In a stark sign of the intense competition for critical AI components, Google has reportedly dismissed a senior procurement executive. The dismissal is linked to a failure to secure sufficient High Bandwidth Memory (HBM) supply for Google's Tensor Processing Unit (TPU) accelerators during a severe global shortage, putting a key AI hardware initiative at risk.

Google sources approximately 60% of its HBM for TPUs from Samsung. However, as the AI boom escalated, both SK Hynix and Micron reportedly refused Google's requests for additional orders, leaving the company exposed. The executive is said to have failed to anticipate the market crunch and sign the necessary long-term supply agreements (LTAs) in advance.

The dismissal is a symptom of a broader, unprecedented shortage of HBM and related memory like LPDDR. Microsoft, Google, and Meta executives have reportedly made personal trips to South Korea to negotiate directly with Samsung and SK Hynix. In one telling incident, a Microsoft executive is said to have angrily left a meeting at SK Hynix headquarters after being denied a supply request.

With HBM production lines at full capacity, the market has shifted decisively to a seller's market. Some hyperscale cloud providers, in desperation, have reportedly offered "blank checks," agreeing to pay any price to secure available HBM stock. This panic buying is further inflating prices and worsening supply constraints. The shortage has even impacted consumer electronics, with Apple reportedly paying a 230% premium for LPDDR5X memory, affecting the cost structure of products like the upcoming iPhone 17.

ICgoodFind's Insight
Google's firing of a procurement head over HBM shortages highlights the extreme pressure and strategic importance of securing semiconductor supply chains in the AI era. It underscores a shift where procurement competency is as critical as technological innovation for maintaining competitive advantage.