Portfolio actions today.

AKO1L confirmed my suspicions about weakening outlook. Basically no profit in Q2. However, CFO today reminded, that QoQ comps can be misleading. They hold quite a lot of inventory in hopes to sell at higher prices in FY H2.

Reduced further by 33%, locking FIFO +61.8% profit at +25.2% CAGR

Increased VLP1L by 50% <€12.2. No expectations on Q4, simply maintaining exposure to OMXB. Focus on dividends, expecting continued management execution turning around MPK, ramping Bauska plant, core business.

And finally, increased $FIGR once again (still a tiny position though). 100%+ revenue growth YoY suits me well. Will be looking for earnings call and diluted share count next week.

Don't follow LLM research and development closely, but today it's first time Im hearing OAI talk aboit prompt caching.

Imo, one of those techniques, that has a step change in efficiency.
#AI

Comparing 2025 vs 2024 $CDNS:
-Revenue +14.1%
-Backlog 7.8B (+14.7%)
-GAAP EPS +5.4%
-OPM -90bps
-Headcount +8.6%
-SBC +16.3%

China revenue share (21'-25'):
13%->15%->17%->12%->13%

That SBC really weighing down on stock.

Marginally reduced $NET at $199.09 last week. Locked 166.3% profit, or 20.8% CAGR based on FIFO. Still holding more than I want to currently. Successfully captured market spread between recent buy and this sale of ~$30.

$ALAB down AH on Q4 92% YoY growth

My guess: weak lower end guidance for QoQ growth of 5.7%. Upper bound of 9.8% - same as 2025 Q2. So they are giving more room. Will listen to call next wk..

Overall, market reaction mirrors that of hyperscalers - caution. Multiple compression

Cloudflare 2025 revenue growth YoY:
Q1: 27%
Q2: 28%
Q3: 31%
Q4: 34%

Last week posed a question. Which SaaS stocks will benefit, which models are threatened by AI. $NET answers.

BOND INVESTORS COULD TURN ON AI COMPANIES

Bond markets may grow impatient with AI firms if their massive investments don’t generate clear returns, warns Michel Saugne, CIO at La Financiere de l’Echiquier.

Investors will expect companies to prove they can monetize AI spending, or risk skepticism as they raise debt to fund growth. Saugne likens it to selling “picks and shovels” in a gold rush—you need proof there’s gold to be found.

They could, but not today.
$GOOG 15B bond issue attracted 100B investor interest.
#AI

Active last week. Outlined actions taken on:
SFG1T, APG1L, $NET, $FIGR, $MU, Robur Corp. Bond Nordic (Fund)
https://www.generalistlab.com/insights/finance/portfolio/portfolio-moves-2026-week-6/

Agree with Gerstner. Years ago you could buy $CRM at 35 FCFs, you knew what you were getting. AI brings fog around the real discount rate and you get a correction in SaaS. Market wants to see winners (acceleration) and losers (deterioration).

One thing is clear though .. every vibe or not coded app is going to need an infra. $DOCN is SaaS too, where was it's dip this week? Non deterministic behavior - backups, snapshots - all infra.

My hurdle for artificial general intelligence:
-sustained profitability in market
-fix plumbing
#AI

Made some movements in portfolio this week. Hopefully will find time on Sunday to write a short blogpost.

24% AWS growth in Q4 - fastest in past 13 quarters and staggering 200B capex guide for 2026, toping yesterday's 185B $GOOG guide.
$AMZN