The jobs report was a barnburner. Nonfarm payrolls increased by 172,000 versus expectations for 88,000, while prior months were revised higher by 93,000. Wage growth came in at roughly 0.3%. Yet the market sold off. In our view, the market is misreading the signal. It is assuming that stronger than expected employment and growth will cause a an acceleration in inflation. History would suggest otherwise. Productivity growth is running near 3%, while unit labor costs are hovering around 0.5%. Those are not the hallmarks of an inflationary boom. They are the hallmarks of healthy, productivity-driven growth that will lower inflation. Meanwhile, the yield curve continues to flatten despite a roughly 55% increase in oil prices year-over-year based on a three month moving average. In past cycles, an energy shock of this magnitude steepened the yield curve when the Federal Reserve was accommodating it. Instead, the bond market appears to be discounting something much more powerful: the deflationary impact of technological innovation, particularly artificial intelligence, which is beginning to increase productivity across broad swaths of the economy. If tensions with Iran ease and oil prices retreat, we believe inflation could move into negative territory before year-end. In our view, the Fed made a historic policy error when it raised rates aggressively into what was largely a supply-driven inflation shock in 2022. We do not believe the next generation of monetary policymakers will be eager to repeat that mistake. Notably, gold peaked on the day Kevin Warsh was appointed. The inflation trade may already be behind us. If our research is correct, the next phase of this cycle could be characterized by accelerating growth, declining inflation, falling interest rates, and a strengthening U.S. dollar. That combination would create a remarkably supportive backdrop for innovation-led equities and the technologies driving the next productivity boom. I discuss this framework in greater detail in this month’s episode of In The Know.
Nebius Deep Dive & new 2027 price target is live!
Fuck it, I just spent my entire Sunday morning so I could get it out today!
I spent multiple days working on this. Hope you’ll notice!
This time, I also added a lot of infographics and comics to make the very long post more digestible and fun to read.
Hope you like it, and I would be super grateful if you can share it / recommend it to people you feel would get value from it!
Link to article in the comments.
Here is the problem, over the past 5 years many equities have 10x Bitcoin return, even over 10 years NVDA beat it - the digital gold narrative may be true, but BTC is a retirement asset now, if you want present day lifestyle returns, equities is where its at.
🚀 Space → ASTS +1,558%
🤖 AI → NVDA +1,400%
💊 Pharma → LLY +480%
🏗️ Infrastructure → VRT +1,300%
☢️ Nuclear → CCJ +470%
🪨 Metals → AMR +1,080%
💻 Tech → SMCI +1,100%
₿ Bitcoin → +108%
If you are tired of waiting for the alligator jaws to close, cash a little out and come play with @ShardiB2
Why I Think $NBIS Nebius x Anthropic Is Happening
I believe Nebius and Anthropic will soon make a deal, or already have a deal and will soon announce it.
Quick disclaimer: this is pure speculation and not financial advice. I could be completely wrong. Please do your own research.
My full thoughts below:
1/
Yesterday, Leopold Aschenbrenner, @leopoldasch, already a legendary AI infrastructure investor and a friend of SemiAnalysis founder Dylan Patel, @dylan522p, disclosed a massive 5.6% stake in Nebius.
This is likely his largest long position by far right now.
His wife is Chief of Staff at Anthropic.
I do not think she informed him about any deal. But I do think Leopold is one of the best-connected investors in AI infrastructure. If there are serious rumors around major AI infrastructure deals, he would likely hear them very early.
He might also simply see this as a logical next step.
2/
In the recent $NVDA earnings call, Jensen once again explained that “compute equals revenue.”
He argued that for model makers like Anthropic, every GPU they can’t get is a direct hit to their top-line growth.
If legacy hyperscalers like AWS and Google are bottlenecked, and they are, Anthropic has no choice but to go to “neoclouds” like Nebius that have the specialized high-density clusters ready to go.
Jensen also explicitly mentioned that Nvidia is prioritizing supply to “AI-native companies” and specialized providers who can deploy at maximum speed.
3/
Nebius recently raised its 2026 CapEx guidance from the previous $16-20B range to $20-25B.
Management stated that the increase was to “accelerate capacity buildout particularly for 2027 demand that is already partially committed.”
Given that Meta already has its $27B deal and OpenAI is anchored to Azure, Anthropic is the most logical frontier client left who needs that specific scale.
4/
On March 2, 2026, the Independence City Council approved a $150.6B Chapter 100 bond issuance for Nebius.
Officials call it a legal ceiling.
Even though it is a legal ceiling, you don’t set a ceiling at $150B unless you are planning for multiple generations of frontier-scale hardware.
To finance a bond of this magnitude, Nebius likely needs a guaranteed tenant.
In the industry, this is called a take-or-pay contract, meaning the tenant pays for the power / compute whether they use it or not.
Very few companies currently have the capital and the compute hunger to anchor a $150B roadmap:
5/
Nebius recently made headlines by breaking ground on a massive 1.2 GW AI factory in Independence, Missouri.
They are also working on multiple data centers in Europe, where Anthropic is probably insanely short compute.
For inference, data centers should be close to end users. What are the actual options in Europe, other than Nebius?
There are basically only the hyperscalers and Nebius.
I think it would be smart for Anthropic to not depend 100% only on GCP and AWS and to diversify more across $CRWV, which they are already doing, $IREN, yes, I think they will clearly get an Anthropic deal as well, and obviously $NBIS.
6/
Nebius is also aggressively expanding into APAC.
Just a side note. Anthropic will also need capacity there as well.
On May 27, 2026, Anthropic signed a massive strategic partnership with Fujitsu to deploy Claude to 100,000 employees and mission-critical government systems in Japan.
Japanese enterprises and government agencies are notoriously strict about data residency. They want the AI to run on hardware located in or near their region.
Again, this is needed for inference anyway.
Just two months ago, Nebius officially launched its APAC expansion, appointed a General Manager in Singapore, and targeted specifically Japan and South Korea.
Anthropic just promised Japan a sovereign AI experience, and Nebius is building out APAC capacity.
7/
Nebius was officially named as a primary NVIDIA Cloud Partner for Vera Rubin among an elite group: AWS, Azure, Google, and Nebius.
During the launch of NVIDIA’s Vera Rubin platform in March 2026, Anthropic CEO Dario Amodei was one of the few lab leaders to provide a direct quote for the press release.
He specifically noted that Anthropic needs infrastructure that can “keep pace” with Claude’s agentic workflows.
8/
For years, Anthropic was seen as “Amazon’s lab.”
That changed on April 6, 2026, when they signed a 3.5 GW TPU deal with Google and Broadcom.
Anthropic’s revenue run-rate has reportedly exploded from around $9B in late 2025 to $30B+ in May 2026.
At this size, they are no longer a startup that can live in one cloud.
Anthropic must make sure they are not held hostage by one infra provider. Diversification is key to maintain good deal conditions.
9/
Of all the frontier labs and hyperscalers, Anthropic seems to be by far the most compute-constrained.
Rumor is they want to IPO soon.
There are many very strong hints, I would almost call it a forced IPO:
a) They have engaged the law firm Wilson Sonsini, which has a deep history with tech IPOs.
b) Preliminary talks have reportedly taken place with major investment banks, including Goldman Sachs, JPMorgan Chase, and Morgan Stanley, to serve as lead underwriters.
c) Insiders have pointed to October 2026 as a feasible window, though some caution that the timeline remains fluid depending on market conditions.
d) Then there’s just the obvious fact that the market is very hot right now and you never want to miss the IPO window.
There are more reasons, like their valuation forcing them to go public, as private markets are bigger than ever before, but they are not infinitely deep.
Basically: they will have to go public, and I think soon.
One major risk they’ll have to explain and talk about constantly:
Where will you get more compute from?
Isn’t capacity essentially your growth bottleneck?
Just like Nebius was almost forced to buy Bloom Fuel Cells, Anthropic will be forced to buy as much capacity as possible.
No capacity, no growth.
No growth = DEATH for a hypergrowth company like them.
Since their entire business is also insanely capital-intensive, a high valuation that offers access to cheap capital is obviously super important.
Long Nebius $NBIS baby.