Cost of using AI models has collapsed according to JP Morgan.
While this is good for the diffusion of the technology, it doesn’t bode well for the foundational model providers as it signals commoditization.Remaining margins will be competed away once the model development flatlines.Value creation will shift to the application layer, and big AI labs supplying foundational models will be more like internet service providers.The technology itself will get occasional upgrades (like we had in telecom with 3G, 4G, 5G) but it won’t be exclusive to any model provider so it’ll remain as a commodity.I think their faith will be parallel to internet service providers. They’ll make massive investments and really change the world, but they won’t see high return on investment as their product will have already become a commodity when they decide to harvest.ImageFor
1.We might get a big sell-off in the US market if Japan raises interest rates.They approved a stimulus because they finally have positive inflation so companies can invest and grow and people are better off spending then saving. They don’t want to lose it.But they are also considering raising rates because their Debt/GDP ratio is above 200% and the central bank can’t print money at no cost anymore as inflation isn’t 0.Thus, any increase in government bond yields skyrockets their interest payment and yields are surging as investors are pricing higher inflation due to stimulus and internal spending.They are stuck between a hard place and rock.If they raise rates, many investors who borrowed cheap in Japan and invested in the US will have to liquidate their positions and close their debt befo
I’ll tell you why Burry’s $Palantir Technologies Inc.(PLTR)$ short won’t work.Actually, no PLTR short will work soon.Those who are shorting the stock are the same idiots who shorted Bitcoin at $50k.They don’t understand the nature of it.PLTR isn’t about fundamentals or valuation, it’s an accord between the retail investors and the company.Retail trusted it to be a vehicle for the redistribution of wealth.They put their money in the stock, and never sold regardless of fundamentals, valuation etc, so the stock has consistently gone up.Management accepted this as they are also mavericks, they remained loyal to the accord by not doing something stupid like fraud etc.As the business also performed well, traders and institutions joined, creating immense
1/ $Novo-Nordisk A/S(NVO)$ was once a market darling, now everybody hates it.Yet, growth is still strong, the oral Wegovy is coming this quarter, and it has the strongest pipeline in weight-loss drugs.Here is why NVO is a 2x opportunity now: 🧵2/ Let's set the stage first: What does NVO do?NVO is the global leader in diabetes treatment with a 33% market share.This market has three main segments:- Oral Anti-Diabetes drugs- Insulin- GLP-1NVO holds the leadership position in all these markets.Image3/ It currently controls 42% of the global insulin market.This has traditionally been Novo Nordisk's biggest business.However, the balance is shifting.Last year, insulin made up only 18% of the shares, while GLP-1 products made up 50% of all sales.Image4/ NVO
1/ $Soluna Holdings, Inc.(SLNH)$ is the cheapest data center stock in the market.It has nearly 1GW capacity under development with 2GW of pipeline.If they can execute, the stock can easily make 6x from here.Here is my SLNH investment thesis: 🧵Mountainous landscape with layered ridges under a hazy sky features several white wind turbines on green hills in the foreground and a winding road leading into the distance. Prominent overlaid text reads SOLUNA in large blue and orange letters with a triangular symbol.2/ SLNH builds data centers near renewable power plants to benefit from excess generation.One-third of the power generated from renewables is wasted due to grid congestion and a lack of storage.Soluna uses excess generation to power data center
$SoFi Technologies Inc.(SOFI)$ earnings first impressions:1) Blockbuster quarter from every angle. Revenue growth accelerated from 34% a year ago to 38%, adding record mumber of users and products.2) It’s becoming increasingly capital-light business. Fee based revenue share has reached 43%, which was just 37% in 2024. This will drive significant margin expansion going forward.3) Lending is up around 10% Q/Q and 23% YoY. This segment is still under its potential and I think it’ll accelerate with rate cuts. Lot of room for growth here.4) Personal loan charge-off rates declined by 92 bps, signaling consumer strength and conservative underwriting.5) It’s now funding 92% of all loans and financial products using its deposit base. As its cost of capital
I’m long $Amazon.com(AMZN)$ , but this is totally unacceptable.AWS is failing to attract AI related business because management thought making their own chips would reduce future costs. Only problem? Those chips have zero developer ecosystem, which wipes out any naked cost advantages.When you account for development speed and performance, $NVIDIA(NVDA)$ GPUs end up having a lower total cost of ownership than custom AWS chips. While AWS tries to push chips that lack a developer ecosystem, rivals win the new business and lock customers in, making any future cost advantages from custom chips irrelevant.Stop trying to challenge the industry standard. You can’t do that when the race is this fast and the stakes
The market is just wrong about $Novo-Nordisk A/S(NVO)$ ..It’s still ahead of $Eli Lilly(LLY)$ in the weight-loss market.Novo’s oral GLP-1 will be launched later this year while Lilly’s oral weight-loss pill is expected in early 2026.Novo’s oral drug led to 13.6% weight-loss after 72 weeks while Lilly’s pill showed 12.4% weight drop.It’s not that different for weight-loss injections too.Novo has Amycretin and Lilly has Retatritude as the upcoming flagship drugs.While they both delivered 24% weight-loss in trials, Retatritude did that in 48 weeks while it only took 36 weeks for Amycretin.In sum, NVO is well positioned to keep its leadership position in the market.Given that this market is expected to reach $1
1. $Amazon.com(AMZN)$ is the cheapest mega-cap stock now.It’s only up 30% since 2020 despite doubling the revenue.It has grown into its overvaluation.Now it’s cheap again, trading at just 19 times operating cash flow.The stock didn’t move because the market believes that AWS is losing the AI infrastructure race as its growth lagged other hyperscalers.However, AWS will reaccelerate as Anthropic’s gigawatt scale Trainium clusters start operation this quarter.Once AWS accelerates, the market will reprice it. Long $Amazon.com(AMZN)$ .Dark-themed financial chart from TrendSpider displaying Amazon AMZN weekly stock performance from 2021 to 2025 with candlestick patterns in green and red, price scale on right fr