The market has it backwards: Meta is the only hyperscaler overspending...

Microsoft, Google, and Amazon sold off after announcing large capex increases, while Meta rose ~10%. Our analysis suggests the market has this backwards. Please see the link for our take, summarised on a single slide.

Hyperscalers to nearly double AI capex again in 2026. Implications

All four hyperscalers have now reported. AI capex is set to nearly double for a second consecutive year. Click the link below for a single slide on the implications for XPU spending.

Jabil Inc: 1 director

A director at Jabil Inc sold 15,000 shares at 246.907USD and the significance rating of the trade was 68/100. Is that information sufficient for you to make an investment decision? This report gives details of those trades and adds context and analysis to them such that you can judge whether these trading decisions are ones worth following. Included in the report is a detailed share price chart which plots discretionary trades by all the company's directors over the last two years clearly show...

Reconciling TSMC guide with total AI Infra. Spending in 2030, on a sin...

TSMC raised its AI revenue growth outlook last week to mid-to-high 50% CAGR from 2024 to 2029. Click on the link for a reconciliation with our forecast for total AI Infrastructure Spending.

OpenAI deploying 750MW of Cerebras wafers. Our quick take.

OpenAI signed a multi-year deal to deploy 750 MW of Cerebras wafer-scale compute, ramping from 2026 through 2028.

Moody's Ratings assigns Baa3 rating to Jabil Inc.'s proposed senior un...

Moody's Ratings (Moody's) assigned Baa3 ratings to Jabil Inc.'s (Jabil) $900-$1,000 million proposed senior unsecured notes, which will be comprised of 2029 and 2033 maturities. The net proceeds from the proposed offerings will be used for general corporate purposes including repaying $500 million o...

Technology infrastructure in 2026: Top picks & updated perspective - C...

AI infrastructure in 2026: Updated perspective & top picks

As usual, to start the year, we have pulled together our favorite ideas in our coverage for 2026. We have selected names with our highest conviction.

The Tech Infrastructure Quarterly Bible 3Q25

We publish today our comprehensive quarterly bible: 242 pages of detailed analysis on what happened in the last 3 months, and how we interpret it, in light of our current convictions. The first section acts as a PM summary, outlining our key findings, and latest thoughts on the semi cycle, in 6 slides:

Gen AI Big Idea Series – The Broadcom Platform – ​Hock Tan, CEO

We hosted another session of the 5th annual BCG × NSR Gen-AI Big Idea Series this week. We will host a slate of exceptional speakers through Feb-26 (full line-up here). This session was a conversation with Hock Tan, CEO of Broadcom, along with Kirsten Spears, CFO, and Charlie Kawwas, President of the Semiconductor Solutions Group. Please see the link below for our takes.

Broadcom 4FQ25: AI momentum confirmed; 2027 visibility gets into expec...

Broadcom reported 3FQ25 yesterday.

Oracle 2FQ26: our quick take.

ODDO BHF SECURITIES Small & MIDCAP MORNING NEWS – 12/05/2025

ODDO BHF SECURITIES Small & MIDCAP MORNING NEWS – 05/12/2025

ODDO : More Rack ‘n’ Roll ahead from 2026 onwards

We initiate coverage on two Swiss Industrials midcaps, Kardex (Outperform, target price CHF 330) and Interroll (Neutral, CHF 2,375). Both niche players should continue to benefit from steadily growing automation needs (low ~15-20% penetration) in the increasingly global intralogistics market where client activities indicate a gradual revival in growth of the e-commerce / project business (attractive payback times of only 12-36 months). Both high-quality names should see ongoing high growth inves...

OpenAI Losing the Lead to Google? Implications for Microsoft

The release of Gemini 3 has triggered a wave of strong reception across the industry. Please click the link for our thoughts on what this means for Microsoft.

TPU v7 vs. Blackwell & Rubin: what we know so far about efficiency.

We published earlier this year an analysis comparing efficiency of GPUs and TPUs. Today, we extend the analysis to TPU v7 (Trillium) and Nvidia Rubin.