer=rtsu-engagement-25%2Crtsu-featured-articles-pro&cio_link_id=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__f25baf539a1e92aac8bbdf5f7219971f4a85e2bde90ab7584453ee595c9b6265&utm_campaign=RTSU%3A+N55+Characters&utm_content=13999&utm_medium=email&utm_source=cio&utm_term=9699)in a chip push as merger chatter swirls.

All the talk of “SaaSpocalypses” starts with a core assumption: very soon, “vibecoding” will be able to replace the kinds of sophisticated software solutions on which enterprises once invested millions each year. But is any of this actually TRUE? Consulting firm Bain & Company is reportedly trying to find out. Financial Times reports that B&C are attempting to vibecode their own versions of products from hundreds of potential acquisition targets, to determine both the companies’ long-term value and how their offerings might evolve in the future. The project started with a small team of software engineers in 2023 but has now reportedly exploded, and involves a large number of the firm’s consultants and associates. According to FT, this has already moved out of the theoretical space, and Bain’s vibecoded alternatives are influencing acquisition and investment decisions in real-time. One private equity exec told the paper that they’ve slowed their deal making in light of Bain’s results, and are applying additional scrutiny to any and all SaaS ventures, while an other investor said that they dropped out of the bidding for an analytics company after seeing Bain’s vibecoded mock-up. But not everyone is convinced that the strategy is entirely sound. On X, Chamath Palihapitiya suggested that vibecoding a mock-up of a powerful piece of software doesn’t necessarily give Bain or its clients a deeper or more nuanced understanding of the underlying business or technology. He asks “Do you think vibing the UI of Google search would even scratch the surface of helping you understand the complexity that sits behind it?” Ultimately, Chamath dismisses the initiative as “busy work,” and another example of the already-outdated push for AI “tokenmaxxing.” https://www.ft.com/content/e5bac4d1-b1f8-43a4-bd54-b182d5357af0?_bhlid=1136a3aaa633ea45fdaa15eab9f79128d5ff2ad8

Maybe, just maybe, big tech firms need to do a better job of explaining their AI investment strategies. Shares of Google parent Alphabet fell 7% on Thursday after the company disclosed a further increase in capital expenditures this year—money it will spend on new AI chips, servers and data centers. The drop all but wiped out the gains the stock had enjoyed so far this year. Sure, the entire stock market sold off, presumably as a result of rising oil prices and the intensification of the Iran war. But Alphabet was among the worst hit, along with Tesla, which also reported sharply higher capex on Wednesday and whose shares fell 15% on Thursday. It’s tough to consider Tesla’s stock valuation on any kind of rational basis, given that it trades based more on its connection to Elon Musk than reality. But Alphabet is a different animal. There’s little doubt Google is among the best positioned of big tech companies in terms of its ability to take advantage of AI advances, given its strength in chips, AI models and consumer apps that should benefit from AI. But even for Alphabet, you have to wonder how much more money the company will have to spend to reach the promised land. Alphabet’s projected capex for this year—now as much as $205 billion—is more than double what it spent in 2025, which in turn was nearly three times what it was spending two years before that. Next year, Chief Financial Officer Anat Ashkenazi said on Wednesday, capex will “increase significantly” again! Dan Flax, an analyst at Neuberger Berman, said today that Alphabet’s capex “will likely be well over $300 billion.” For context, analysts expect the company to generate $259 billion in cash from operations next year, according to S&P Global Market Intelligence. That suggests Alphabet could burn $40 billion in cash next year, which is quite a turnabout for a company that before this spending surge routinely generated $60 billion to $70 billion in free cash flow. And when will this end? Alphabet executives are vague about it. CEO Sundar Pichai keeps talking about the industry being in the “early innings” of the AI transition, which implies the heavy spending could go on for a while. Ashkenazi said on Wednesday that “our goal is to invest as long as we see an attractive return on that investment.” In other words, trust us. That may not be enough anymore.

Advanced Micro Devices said Wednesday it would make an equity investment of “up to $5 billion” in Anthropic “in the future,” and the Claude AI maker would use AMD’s AI server chips starting next year. Anthropic has been steadily diversifying its AI server sources and is evaluating chips from other providers including startups and Microsoft, The Information has reported. AMD implied that Anthropic would agree to purchase a substantial number of its server chips, which would eventually consume up to 2 gigawatts of power, and their announcement also implied AMD would purchase access to Claude models to help it design better chips. AMD shares rose 2% and have already risen nearly 150% this year. AI developers OpenAI and Meta also have announced agreements to use AMD chips, though it isn’t clear how much they will end up using or when. AMD’s stock has surged as numerous cloud providers purchase its hardware to diversify from Nvidia’s. However, Nvidia is nearly seven times bigger than AMD in revenue and is growing its revenue about 30 percentage points faster than AMD (while also experiencing accelerating sales), underscoring the gap in performance. Besides Google’s DeepMind, Anthropic is one of the other major AI developers that doesn’t heavily rely on chips from AMD rival Nvidia; it has developed its models using chips from Google and Amazon, for instance. But even Anthropic has started paying billions of dollars a year for Nvidia chips as it races to get more computing capacity to meet customer demand. Anthropic’s private valuation is nearly $1 trillion, meaning the AMD investment would represent about 0.5% of Anthropic shares at the current price. Anthropic has filed to go public as soon as this fall. https://ir.amd.com/news-events/press-releases/detail/1292/amd-and-anthropic-announce-strategic-partnership-to-deploy-up-to-2-gigawatts-of-amd-instinct-mi450-series-gpus

Talk about a nervous market! Shares of Nvidia fell 5% on Monday on news of a $500 billion partnership the AI chip giant struck with South Korean conglomerate SK Group, along with a Wall Street Journal report that Nvidia was separately in talks to help OpenAI finance a $500 billion data center campus in Ohio. Investors should take a chill pill: Neither of these deals was actually news. Nor were they fully fleshed out. For instance, Nvidia and SK have only signed letters of intent to formalize a partnership, they said on Friday. Letters of intent are not worth the paper they’re written on. Remember, Nvidia and OpenAI signed a letter of intent last September for what they called a “landmark strategic partnership,” only to abandon the idea a few months later. Moreover, the SK Group–Nvidia announcement from Friday was mostly a reprise of two other announcements from early June. In one, Nvidia and SK Telecom (a part of SK Group) had announced plans to build a “gigawatt-scale AI Cloud in Korea” using Nvidia’s DSX technology. (Funnily enough, while Friday’s press release linked to that announcement, the latest update doubled the scale of the AI cloud in Korea to 2 gigawatts.) In the other, Nvidia and SK Hynix (another part of SK Group) announced a memory chip partnership. As to details of the $500 billion, such as which company is putting up how much, Nvidia isn’t saying. My colleague Phoebe Liu reported on Friday that executives at a press briefing wouldn’t talk about those details. Then there’s the OpenAI-Nvidia Wall Street Journal report that drew attention on Monday, which wasn’t news: It was outlined in a report in The Information in early June! As was the case back then, the details aren’t clear. It does imply Nvidia would take on a lot of risk to back OpenAI. But given the history of OpenAI and Nvidia, there’s no guarantee this deal will get done. Nvidia likes to make announcements. We got two more today, with news that it is investing in Ilya Sutskever’s AI startup, Safe Superintelligence, and helping launch an open-source software cybersecurity alliance. Keeping track of what’s new and meaningful about AI isn’t easy, and the media’s reporting often makes things more confusing by recycling different versions of the same news. But before investors panic about Nvidia’s financial commitments, they need more details about what’s real and new and what’s not. https://www.wsj.com/tech/ai/nvidia-in-talks-with-openai-to-guarantee-250-billion-financing-for-data-center-3dd6eae3?mod=tech_feat1_ai_pos4&utm_campaign=article_email&utm_content=article-17527&utm_medium=email&utm_source=sg

Roger Lee changed their vote · 2 weeks ago

a judge blocked the Ellisons’ Paramount Skydance from completing its $110 billion purchase of Warner Bros. Discovery for 14 days. Closing of the deal had appeared imminent: Paramount had received a spate of regulatory approvals from countries around the world. The U.S. government had already signed off on the deal, but a bunch of states went to court to block it. And Monday’s ruling, by the judge hearing the case, suggests the states have a decent chance of winning. That would be disastrous for Paramount. It’s already paying a high price for WBD—Netflix abandoned a previously arranged purchase of half of WBD rather than try to match Paramount’s offer. If the court case delays Paramount’s completion of the purchase after Sept. 30, Paramount has to pay a fee of $650 million for every quarter the closing is delayed. And if regulators kill the deal, Paramount will have to pay a $7 billion termination fee to WBD. That’s on top of the $2.8 billion it paid earlier this year to break up the WBD-Netflix deal. So you can imagine a scenario where the current deal falls apart and Paramount is out roughly $10 billion. Paramount shares fell 2% to $8.57 today. That’s well below the $12 to $16.02 price at which a group of investors is putting in new equity to help fund the WBD purchase. The Ellisons’ investment in Paramount is suddenly looking a lot less promising. https://www.wsj.com/business/media/judge-temporarily-blocks-paramount-warner-deal-a9c6af43?mod=hp_lead_pos3&utm_campaign=article_email&utm_content=article-17490&utm_medium=email&utm_source=sg

The recent memory chip selloff offers investors a new opportunity to jump into the sector. Samsung is the best bet (https://www.theinformation.com/articles/samsung-new-memory-chip-underdog-now?offer=rtsu-engagement-25%2Crtsu-featured-articles-pro&cio_link_id=eyJlbWFpbF9pZCI6ImRnU2kwUVlEQUplSkJwYUpCZ0dmZ0FxWnp2c2hZSnVSTnB5WUJrcz0iLCJocmVmIjoiaHR0cHM6Ly93d3cudGhlaW5mb3JtYXRpb24uY29tL2FydGljbGVzL3NhbXN1bmctbmV3LW1lbW9yeS1jaGlwLXVuZGVyZG9nLW5vdz9vZmZlcj1ydHN1LWVuZ2FnZW1lbnQtMjUlMkNydHN1LWZlYXR1cmVkLWFydGljbGVzLXByb1x1MDAyNmNpb19saW5rX2lkPUNJTy0tTElOS0lEXHUwMDI2dXRtX2NhbXBhaWduPVJUU1UlM0ErU2Ftc3VuZytJcyt0aGVcdTAwMjZ1dG1fY29udGVudD0xNDIzMFx1MDAyNnV0bV9tZWRpdW09ZW1haWxcdTAwMjZ1dG1fc291cmNlPWNpb1x1MDAyNnV0bV90ZXJtPTk4NzkiLCJpbnRlcm5hbCI6ImEyZDEwNjAzODY0YTk3ODkwNiIsImxpbmtfaWQiOjg2NDM2fQ__78ca81c10356e79c74244627892e95b58455d370dcff128d8b5b345ed5d6e7a7&utm_campaign=RTSU%3A+Samsung+Is+the&utm_content=14230&utm_medium=email&utm_source=cio&utm_term=9879).

AI tokens aren’t just changing how corporations do business. The massive explosion in token usage has fundamentally changed the dynamics of the memory chip industry. For many decades, the DRAM industry suffered cyclical downturns every two to four years when market demand for computers would stall. Chip equipment makers, in particular, were hit hard. Teradyne co-founder Alex D’Arbelloff once told me: “We had been growing very, very rapidly through the '60s, almost doubling every year, and then came the recession of 1970. For three months starting in mid ‘70 we didn't get a single order. It was really a shock, the first of the cyclical downturns.” Today, nobody is forecasting a cyclical downturn in memory anytime soon. It’s a real eye-opener to see the share price history of Micron Technology (DRAMs) and Teradyne (test gear) over the past five years. Both languished until a year ago, and have since skyrocketed. Chinese memory makers are also taking advantage of the newfound investor fervor for semiconductor memory. More than 9.4 million investor accounts applie d for shares in the blockbuster initial public offering (IPO) of ChangXin Memory Technologies (CXMT), with retail demand reaching about 212 times the shares initially available online. “For A-share IPOs in China, winning an allocation [usually] means making money,” said Kevin Chen, a 35-year-old employee at a state-owned enterprise in Shanghai, who also subscribed for shares. Professional investors also showed strong interest. Among them,153 private fund products managed by High-Flyer Quant, the hedge fund co-founded by DeepSeek founder Liang Wenfeng, submitted bids for a combined 12.55 billion shares. The Hefei-based CXMT is seeking to challenge Samsung Electronics, SK Hynix and Micron. The Chinese company currently holds a 9 per cent global share of DRAMs measured by bits, according to Counterpoint Research. https://www.semi.org/en/Oral-History-Interview-Alex-d%E2%80%99Arbeloff

A Chinese company has begun manufacturing one of the key pieces of equipment used in chip manufacturing (https://www.theinformation.com/articles/china-starts-mass-producing-homegrown-duv-chipmaking-tools-advance-local-chip-industry?offer=rtsu-engagement-25%2Crtsu-featured-articles-pro&cio_link_id=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__8f2b3d61f0380cd8920b03394d74226b6f59db594e301dda6ed60c27592ff1e2&utm_campaign=RTSU%3A+China+Starts+M&utm_content=14318&utm_medium=email&utm_source=cio&utm_term=9951), a key step towards self reliance in chips.

PayPal has a message for Wall Street and would-be buyers: “We got this.” That’s the upshot of CEO Enrique Lores’ response to what he called “recent M&A speculation”—also known as Stripe’s reported $53 billion takeover bid for the struggling payments firm. Lores wouldn’t comment about anything specific, but his comments implied Stripe’s offer didn’t meet PayPal’s bar, and the board was sticking with its existing plan for managing the business. Lores said the right things, to be sure. The board is “open” to offers and would compare any with its own plan and choose “the option that creates more value.” His plan, which he said “will create significant value for our shareholders,” involves things like speeding up Venmo’s growth and expanding its payment services business (which includes Braintree) and getting affluent consumers to use PayPal more. When Lores was asked why these efforts would succeed where PayPal’s previous turnaround measures hadn’t, he repeated what he’d said a few moments before and added that he was “improving execution, improving accountability.” In other words, he would run the company better than in the past. It’s not exactly a persuasive answer. The Stripe offer, reportedly made with private equity firm Advent, was too low—there’s no doubt about that. It’s possible it has elements that made it unappealing, such as a lack of secured financing. Perhaps the board is negotiating with Stripe and Advent behind the scenes to lift the offer. The board might be quietly exploring other options. But it’s common in these situations for a board to publicly open an auction up to all comers as a way of seeing what someone might pay. PayPal hasn’t done that. It’s leaving shareholders to trust that the board knows what it’s doing. That’s asking a lot. After all, PayPal has been a dumpster fire in recent years, with its stock falling roughly 80% since 2021 as growth has slowed to a crawl and rivals like Apple Pay have dug in. Lores, a former CEO of HP, has only been in the job since March. He has no prior experience running a payments firm. It may be relevant that no one on the board has a significant stake in the company. As of earlier this year, the entire board’s combined stake was 5.66 million shares, or 0.6% of the shares outstanding. Would PayPal’s board take a different stance if directors had a bit more skin in the game? Just asking. https://www.theinformation.com/newsletters/the-briefing/stripes-paypal-bid-puts-payments-firm-play-musk-jump?utm_campaign=article_email&utm_content=article-17536&utm_medium=email&utm_source=sg&rc=gzsl71

Intel secures the Apple bag? POTUS announced  that Apple will use Intel fabs in the United States to make its chips in the future. Recall that Apple has become a quiet silicon giant, designing its own processors for its computers (both handheld and desktop-oriented). If Cupertino shifted its chip manufacturing from TSMC to Intel, it would be a coup for the American, and partially state-owned, company. Thus far, the companies aren’t commenting. Let’s see if Trump jumped the gun, or is wrong altogether. https://truthsocial.com/@realDonaldTrump/posts/116769225357410422?_bhlid=3b17e04a85a591c0ccfff52f006af80510260980