Kim Kahn

Good morning.

Today is Wednesday, December 10th, and I'm your host, Kim Kahn, filling in for Julie Morgan.

Bitcoin has wiped out its 2025 gains, and one of Wall Street's biggest bulls is cutting forecasts across the board.

Standard Charter's Jeff Kendrick now sees Bitcoin at $100,000 at the end of this year, down from $200,000, $150,000 at the end of 2026, down from $300,000, and $225,000 in 2027, down from $400,000.

The bank also pushed out its long-term target of $500,000 to 2030, two years later than previously expected.

Recent price action has been challenging, to say the least, said Kendrick, who heads digital assets research for the bank.

Kendrick said Bitcoin's run since 2024 was powered by ETF inflows and digital asset treasury companies buying aggressively, but many of those DATs now trade below the value of the crypto they hold, making further accumulation harder to justify.

He expects consolidation rather than capitulation, but also says debt buying is largely done.

On the other side of the ledger, Bernstein said Bitcoin is now in an elongated bull cycle, with stickier institutional demand helping to absorb retail selling pressure.

Despite a 30% correction, ETFs have seen outflows of less than 5%.

Bernstein now targets 150,000 for 2026, with a potential cycle peak of 200,000 in 2027.

Its long-term 2033 target of roughly $1 million remains intact.

How about LoJack for your GPU?

Nvidia has developed a location verification feature that can indicate which country its chips are operating in, a move designed to stop high-end AI GPUs from being smuggled into restricted markets.

Reuters reports that NVIDIA has privately demoed the technology, though it hasn't been released yet.

The tool would be an optional software agent customers can install, tapping into the confidential computing capabilities already built into NVIDIA's GPUs.

Originally created so data center operators could track performance and manage large fleets of processors, the system uses communication latency with NVIDIA-run servers to estimate where a chip is physically located.

similar to how other internet-based geolocation services work.

NVIDIA said in a statement, "...we're implementing a new software service that empowers data center operators to monitor the health and inventory of their entire AI GPU fleet, leveraging GPU telemetry to monitor fleet health, integrity, and inventory."

And silver's bull run isn't losing its shine.