SEMI

Columbia Select Technology ETF

Columbia Threadneedle · $44.1m in assets, October 8, 2026

SEMI, the Columbia Select Technology ETF, is a $44 million ETF whose main story is the AI compute core. Over the past year it returned +34.6%, against +16.8% for broad US ETFs. Figures to October 2, 2026; past performance, not a recommendation.

SEMI · the ETFFRI OCT 2Market price total return
+19.0% a yearTotal return since launch, March 2022
+34.6%Last 12 months
+33.4%Year to date
31.7%Volatility since launch

The narrative it carries

This is its only narrative.

The AI compute core →

The chips, memory and data centers AI runs on, and the big technology companies that build and buy them.

The AI compute core · the narrative this week
57.0WARM level, 4 of 22
last four weeks
396ETFs carry it
68.5%held by the top five
VGTlargest holder, 23.8%
$682 bncapital, 5th of 22

The band is where this narrative’s level sits among the 22 this week. QUIET means less talked about, not falling.

Who owns the capital
VGT 23.8%XLK 20.5%SMH 12.1%SOXX 7.8%IYW 4.3%391 other ETFs 31.5%68.5%top five
#ETFSHAREASSETS1 YEARCOST
1VGT23.8%$147.56bn+36.0%0.09%
2XLK20.5%$127.30bn+40.4%0.08%
3SMH12.1%$74.88bn+87.0%0.35%
4SOXX7.8%$48.43bn+109.7%0.33%
5IYW4.3%$26.53bn+36.0%0.37%
391 other ETFs31.5%
SEMI: 0.0%

Share is each ETF’s slice of the capital carrying this narrative. 1 year is market price total return; cost is the expense ratio. See all 396 ETFs on the narrative page →

FRI OCT 2MARKET PRICE TOTAL RETURN

1M3MYTD1Y3Y*5Y*
SEMI+8.3%+7.7%+33.4%+34.6%+30.9%–
The AI compute coreits narrative+9.5%+8.7%+47.8%+50.2%+41.6%+24.5%
The marketbroad US ETFs+0.8%+2.6%+14.6%+16.8%+22.9%+12.9%

1M, 3M, 1Y, 3Y, 5Y: one month to five years. YTD: year to date. * A year, annualised. A dash means the ETF has not traded for the whole period. Since it started trading on March 30, 2022: +19.0% a year.

$10,000 invested March 31, 2022

SEMI $22,150The AI compute core $29,299The market $17,865
$0$10,000$20,000$30,000Mar 2022Mar 2023Mar 2024Mar 2025Mar 2026$0$10,000$20,000$30,000Mar 2022Mar 2024Mar 2026

Year by year

SEMIThe AI compute coreThe market
0%20%40%60%SEMI 2023: +45.4%Narrative 2023: +57.1%The market 2023: +27.9%2023SEMI 2024: +15.9%Narrative 2024: +27.5%The market 2024: +23.1%2024SEMI 2025: +24.8%Narrative 2025: +29.0%The market 2025: +16.8%2025SEMI YTD: +33.4%Narrative YTD: +47.8%The market YTD: +14.6%YTD0%20%40%60%SEMI 2023: +45.4%Narrative 2023: +57.1%The market 2023: +27.9%2023SEMI 2024: +15.9%Narrative 2024: +27.5%The market 2024: +23.1%2024SEMI 2025: +24.8%Narrative 2025: +29.0%The market 2025: +16.8%2025SEMI YTD: +33.4%Narrative YTD: +47.8%The market YTD: +14.6%YTD

Total return in each calendar year; this year is year to date. A year appears only if SEMI traded for all of it.

Total return on the market price, with each distribution reinvested on its ex-date. Not the NAV return the issuer publishes; the two differ when the ETF trades above or below the value of its holdings. The narrative is the asset-weighted return of the unlevered ETFs that carry it. The market is the asset-weighted return of broad US index ETFs. Past performance is no guarantee of future results.

Common questions

What story is SEMI selling?

SEMI’s main story is the AI compute core, which covers the chips, memory and data centers AI runs on, and the big technology companies that build and buy them.

Which ETFs compete with SEMI?

The ETFs whose weekly returns move most like SEMI’s over the past 3 years are JHAI (0.97), SHOC (0.95) and TEK (0.95).

How has SEMI performed against its story?

Over the past year SEMI returned +34.6%, the ETFs carrying the AI compute core +50.2% and broad US ETFs +16.8%, to October 2, 2026. Its weekly returns have a correlation of 0.94 with the story. Past performance only.

Run an ETF in the AI compute core? StoryVector shows issuers where their ETFs stand against the closest rivals, every week.

Request a pilot
Narrative figures are from the weekly board. Performance is market price total return as of October 2, 2026. Assets are StoryVector’s weekly figure. Classification and measurement. Not investment advice. How the measurement works →