BNP Paribas Raises Google Target Price to $450, Betting on TPU External Sales Potential
nashnova research
BNP Paribas lifted its Alphabet price target to $450, betting that Google's in-house AI chip — the TPU — can generate $80–100 billion in cumulative external revenue by 2028. The catch: near-term margin drag and a supply chain that still has to deliver.
Why is BNP raising the target now?
The new $450 target implies roughly 28% upside from Alphabet's Tuesday close of $351.40.
The thesis is not about search or advertising. It is about the TPU — Google's tensor processing unit, a chip purpose-built to run AI models — moving from internal use to external sales.
This means → BNP sees the TPU not as back-end infrastructure but as a standalone, priceable chip business.
Where does the TPU revenue come from?
BNP estimates TPU external sales could generate $80–100 billion in cumulative revenue through 2028.
Disclosed customers include AI firm Anthropic and asset manager Blackstone.
Alphabet has locked in over 5 gigawatts of identifiable TPU capacity commitments; most deliveries are expected to begin in 2027.
In plain terms = the orders are signed and the capacity is reserved — the question is whether it ships on time.
What does the revenue ramp look like?
BNP's year-by-year forecast: 2026 >$4 billion → 2027 >$40 billion → 2028 ~$30 billion.
2027 is the breakout year — a tenfold jump from $4 billion to $40 billion.
This means → if large-scale 2027 deliveries slip, the entire revenue curve has to be redrawn.
How did the valuation method change?
Old method: 30× P/E on 2027 EPS estimate of $13.85.
New method: 26.5× P/E on 2028 EPS estimate of $17.
In plain terms = the multiple came down but the earnings forecast went up — so the target actually rose. BNP is betting TPU sales can grow the profit pool.
How much margin drag should investors expect?
TPU external sales will increase operating profit in absolute dollar terms but will weigh on Google Cloud's operating margin.
Estimated dilution: roughly 115 basis points in 2027 and over 50 basis points in 2028.
This reflects a heavy-upfront-investment business — scale first, margin recovery later.
What is the biggest uncertainty?
Supply-chain constraints could slow TPU delivery, and disclosure on the TPU business's unit economics remains limited.
Whether large-scale 2027 deliveries happen on schedule is the single most important test of this valuation thesis.
This means → BNP has drawn an attractive revenue curve, but there is only one verification window — 2027 capacity delivery.
市场有风险，内容仅供研究参考，不构成投资建议。