As mentioned earlier, there is a lot of fear and FUD in the market right now. The disconnection between fundamentals and sentiment is only widening.
Anthropic and OpenAI numbers alongside falling token prices are being treated like the cycle is rolling over, or that it is the end. Near-term revenue at the two labs does not matter as much as that framing suggests. In 2-4 years there is an extreme chance of multiple OOMs of advancement in capability of models, in usability of models, and of course, the size of the user base. The one thing that stays certain is the need for compute. The buildout is not going to stop because of a quarter or two of financing noise.
What is being priced is the financing of the current leg, layered on the broader yield and inflation backdrop. CRWV paper in the mid-teens, other AI-linked CDS and yields wider, a chunk of data center debt now junk-rated, equity windows narrower for pure play labs and for things like the Solidigm IPO. Oil higher, inflation expectations, rate cut odds nonexistent, etc. Capital is selective and expensive.
Though, that is also the part that can and will reverse. Physical orders have not disappeared and Nvidia is still prepaying packaging and substrate capacity through 2028. Sentiment has run well ahead of any actual structural break in demand. If the next model leap shows up on schedule (my timeline is 6-8 months max), or if cash flows at the labs just improve enough to reduce the external financing burden, the credit currently pricing distress on the leveraged edge can tighten. The demand for the chips, the power, packaging, etc., does not require the current fear or the current yield structure to stay intact.
$CRDO dropped on Friday after $MXL 's CEO mentioned on their earnings call that they are targeting ~25% share of 1.6T AEC market despite not expecting their first revenue until year 2027.
For context, MaxLinear have not yet shipped their 1.6T DSP Rushmore after a few engineering samples didn't offer good-enough results. Revenue for that product is expected to ramp in 2028 so definitely an execution issue there, but that's just from channel checks and industry discussions. I can't see them executing and taking a large chunk of market share from Credo's AEC business, but hyperscalers prefer multi-sourcing and at the end of the day, the AEC market , and Credo will maintain the lead in that growing market.
Credo's stock droppe mostly on Edgewater mentioning that SpaceX might not use 1.6T AECs anymore, as their new data center architecture requires 10m 1.6T AECs when the max they can reach is 6.5. Something ALCs can resolve. Perplexing dynamic to me, but definitely something worth digging more into.
ICYMI: recap of this week at @PredictQmkt
New tools / site improvements
- Parlay/Combo fair price tool launched
- $10k freeroll + Survivor Revival Optimizer added
- Trading dashboard UI improvements
The following articles:
- Model output for fantasy football markets
- Week 7 NFL Predictions
- NFL Futures Markets recap
- Iowa Senate race breakdown
- Midterms analysis
- Predicting when the ski slopes will open
- Culture Radar
And two shows:
- an interview with @samschwartz71
- Thursday night NFL breakdown for the weekend
Thoughts on memory:
Anyone who thinks compute will stay in shortage while memory slides into oversupply must be smoking something. Compute and memory are destined to tighten together or deflate together. One cannot die while the other survives.
By the same logic, the pods shorting memory and going long HDD are out of their minds. How do you think HDD isn't joined at the hip with memory? You really think HDD stays safe if this cycle rolls over?
In the end, all of this comes down to nobody having conviction about what happens after 2027. Historically, we have watched P peak one to two quarters before supply ramps, then Q expands and the entire profit pool collapses. But everyone forgets that this cycle runs on nonstandard products, unlike PCs, smartphones, and general purpose servers, and that it was triggered by factors different from past cycles.
This is because of the boomers who still see memory as an industry where you make a killing in the upcycle and lose your shirt in the downcycle. These self proclaimed veterans view memory through the lens of their past experience, and since they drive the market and set the stock prices, memory trades at such a low valuation. What I keep saying is that a sharp collapse like in the past is impossible. A gradual decline is possible. But even if Samsung and SK Hynix margins were cut in half from current levels, they would still trade at a P/E in the teens. That is still roughly half of HDD valuations.
Personally, I think that as memory prices fall, other sources of demand will absorb the supply at each step down. In particular, I expect personal AI demand, led by local AI, to emerge. Looked at another way, for local AI to take off, the hyperscaler driven spike in memory prices has to be brought under control. At current memory prices, local AI can never diversify.
Just as mainframes gave way to PCs and the cloud gave way to smartphones, each time carrying the cycle from P into a Q expansion phase, I think this AI cycle will also move from frontier labs to local open source AI. Based on these historical precedents, I believe local AI will take over the memory cycle from the frontier labs and carry memory demand.
I am not saying memory prices will hold for more than a decade. Memory P will fall at some point. But what is different from the past is that new consumer devices and new consumer demand, such as open source driven local AI, will support the floor for memory.
So the market right now keeps discounting these stocks out of fear of a memory peak nobody can time. This discount is being driven by boomer investors who dismiss the idea that AI supports the downside and keep pushing the view that earnings will collapse and the trough will revert to past levels too. My point is that this is completely wrong.
I don't think that EDA for semis will get "oneshotted" anytime soon.
You need special relationships with companies to get access to data.
People don't understand, PrimeTime is valuable not because the software exists but because it is in the TSMC ecosystem and has been used in 100K+ tapeouts(lower bound).
Why is OpenAI working with Synopsys instead of making its own tools.
My conversation with @alexandr_wang, chief AI officer of Meta.
0:00 The vision behind Muse
1:32 Discovering the potential of personal AI agents
3:34 Three years of therapy in one AI session
5:46 Building Muse in 7 months
9:16 Why Meta bet on personal AI
16:51 Mark Zuckerberg's obsession with improving products
20:15 How Meta knew Muse was ready to launch
23:32 Why great products need a single point of view
25:56 Making Meta cool again with memes
34:38 Building trust with AI agents
38:43 The future of human ambition
45:22 The call from Mark Zuckerberg that changed everything
52:13 Leaving Scale AI for Meta
55:19 Rebuilding Meta's AI lab from scratch
58:13 From founder to coach
1:05:55 The future billionaires at Alexandr's 19th birthday party
1:08:04 Diamond mining vs. building skyscrapers
1:09:21 Why Alexandr has more than 200 direct reports
Includes paid partnerships.