TSMC's record Q2 earnings reveal the real AI bottleneck isn't wafer output — it's CoWoS advanced packaging, with lead times stretching 52–78 weeks while DRAM supply is sold out through 2027. Plus: the UAE gets license-free Nvidia access as the US reshapes its semiconductor alliance strategy.
Audio is available on Spreaker — see link below.
TSMC's chief executive just said out loud what the supply chain has been showing for months: the bottleneck in AI hardware isn't wafer production anymore. It's packaging.
Here's what that shift means in practice. The conversation in this industry has been dominated by node progression, two-nanometer yields, three-nanometer capacity, how fast TSMC can ramp N2.
The packaging constraint doesn't sit alone. Memory is now equally tight.
On the investment side, TSMC added another hundred billion dollars to its Arizona commitment, bringing the total to two-hundred-sixty-five billion dollars. Capital expenditure guidance for twenty-twenty-six was raised to between sixty and sixty-four billion dollars.
Away from TSMC's earnings, the US Commerce Department elevated the UAE to A:5 classification, granting license-free access to advanced Nvidia chips. The immediate beneficiary is the Stargate UAE data center project, a thirty-billion-dollar joint venture targeting five gigawatts of AI compute capacity.
The near-term signals to watch are specific. First, whether TSMC discloses targeted CoWoS capex in its next investor communications.
Chapter summary auto-generated from the verified script. Listen to the full episode for the complete content.