Not while JGBs keep pressuring

Bitcoin has become the first thing I look at when I turn my computer on since Monday, since it tells about the risk appetite for the riskiest and most speculative pockets of the market. Unsurprisingly, its correlation to the USDJPY has been notably close in the two sessions that followed Bank of Japan (BoJ) Governor Ueda’s hint that the BoJ was cooking a rate hike for the December dinner.

This morning, Bitcoin looks better, the Nikkei returned above the 50K psychological level and the US and European futures are pointing at a positive start, yet the pressure from the JGBs remains firm. The Japanese 10-year yield is up to a fresh multi-decade high, at 1.88% this morning. And that’s no good news: these levels mark the end of the Japanese free liquidity that has been propping the markets over the past decade, as Japanese investors put trillions of dollars of funds into US Treasuries among other assets to get better returns. Above the 1.71% level for the 10-year JGB, the math doesn’t work: the Japanese don’t get a better return outside when you take the currency hedging costs into account. The risk is another episode of reversal of carry positions in the yen – where investors will sell their foreign assets and return to yen – strengthening the yen and triggering a fresh selloff across global financial markets, including US Treasuries. And US Treasuries are the basis of valuation. This is what we use as the risk-free rate in pricing models.

So the chain goes like this: the BoJ hikes rates, the Japanese yields climb further, the Japanese sell their foreign assets and repatriate their funds, the yen appreciates and the global financial assets depreciate. This is what happened on Monday. On Tuesday, a strong 10-year JGB auction gave relief, allowing the Bitcoin rebound. Today, we’re back to square one. The Japanese yields are pushing higher again. If we follow the reasoning, we should see cautious risk taking and perhaps a bearish session-

I come to believe that the biggest driver of the year-end sentiment may not be the Federal Reserve (Fed), but rather the BoJ – or the cocktail of Fed/BoJ. The Fed is widely expected to announce a 25bp cut when it meets next week. The ADP report due today is expected to print 5’000 new private-sector job additions in the US during last month – a number small enough to back a December Fed cut. But Friday’s PCE data will hint at what will happen next. And given that it’s still well above the Fed’s inflation target, we could hear a hawkish cut from the Fed next week that could reshuffle expectations for next year – toward a more cautious path of easing. Right now, the expectation is 2–4 rate cuts next year. The risks are skewed toward fewer, not more.

Then, the BoJ will announce its decision a week after the Fed, and markets assign roughly an 80% probability to a hike. Mr. Ueda has made it clear he sees more room to go in terms of policy normalization. I mean, look, the BoJ’s policy rate sits at 0.50% while Japanese inflation sits around 3%. You bet, there is room to normalize. It’s the end of Japanese deflation, and it’s the end of the Japanese free cash – for good.

The S&P 500 was slightly up yesterday. Nvidia made a positive attempt, though most of the session gains were quickly given back, perhaps on fresh news that Amazon also started selling its own AI chip – the Trainium 3 – to compete with Nvidia and Google in the hardware segment. Interestingly, the market reaction to Amazon’s Trainium has been much more muted compared to the reaction to Google’s TPUs. But wait until a big name inks a deal with Amazon for their chips. Amazon – like Google – has the data centers and the chips. They don’t have a Gemini-like chat buddy, but AWS supports training and deployment of AI models on its custom AI chips. So keep an eye on Amazon.

I stumbled on an interesting Bloomberg piece — and separately a Visual Capitalist chart — that are worth commenting. The Bloomberg report highlights how the winning tech business models of companies like Meta and Amazon were historically capital-light, which helped deliver the kind of juicy profit margins investors love. The Visual Capitalist chart then compares profit margins to revenue — and you'd expect higher margins to neatly align with higher revenues, but the relationship is not diagonal. Some companies scale revenue with big margins; others don’t. And unsurprisingly, your capital-light tech buddies sit among the most profitable companies in the world precisely because they’ve been capital-light.

But that’s changing with AI. These companies are now pouring billions into data centres and chips just to stay in the game. And that capex has already been eating into free cash flow — and may soon start nibbling at their margins. Margins won’t collapse overnight, but the direction is clear: AI is rewriting the cost structure of Big Tech, and investors will eventually have to price that in.

And the size of the repricing will ultimately depend on yields — because yields are the backbone of every major valuation model. Lower yields mechanically lift valuations; higher yields compress them. And right now, yields are under upward pressure.

So, all eyes turn to the Fed, and to the inflation data that will dictate how fast — and how deep — the Fed can realistically cut.