Warren Buffett is sitting on nearly $400B in cash, and that’s hard to ignore.
The last time Berkshire built a cash pile this large was around the 2007–2008 period. It doesn’t guarantee a crash, but it does suggest Buffett sees more value in waiting than chasing current prices.
I’ve been watching this Fed setup closely, and the labor market + inflation combination now looks harder to ignore.
August payrolls came in at 162K, far above the roughly 56K expected, while unemployment stayed at 4.1%. That already removed some of the argument for the Fed to stay cautious.
Now CPI has added another layer.
Headline inflation came in at 3.4% YoY, with prices rising 0.4% MoM. Core CPI also rose 0.3% MoM, keeping underlying inflation pressure alive. Markets are now pricing a very high probability of a September rate hike.
My read is that this is not automatically “bearish everything.”
Higher rates are a headwind for risk assets, but the bigger signal is that the Fed may have less room to ignore persistent inflation while the labor market is still holding up.
That makes me more cautious on high-beta equities, while gold remains the asset I’m watching more closely as inflation, geopolitical risk and rate expectations pull in different directions.
For me, the next move is less about chasing one CPI candle and more about how markets reprice the September 16 Fed decision.
$BTC is starting to shift its 1H structure back toward an uptrend. 🔥
The key move wasn’t just the bounce from the $76K area — BTC reclaimed the local resistance around $77.4K–$77.6K with a strong expansion candle and noticeable volume.
Now the important part is holding that breakout instead of slipping straight back below it.
If $77.5K starts acting as support, the structure opens room toward $79K and then the previous 1H supply above.
For me, this is the first meaningful sign that buyers are taking control of the short-term structure again.
I was reading through Dusk’s cryptography stack when one less glamorous page caught my attention: the hardware guide for prover nodes.
That page changed the question for me.
Dusk can hide transaction details and still prove that a transfer is valid. But those proofs do not appear for free. Someone has to perform the computation.
Dusk’s operator docs describe proof generation as a compute-heavy job. Each prover worker needs its own CPU core, and strong single-core performance matters because individual proofs are generated in a single-threaded process.
This sounds like an infrastructure detail until you imagine actual financial traffic.
A tokenized fund may process subscriptions during the day, then calculate redemptions and ownership updates together. A bond issuer may distribute interest to thousands of holders at once. A market stress event could trigger many private transfers or compliance checks within the same short window.
That traffic would not arrive evenly.
It would arrive in bursts.
So the privacy question is not only whether Dusk can generate a valid zero-knowledge proof. It is whether the prover network can absorb a sudden concentration of proof requests without turning confidentiality into a queue.
Hedger says its lightweight circuits can generate client-side proofs in under two seconds. That is encouraging, but one clean proof and a full market workflow are different tests.
I would want to see how proof time changes when many users submit private transactions together, and whether applications can add prover capacity without making the experience unpredictable.
Public chains usually measure scale through transactions per second.
For Dusk, another number may matter just as much: proofs completed per second during peak demand.
The cryptography can be correct while the user experience still becomes slow.
That is the part of Dusk’s privacy stack I would benchmark before the institutional traffic arrives.
I went into the DuskEVM bridge guide looking for withdrawal timing.
The fee line was what made me stop.
Moving DUSK into the current DuskEVM testnet flow is fairly simple: submit the deposit on Dusk L1, then wait for the balance to appear on DuskEVM.
Coming back is different.
A withdrawal requires three separate actions:
Initiate it on DuskEVM. Prove it on Dusk L1. Finalize it on Dusk L1.
The wallet can move through “Waiting for output proposal,” “Ready to prove,” “Proof submitted” and “Waiting to finalize” before the DUSK is released.
That part makes sense. The return journey is where DuskDS verifies the state coming from the EVM layer.
But then I noticed the user also needs enough unshielded DUSK on the L1 to pay for both the proof and finalization transactions.
So someone can hold DUSK inside DuskEVM, start moving it back, and still need a separate usable L1 balance to complete the exit.
The funds are not necessarily stuck. The user may simply be missing the asset required to finish proving that they are allowed to retrieve them.
That feels like a small technical requirement until DuskEVM is used by someone who has never touched Dusk L1 directly.
A Solidity developer may understand contracts and MetaMask perfectly, while the end user only sees a withdrawal sitting at “Waiting to finalize.”
Dusk’s docs are careful here. They tell users to follow the wallet status rather than estimate readiness from elapsed time alone.
Now I am watching how much of this survives into the final production experience.
Will users still manually manage three transactions and two L1 fees, or will the wallet abstract the process without hiding which settlement stage their money has actually reached?
The bridge is not only moving DUSK between layers. It is where DuskEVM’s familiar front end meets Dusk-native settlement and where that difference becomes impossible to ignore.