Serenity
AI & Semiconductor Supply-Chain Researcher, Trader
Higher-power CW lasers command much higher prices and CPO margins. ASPs rise from ~$3-4 at 70mW to ~$50-75 at 400mW NPO/CPO (Mizuho); CPO laser gross margins exceed 60%. Coherent is probably the best risk-adjusted laser play on its $20B LTA. Lumentum looks unattractive now due to a Marvell-acquisition valuation overhang despite strong operations. Sivers' target of 300m lasers could be enormous against its market cap, but qualification remains uncertain. x.com
Contracted demand makes AI hardware bottlenecks look too cheap for a demand-collapse thesis. POET traded $97M below net cash; memory names from Samsung to ESMT trade at 1-3x forward earnings, while HBM suppliers have multiyear price floors. Coherent and Lumentum have $20B and $5B agreements. A recovery looks likely as earnings and cash flow accrue, especially in bottlenecks. Macro and capex cuts remain risks; Applied Digital's CEO cites build timing despite robust demand. x.com x.com
This signal is no longer available. Showing the latest signals.
Jukan
Semiconductor & AI Infrastructure Research at Citrini
What’s even more surprising is that the White House itself reportedly asked Congress to include a provision in the bill requiring the Commerce Department to issue rules blocking remote access to certain chips. That’s a shift in the White House’s position. If that provision passes, GPU spot prices could take a hit, too. x.com
AlvaChallengesThe Information reportedly says the White House sought rules blocking remote access to certain chips; Jukan warns passage could lower GPU spot prices, pressuring the demand backdrop for hardware investment.Jukan
Semiconductor & AI Infrastructure Research at Citrini
According to The Information, ... The fallout has reportedly made it easier to advance three bills. The MATCH Act would restrict shipments of lithography equipment to China. The Overwatch Act would give Congress greater oversight of the Commerce Department’s decisions to approve AI chip sales to countries such as China. The Chip Security Act would require AI chips to have built-in tracking capabilities. x.com
AlvaChallengesReported momentum for lithography restrictions, congressional chip-sale oversight and chip tracking broadens policy risk to hardware sales and deployment; the measures remain proposals.
First Squawk
Financial Audio News Service
Senators Jim Banks and Elizabeth Warren have urged the Commerce Department to reconsider NVIDIA’s export licenses to China and conduct a full national security review, and are working on legislation that would temporarily restrict China from purchasing NVIDIA chips more advanced than the H200, seeking to include the measures in the annual NDAA x.com
AlvaChallengesThe proposed restriction could limit NVIDIA’s China sales above the H200, adding a policy risk to the thesis’s AI hardware demand outlook; it is not yet enacted.Olaolu Aganga
other
There are between three thousand and 5000 data centers today operating in the U.S. That same amount is in different phases of planning. So yes, our job is to be able to look through the noise and right now the signal is strong. youtube.com
AlvaSupportsEstimates 3,000–5,000 U.S. data centers are operating and a similar number are in planning, supporting a substantial prospective hardware-demand pipeline; those projects remain at varying planning stages.
Rob Arnott
Chairman of Syzygy Asset Management, Founder of Research Affiliates
And what happens if you find that you can, for considerably less energy consumption, deliver the same compute power with new chips, then you suddenly have to write off half of that trillion dollars. There's a good chance that some of these earnings are gonna be reversed in three or four years. I don't think that gets enough attention. I think there are risks that are very real. youtube.com
AlvaChallengesWarns that chips delivering the same compute with less energy could force write-offs of half a $1T investment base and reverse some AI earnings in three or four years, challenging earnings durability.
Deedy Das
Partner at Menlo Ventures
The compute demand isn't going anywhere. ... People are using enterprise AI, AI has come to the enterprise and led to these fantastic productivity gains. ... Whether or not something happens with Jensen, I just think it's a stark reminder when we look at this $177 billion number that the compute will come from Nvidia whether Jensen's at the helm or not. youtube.com
AlvaSupportsReported enterprise AI deployment spend of $177B by September 2026 versus $33B in 2025 supports continued compute demand behind hardware investment; Das expects that demand to persist despite Nvidia's leadership uncertainty.Adam Blake
Co-Founder & CEO of Mavria
I think the thesis is that the demand for compute is coming in all shapes and sizes. ... There are thousands of sites across The US that exist on existing distribution infrastructure. I mean, there's probably a 100 plus gigawatts that if you could use batteries and agree to curtailment, that you could deploy our product on existing infrastructure. podcasts.apple.com
AlvaSupportsEstimates 100+GW could be deployed on existing U.S. distribution infrastructure with batteries and agreed curtailment, offering a conditional route around grid constraints that delay hardware deployment.
Dow Jones
Real-Time Financial News Wire
Trivariate Research Founder Adam Parker ... argues that political backlash to AI is obscuring the sector's growth outlook ... "Earnings per share growth is likely to remain robust for the next few quarters," he writes. tradingview.com
AlvaSupportsExpects robust EPS growth over the next few quarters despite political backlash, supporting earnings accrual as a catalyst for recovery in sold-off AI hardware.
Oksana Aronov
Head of Market Strategy for Alternative Fixed Income at JPMorgan Asset Management
I think the buildout will have to get at least stretched out ... perhaps what was earmarked to be spent through 2030 will take a longer amount of time. I think definitely a re-pricing is very likely here. youtube.com
AlvaChallengesExpects AI buildout spending earmarked through 2030 to stretch over a longer period, delaying the hardware revenue and cash flow underpinning the recovery case and raising repricing risk.
Mandeep Singh
Global Head of Technology Research at Bloomberg Intelligence, Host of Tech Disruptors
[Google's TPU line] could be the largest part of the Google Cloud business, and Google Cloud business could double in 2027. ... We are modeling it to be over $100 billion business in 2028. youtube.com
AlvaSupportsForecasts Google Cloud could double in 2027 and exceed $100B in 2028, with TPUs potentially its largest segment, supporting expanding AI compute demand behind hardware investment.Dow Jones
Real-Time Financial News Wire
Needham analyst John Todaro and his team think that miners face a couple of conditions before they can actually cash in ... around five gigawatts of capacity across major operators like Iren and Hut 8 are waiting on critical regulatory decisions on Dec. 10 and Dec. 17. ... Riot Platforms broke what he called an industrywide dry spell by signing a major lease valued at over $28 billion last week. ... TeraWulf posted third-quarter revenue for its high-performance computing lease of $62.9 million tradingview.com
AlvaMixedFor AI-hardware bottlenecks, the miners' 5 GW of pending Texas decisions and financing needs signal buildout-timing risk, while Riot's $28B lease and TeraWulf's $62.9M HPC revenue show early demand traction.
Paul Christopher
Head of Global Investment Strategy at Wells Fargo Investment Institute
The earnings picture does look good to us through 2027. The problem is the expectations for earnings have to come down a little bit more. ... Some of those order books for chips and other industrial components are full all the way into 2028. ... We are seeing activity on the industrial side to expand capacity. But that can't happen overnight. youtube.com
AlvaMixedSees some chip and component order books full into 2028 and a good earnings outlook through 2027, supporting durable demand, but expects earnings expectations to fall further before recovery.
Hedgeye
Independent Investment Research and Financial Media Company
Four of the six biggest US tech companies posted negative free cash flow last quarter, only $AAPL and $MSFT stayed positive x.com
AlvaChallengesNegative free cash flow at four of six major US tech companies raises the risk that sustained investment needs external funding, testing the durability of customer spending behind hardware demand.Dow Jones
Real-Time Financial News Wire
Morgan Stanley sees a buying opportunity in semiconductors ... "A shift in some projects from 2028 to 2029 could drive greater visibility into a longer cycle and improve the multiples for semis already pricing in a more near-term peak," they add. "We see DRAM, NAND, CPUs, etc. as tight as any point this cycle." tradingview.com
AlvaSupportsMorgan Stanley says tight DRAM, NAND and CPU supply supports the thesis that memory bottlenecks are undervalued; a shift of some projects from 2028 to 2029 may extend the cycle, despite financing-headwind risk.Dow Jones
Real-Time Financial News Wire
AI hyperscalers including Google, Amazon, Meta Platforms, Microsoft, and SpaceX are all well-positioned to handle a new environment for data-center construction following the midterms ... Morgan Stanley analysts write in a note. They leave their forecasts for hyperscaler capital expenditures unchanged, writing that hyperscalers will manage to execute on their projects, either by working with Democratic governors or adjusting to friendlier states. tradingview.com
AlvaSupportsMorgan Stanley expects hyperscalers to execute planned projects despite the midterms, leaving capex forecasts unchanged; this supports contracted AI demand against a demand-collapse thesis.Dow Jones
Real-Time Financial News Wire
If the Democratic midterm victories lead to a slowdown in the AI infrastructure buildout, investors should watch for signs of large AI companies increasingly improving the efficiency of their existing compute capacity, Morgan Stanley analysts write in a note. "We are highly focused on increased (exponential) compute efficiencies from leading labs in '27 to allow products to scale with less compute," the analysts write tradingview.com
AlvaChallengesExpects greater emphasis on compute efficiency in 2027 if builds are delayed, allowing AI products to scale with less compute and posing a risk to incremental hardware demand behind bottleneck earnings.
Eric Balchunas
Global Head of ETF Research at Bloomberg, Co-host of Trillions and ETF IQ
Is Micron a value stock all of a sudden? This seems impossible. SPX p/e is 25 for context. x.com
AlvaSupportsThe attached roughly 9–10x trailing P/E for Micron, versus the cited 25x S&P 500 multiple, supports relative memory cheapness; Balchunas questions whether it now qualifies as a value stock.Dow Jones
Real-Time Financial News Wire
Some hyperscalers may try to shift data centers to friendlier states, and the analysts project potentially 20 gigawatts of capacity spilling over into international markets. ... "Thus the impact of the election skews toward a delay, rather than a reduction, in data center volumes." tradingview.com
AlvaSupportsMorgan Stanley expects 2028 Texas and Ohio projects to slip into 2029 and potentially 20GW to shift internationally, supporting delayed or relocated hardware demand rather than lost volumes.
SemiAnalysis
Independent Semiconductor and AI Research Firm
CXMT is talking up its 4F² DRAM roadmap again. At its October 30 earnings briefing, the company mentioned its first 4F² DDR5 RDIMMs are targeted for sampling by year-end. x.com
AlvaChallengesCXMT's targeted year-end sampling of 4F² DDR5 server-memory modules advances potential DRAM competition, posing a future supply risk to the memory pricing and earnings underpinning low valuations.
Constellation Energy Corporation
Power Generation Company, Largest U.S. Nuclear Fleet Operator
Year to date, we have contracted more than 8,500 megawatts well into the 2040s at compelling economics.
AlvaSupportsMore than 8,500MW contracted into the 2040s, with Google and Amazon among counterparties, supports durable power demand behind AI infrastructure investment and the case against demand collapse.
