On Monday, Nov. 10, 2025, Wall Street chatter is still centered on Jefferies’ high‑profile call: Broadcom (NASDAQ: AVGO) replaces Nvidia (NASDAQ: NVDA) as the brokerage’s top pick in semiconductors, underpinned by a surge in demand for custom AI chips. The move, first detailed last week, comes alongside a fresh five‑year scenario analysis for Broadcom making the rounds in investor media—and a separate downgrade of Pinterest in the same analyst roundup. [1]
Key takeaways
- Broadcom named Jefferies’ new top pick; price target to $480 (from $415). Analysts see “outsized upside” as custom AI silicon hits an inflection point at hyperscalers like Google and Meta. Nvidia remains rated Buy, but it’s no longer the firm’s top franchise pick. [2]
- Why the pivot? Jefferies highlights accelerating orders for application‑specific integrated circuits (ASICs)—including checks that point to sharply rising TPU volumes and major next‑gen AI deployments through 2026–2028. [3]
- 5‑year view: A widely shared analysis posits how AVGO could compound if it captures a meaningful slice of a multi‑trillion‑dollar data‑center buildout through the decade. It’s a scenario—not guidance—but it frames the bullish debate. [4]
- Elsewhere in the roundup: Pinterest (NYSE: PINS) was cut to Neutral by Monness, Crespi, Hardt after soft Q4 commentary and intensifying competition. [5]
Why Jefferies flipped its top pick from Nvidia to Broadcom
Jefferies’ semiconductor team, led by Blayne Curtis, argues that custom AI chips are moving into a “step‑function” growth phase, and that Broadcom’s ASIC franchise stands to benefit most. The note raised AVGO’s price target to $480 (from $415), added the stock to Jefferies’ Franchise Picks, and explicitly removed Nvidia from that list—while keeping a Buy on NVDA and lifting its target to $240. [6]
The call is grounded in hyperscaler roadmaps. Jefferies’ checks point to Google’s TPU shipments nearing ~3 million units in 2026, supported by a new ~$10 billion order from Anthropic (~250,000 units) and potential follow‑ons. The firm models ~$10B of AI revenue for Broadcom in calendar 2027, with a pathway that “could easily scale to $40–$50B per year in C28+” as deployments broaden. The analysts also flagged Meta’s first HBM‑equipped AI chip in 3Q26 and an OpenAI ASIC later in 2026 as additional catalysts. [7]
On the numbers, Jefferies lifted AVGO’s revenue estimates to $100B (2026) and $130B (2027) and bumped EPS forecasts to $10.31 and $13.88, respectively. In an upside case tied to hyperscaler capacity expansions, the team sketches EPS of ~$20 in 2027 and ~$30 in 2028 if OpenAI’s build scales 2–3 GW annually. [8]
Bottom line: Jefferies’ view isn’t that Nvidia stops winning—far from it—but that Broadcom has more room for estimate revisions upward over the next 12–24 months as custom silicon proliferates. That’s why the baton for “top pick” is moving. MarketWatch and Barron’s echoed the same rationale last week, underscoring a market narrative that AI leadership is broadening beyond GPUs alone. [9]
The 5‑year AVGO scenario investors are circulating
A weekend article (syndicated from The Motley Fool) models what Broadcom’s stock could look like in five years if the AI data‑center cycle matches even conservative estimates. The piece notes that industry leaders see $3T–$4T (or more) of data‑center spending by 2030, with a sizable share flowing into AI‑capable silicon. In one back‑of‑the‑envelope scenario—if Broadcom took ~20% of AI‑chip revenue by 2030—the author estimates ~$234B in annual sales and outlines how the share price could compound assuming valuation multiples stay intact. It’s a thought experiment, not a forecast, but it captures how big the runway might be if custom accelerators continue to gain share. [10]
Context matters: the article also points out Broadcom’s outsized run since early 2023 and flags valuation sensitivities (P/E and PEG), reminding readers that execution and competitive dynamics—especially with Nvidia—will determine whether that scenario is achievable. [11]
Where this leaves Nvidia right now
Jefferies didn’t turn bearish on Nvidia; it raised NVDA’s target to $240 and kept a Buy rating. The thesis is that AI infrastructure is not a zero‑sum game: GPUs remain central to training, while custom ASICs increasingly address cost, energy, and latency for scaled inference and specialized workloads. That leaves room for both companies to post strong results even as leadership diversifies across the stack. [12]
The other headline in the same analyst roundup: Pinterest cut to Neutral
Within the same “5 big analyst AI moves” round‑up, Monness, Crespi, Hardt downgraded Pinterest to Neutral, citing a tougher Q4 outlook, fierce competition, and macro headwinds. Separate notes and filings around the downgrade tracked a sharp move in the shares, reflecting the Street’s caution on ad‑market execution amid rising AI‑driven competition across social platforms. [13]
What to watch next
- Customer disclosures & PO cadence: Any confirmations from Google, Meta, or OpenAI about custom silicon roadmaps or multi‑year purchase commitments would be key signposts for the ASIC ramp Jefferies models. [14]
- AVGO guidance updates: Watch for revenue/EPS bridges into FY26–FY27 that reconcile networking, ASICs, and software contributions. [15]
- Nvidia product cycles: NVDA’s next‑gen platforms (and software attach) still anchor the training market. Upside to inference could tighten the “ASIC vs. GPU” narrative into a “both/and” dynamic. [16]
Quick FAQ
Is Jefferies bearish on Nvidia now?
No. Jefferies still rates NVDA Buy with a higher target; it simply sees more upside revision potential at AVGO over the next year as custom chips scale. [17]
How confident is the 5‑year AVGO price call?
It’s scenario analysis, not company guidance. It frames the possibility set if AI data‑center spend and Broadcom’s market share evolve favorably. Treat it as context, not a target. [18]
Disclosure: This article is for informational purposes only and is not investment advice. Do your own research.
Sources: MarketWatch coverage of Jefferies’ call; Barron’s recap of the top‑pick change and targets; Reuters context on Broadcom’s AI pipeline and CEO tenure; Jefferies note details and modeling via StreetInsider; and the five‑year AVGO scenario from The Motley Fool (syndicated on Nasdaq). [19]
References
1. www.marketwatch.com, 2. www.streetinsider.com, 3. www.streetinsider.com, 4. www.nasdaq.com, 5. www.investing.com, 6. www.streetinsider.com, 7. www.streetinsider.com, 8. www.streetinsider.com, 9. www.marketwatch.com, 10. www.nasdaq.com, 11. www.nasdaq.com, 12. www.streetinsider.com, 13. www.investing.com, 14. www.streetinsider.com, 15. www.streetinsider.com, 16. www.barrons.com, 17. www.streetinsider.com, 18. www.nasdaq.com, 19. www.marketwatch.com
