Who is jukan05? The record behind the chip leaker

If you searched this handle and found nothing familiar, start here: @jukan05 is not a new account. It is the renamed home of Jukan, the semiconductor account that built just over 200,000 followers as @Jukanlosreve by translating Korean chip-industry reporting, broker notes, and supply-chain data into English threads, often hours before the same information reached Western outlets. The account dates to September 2024, and its bio today is Citrini-first: "@citrini | Not Investment Advice | DYODD". The rename is verifiable in public records: X redirects the old status URLs to the current handle, and Citrini Research lists @jukan05 as the X account of its semiconductor analyst. We track the account and grade its stock calls, so this page covers both who Jukan is and what the graded record shows.

Who they are

Jukan is pseudonymous, publishing under a pen name, and this profile stays within what the account states publicly. The output speaks for itself: a high-volume feed of translated Korean media exclusives, memory-market data (DRAM and NAND pricing, HBM roadmaps, Korean export statistics), and summarized sell-side research, plus the consumer-tech leaks that made the account famous. MacRumors keeps a source page on them as a tech news aggregator, and Tom's Guide has called the account a reliable tipster. The display name cycles through conference tags, GTC, COMPUTEX, ICML, and the newsletter's about page claims sources inside the semiconductor industry, including Samsung and SK hynix.

The account's biggest single moment came in January 2025, when its thread on a Korean brokerage's DeepSeek analysis became the basis for Tom's Hardware's widely mirrored story on DeepSeek bypassing CUDA for PTX, one of the most consequential Nvidia stories of that year. Since early 2026, Jukan has also written the Semis Memos series at Citrini Research and publishes a Substack, SemiconSam, on top of the feed.

What they post, in our data

Quantral has tracked the account since April 6, 2026: 838 stock posts in about four and a half months, nearly all of them in and around the semiconductor complex. The most-posted names, with each one's bullish-to-bearish split among the posts that carry a direction:

Note what is missing from that table: conviction. A third or more of the account's posts on its top names carry no direction at all, which fits what the feed is. Jukan relays and translates; the market decides what it means. AAPL is the tell, a near-even 10 to 11 split from an account that mostly moves information rather than takes sides.

The graded record

Relaying information is still making calls, and the calls grade out well. On Quantral's six-month accuracy board, @jukan05 ranked third of every account we track: 70 graded calls, 58.6% right, with a best graded call of DELL at +43% in seven days. On the Q2 board they placed ninth with 70 graded calls in the quarter.

Third place undersells it. The number-one account on that same board, @citrini at 64.2% across 81 graded calls, is Citrini Research, the shop Jukan has written the Semis Memos for since January 2026. Count it that way and one pseudonymous analyst has a hand in two of the top three accounts we track: the research house they write for at number one, and their own feed at number three.

The comparison worth making is with the thesis-writers above them on the board. The number-two account, @aleabitoreddit, posts concentrated, high-conviction supply-chain theses and graded out at 59.5%. Jukan posts translated scoops and data, hedged and sourced, and graded out at 58.6%. Two opposite styles, one percentage point apart, and both wrong roughly four times in ten. That last number is the honest context for every follower count in this story: the best graded hit rate on our board is about 64%, so nobody in the set, aggregator or oracle, escapes the error bar.

The August controversy

A profile written this week cannot skip it. In early August 2026, posts from the account's past online life surfaced on X. Jukan confirmed some were theirs and published an apology, citing inappropriate remarks, unauthorized use of others' information, and insufficient source attribution, and acknowledging that account growth had sometimes been prioritized over verification. They pledged clearer sourcing going forward. The handle change to @jukan05 appears in public records around the same period, which is likely why you searched this handle in the first place.

We are not the judge of any of that. We grade stock calls, and the grading is indifferent to reputation in both directions: the record stood at third on our board before August, and every call since gets graded the same way. If the sourcing pledge changes the quality of the feed, the number will show it.

Follow the calls, graded

The useful question about an aggregator account is not whether the scoops are exciting but whether acting on them would have been right. That is an answerable question. The live version of @jukan05's record, updated as calls resolve, is on their Quantral voice page, next to every other account we track, including the thesis-writers they share the board with. Check the record before you check the headline.

---

Graded-call figures are Quantral's own methodology (subject calls graded over fixed windows; see how a track record is graded) as published on the boards linked above. Post counts and splits cover the account's tracked stock posts from April 6 to August 16, 2026. Claims attributed to the account or to third-party coverage are theirs, not ours, and are linked to their sources. Quantral surfaces signals and context from public sources to support your own research. Nothing here is financial advice or a recommendation to buy, sell, or follow anyone. Past performance is not indicative of future results.