Market research

# 9 Best Stock Twitter (X) Accounts to Follow in 2026, Ranked by Track Record

Nine fintwit accounts with the strongest scored records, their best and worst calls, and who actually beat the semiconductor ETF.

**By Buzzberg Research**November 12, 2026

![A researcher studies an early investment idea as a chain of AI chips and optical components leads toward data centers.](https://buzzberg.ai/static/blog/ai-infrastructure-winners-2026/cover-1536.webp?v=20261112-2)BUZZBERG RESEARCHbuzzberg.ai

In this research

[01How the nine accounts did](https://buzzberg.ai/blog/best-stock-twitter-accounts-2026#how-the-nine-accounts-did)[02The nine accounts](https://buzzberg.ai/blog/best-stock-twitter-accounts-2026#1-jukan-jukan05)[03Where the returns came from](https://buzzberg.ai/blog/best-stock-twitter-accounts-2026#where-the-returns-came-from)[04How to use this list](https://buzzberg.ai/blog/best-stock-twitter-accounts-2026#how-to-use-this-list)[05How we measured it](https://buzzberg.ai/blog/best-stock-twitter-accounts-2026#how-we-measured-it)

In this research[01How the nine accounts did](https://buzzberg.ai/blog/best-stock-twitter-accounts-2026#how-the-nine-accounts-did)[02The nine accounts](https://buzzberg.ai/blog/best-stock-twitter-accounts-2026#1-jukan-jukan05)[03Where the returns came from](https://buzzberg.ai/blog/best-stock-twitter-accounts-2026#where-the-returns-came-from)[04How to use this list](https://buzzberg.ai/blog/best-stock-twitter-accounts-2026#how-to-use-this-list)[05How we measured it](https://buzzberg.ai/blog/best-stock-twitter-accounts-2026#how-we-measured-it)

Most lists of the best stock Twitter accounts are built on follower counts and reputation. This one is built on records. Buzzberg scores the public calls of the investors it tracks on X and measures what happened after each one. We started from Buzzberg's [Authors leaderboard](https://buzzberg.ai/speakers), kept only X accounts with at least 30 scored calls, and selected those that ranked in the top 20 at **both** the 90-day and 180-day horizons on November 11, 2026. Nine fintwit accounts passed both tests.

The leaderboard ranks authors by how far their calls beat the S&P 500, so all nine beat it on average almost by definition. In 2026 that was also the easier test. Many of their best calls were memory, optical-networking and chip stocks during a year when [SMH](https://www.vaneck.com/us/en/investments/semiconductor-etf-smh/holdings), the VanEck semiconductor ETF, soared too. So we asked a harder question about each account, the same one we put to [AI stock picks against SPY, QQQ and SMH](https://buzzberg.ai/blog/qqq-vs-spy-ai-stock-picks-2026): would a reader have done better simply buying SMH on the same days?

Only one account beat SMH at both horizons. Across the nine, the calls on the AI hardware supply chain did beat SMH on average, but most individual ones did not: memory stocks and a handful of optical winners carried the average. Outside that chain, the calls trailed the S&P 500.

## How the nine accounts did

Imagine reading each post, buying the stock at the close of the next trading day and selling about 90 days after the post, or after the post that ended the call if that came sooner. Across **877 such calls** from these nine accounts, the average return was **+13.4%**. SPY, an S&P 500 ETF bought and sold on the same days, returned **+4.5%**. SMH returned **+18.2%**.

Average return of a reader buying at the close of the next trading day after each call, compared with SPY and SMH on the same dates

| Account | 90-day calls | 90-day return | SPY (90d) | SMH (90d) | Win rate (90d) | 180-day return | SMH (180d) |
| --- | --- | --- | --- | --- | --- | --- | --- |
| 1. Jukan @jukan05 | 44 | +24.1% | +4.9% | +21.9% | 72.7% | +47.7% | +35.5% |
| 2. Bull of Britain @BULLOFBRITAIN | 93 | +20.6% | +6.2% | +23.5% | 52.7% | +43.6% | +36.7% |
| 3. BryzonX @BryzonX | 49 | +14.5% | +5.5% | +23.8% | 51.0% | +34.6% | +38.6% |
| 4. Serenity @aleabitoreddit | 234 | +11.0% | +4.1% | +18.8% | 50.9% | +17.4% | +30.7% |
| 5. Caesar Capital @CaesarCapitalz | 61 | +14.2% | +6.9% | +29.8% | 52.5% | +18.2% | +32.5% |
| 6. TheValueist @TheValueist | 131 | +16.5% | +4.4% | +14.6% | 58.8% | +11.8% | +23.6% |
| 7. Temple 8 Research @Temple_Eight | 128 | +7.1% | +3.3% | +8.6% | 54.7% | +14.9% | +24.6% |
| 8. Pep Invest @Pep_Invest | 81 | +10.5% | +3.5% | +18.7% | 39.5% | +22.4% | +39.9% |
| 9. KaizenInvestor @Kaizen_Investor | 56 | +12.0% | +3.3% | +16.3% | 39.3% | +11.9% | +40.9% |

Accounts are ordered by their combined position on the two leaderboards. Each row compares the account with SPY and SMH on that account's own dates, so the benchmark figures differ between rows. The 180-day columns include only calls old enough to have reached six months. These are averages of separate calls, not the return of a portfolio.

- **Beat SMH at both horizons:** Jukan.
- **Beat SMH at 180 days, not at 90:** Bull of Britain.
- **About level with SMH at 90 days (within two points), behind at 180:** TheValueist and Temple 8 Research.
- **Behind SMH at both horizons:** BryzonX, Serenity, Caesar Capital, Pep Invest and KaizenInvestor. Buying SMH on the same days would have earned more than following their calls.

## 1. Jukan (@jukan05): Asia's memory and chip supply chain

[Jukan](https://buzzberg.ai/speakers/jukan05) posts about the memory and semiconductor supply chain in Korea, Taiwan and Japan, often relaying an industry report with a short note on what it means for a stock. About four in five of the account's scored calls are on the AI hardware supply chain, mostly memory, chip and optical stocks.

On [January 15](https://x.com/jukan05/status/2011620175536398356), Jukan wrote that NAND makers were "basically going to be printing money" over the next three years, tagging [Sandisk](https://buzzberg.ai/trade-ideas/ticker/SNDK) and [Micron](https://buzzberg.ai/trade-ideas/ticker/MU). A reader who bought Sandisk at the next day's close was up **117.9%** after 90 days. SMH gained 14.3% over the same days.

The miss came on [April 15](https://x.com/jukan05/status/2044208909976776936). Reacting to a report on rising glass-fiber prices, Jukan wrote that Unitika, a Japanese materials maker with a glass-fiber cloth business, would go up even more. Over the next 90 days it fell **56.2%**, while SPY gained 7.4% and SMH 32.5%.

Across 44 calls, the account averaged **+24.1%** at 90 days, against +4.9% for SPY and +21.9% for SMH; 72.7% of the calls finished positive. At 180 days the lead over SMH was wider: **+47.7% versus +35.5%** across 34 calls.

**Caveat:** the sample is small. Remove the two best calls and the 90-day lead over SMH becomes a two-point deficit (+19.4% versus +21.6%). The nine calls outside the AI hardware chain averaged −0.1%, behind SPY's +7.6% on those dates.

## 2. Bull of Britain (@BULLOFBRITAIN): AI supply chain lists

[Bull of Britain](https://buzzberg.ai/speakers/bullofbritain) maps the AI hardware supply chain layer by layer: fabs, memory, co-packaged optics, equipment and substrates, including Asian listings that rarely appear on US feeds.

During [Nvidia](https://buzzberg.ai/trade-ideas/ticker/NVDA)'s GTC conference on [March 16](https://x.com/BULLOFBRITAIN/status/2033652087620260011), the account posted a list of "Quick Fire Long positions" covering nearly 30 companies. Buzzberg scored 20 of them as separate calls. From the next day's close, [Marvell](https://buzzberg.ai/trade-ideas/ticker/MRVL) returned **+240.2%** over 90 days, [Arm](https://buzzberg.ai/trade-ideas/ticker/ARM) +224.1% and [Intel](https://buzzberg.ai/trade-ideas/ticker/INTC) +190.2%. The 20 calls averaged +78.4%, against +60.2% for SMH on the same days; 11 of them beat it.

The miss came on [June 3](https://x.com/BULLOFBRITAIN/status/2062255649354883093), in a post that began "Bought $MRVL $NVTS $Kioxia here." Navitas Semiconductor fell **64.3%** over the next 90 days and [Kioxia](https://buzzberg.ai/trade-ideas/ticker/285A.T) 33.7%, while SMH fell 13.1%.

Over 90 days, the 93 calls averaged **+20.6%**, against +6.2% for SPY and +23.5% for SMH. At 180 days the account ranks first on the leaderboard: **+43.6% versus +36.7%** for SMH across 56 calls.

**Caveat:** one list post can add 20 calls at once, and when a whole sector rallies, a long list catches the rally. The record also includes an April [AXT](https://buzzberg.ai/trade-ideas/ticker/AXTI) "short report" that Buzzberg scored as a short, which lost 99.3% as the stock kept rising. In [a later post](https://x.com/BULLOFBRITAIN/status/2062173940714229970) the account explained "I write the short thesis on my own longs" and said it had bought AXT at $13, so that label is open to question. Counting only the first call on each ticker, which drops later re-entries such as that short, the 90-day average rises to +30.2%, against +27.9% for SMH.

## 3. BryzonX (@BryzonX): the CPU call

[BryzonX](https://buzzberg.ai/speakers/bryzonx) writes about the parts of AI infrastructure beyond the GPU: processors, power equipment, batteries and sensors.

On [March 24](https://x.com/bryzonx/status/2036553546255245330), the account argued that CPUs would "have their moment in the supply chain" and named [AMD](https://buzzberg.ai/trade-ideas/ticker/AMD), Intel and Arm as the pure plays. AMD was already an open call, so Buzzberg scored Intel and Arm from this post. Over the next 90 days, Intel returned **+198.7%** and Arm +159.6%, against +67.6% for SMH.

The miss came on [June 15](https://x.com/BryzonX/status/2066345165971915255). The post argued that Nvidia's move to 800-volt DC power in data centers would force operators to buy battery storage, turning Fluence Energy's agreements with large cloud companies into orders. Fluence fell **61.1%** over 90 days, while SPY gained 0.8%.

The 49 calls averaged **+14.5%** at 90 days, against +5.5% for SPY and +23.8% for SMH. At 180 days, 38 calls averaged +34.6%, against +38.6% for SMH.

**Caveat:** just over half the calls were on the AI hardware chain, and those averaged +33.1%, well ahead of SMH's +21.4%. The other 22 averaged −8.3%, behind SPY's +5.4%.

## 4. Serenity (@aleabitoreddit): hunting bottlenecks at scale

[Serenity](https://buzzberg.ai/speakers/aleabitoreddit) is the largest account on this list, with about a million followers, and the most prolific: more than 240 scored 90-day calls since July 2025. The method is to find the small supplier that a much larger industry cannot do without.

On [December 11, 2025](https://x.com/aleabitoreddit/status/1999240464420250042), Serenity explained the reasons for owning [Lumentum](https://buzzberg.ai/trade-ideas/ticker/LITE) and introduced [Applied Optoelectronics](https://buzzberg.ai/trade-ideas/ticker/AAOI), one of two photonics stocks the account said it held, as a small-cap supplier of optical connections for [Amazon](https://buzzberg.ai/trade-ideas/ticker/AMZN)'s Trainium clusters. A reader who bought AAOI at the next day's close made **+296.2%** in 90 days. SMH gained 13.2%.

The miss came on [May 7](https://x.com/aleabitoreddit/status/2052184498620834237). Serenity suggested going long on glass-core substrate maker SKC and inspection supplier HB Technologies, while disclosing only "very small exposure to both." They fell **53.2%** and **55.1%** over 90 days. SMH gained 5.5%.

At 90 days, the 234 priced calls averaged **+11.0%**, against +4.1% for SPY and +18.8% for SMH. At 180 days the figures were +17.4% and +30.7%.

**Caveat:** breadth cuts both ways. The 87 calls on the AI hardware chain averaged **+31.0%**, well ahead of SMH's +18.6%. The other 147, including bitcoin miners and AI cloud companies such as [IREN](https://buzzberg.ai/trade-ideas/ticker/IREN), which the account called eight separate times, averaged −0.8%, behind SPY's +4.1%. Several of Serenity's 2025 posts no longer load on X; for those, Buzzberg's saved copy is the remaining record.

## 5. Caesar Capital (@CaesarCapitalz): one winner the 90-day measurement leaves out

[Caesar Capital](https://buzzberg.ai/speakers/caesarcapitalz) trades a wide mix of stocks, from AI cloud and optical names to brokers, crypto companies and defense, and often discloses holdings in short replies. The scored record starts in March 2026.

On [March 15](https://x.com/CaesarCapitalz/status/2033196624541839616), replying to another investor about optical stocks, the account called Lumentum "probably the safest bet," [Coherent](https://buzzberg.ai/trade-ideas/ticker/COHR) second, and AAOI and AXT the "high risk/high reward plays." The risky pair did best: AXT returned **+128.8%** in 90 days and AAOI +103.6%, against +64.3% for SMH. Lumentum, the safest bet, returned +53.2%.

The miss: on [March 9](https://x.com/CaesarCapitalz/status/2031121759055815149) the account mentioned owning TransMedics. It fell **45.6%** over 90 days, while SPY gained 9.2% and SMH 50.5%.

The 61 calls in our measurement averaged **+14.2%** at 90 days, against +6.9% for SPY and +29.8% for SMH. At 180 days, 47 calls averaged +18.2%, against +32.5% for SMH.

**Caveat:** Caesar Capital ranks first on the 90-day leaderboard largely because of one reply. On [March 19](https://x.com/CaesarCapitalz/status/2034753801328967932) the account said it would add [Sivers Semiconductors](https://buzzberg.ai/trade-ideas/ticker/SIVE) to its portfolio soon. The leaderboard records **+871%** from the price when the post appeared. From the next day's close the stock rose about twelvefold over 90 days. The move looks genuine in both its Stockholm and US listings, but our measurement, like the leaderboard, sets aside moves above tenfold as possible data errors, so the call is not in the 90-day figures above. Counting it, the 90-day average would be +31.8% across 62 calls, against +30.3% for SMH: about level. The 180-day figures do include it (+220.1%); without it, the 180-day average is +13.8% against +32.3% for SMH. Several of the calls quoted here are one-line replies.

## 6. TheValueist (@TheValueist): research notes on AI power and photonics

[TheValueist](https://buzzberg.ai/speakers/thevalueist) publishes long research notes on power producers serving data centers, photonics suppliers and large technology companies such as [Oracle](https://buzzberg.ai/trade-ideas/ticker/ORCL) and [Alphabet](https://buzzberg.ai/trade-ideas/ticker/GOOGL). Every one of the account's scored calls is a long.

On [January 18](https://x.com/TheValueist/status/2012717399074570734), TheValueist wrote of liking Lumentum and Coherent "as long term longs" and pointed readers to earlier threads. Over the next 90 days, Lumentum returned **+150.9%** and Coherent +79.6%, against +18.8% for SMH.

The miss came on [April 30](https://x.com/TheValueist/status/2049944315271541174), when the account shared AXT's first-quarter results alongside its February photonics supply-chain thesis, which included the stock. Buzzberg scored a new long from this post. A reader buying at the next day's close lost **61.5%** over 90 days; SMH was roughly flat (−1.1%).

The 131 calls averaged **+16.5%** at 90 days, against +4.4% for SPY and +14.6% for SMH; 58.8% finished positive. At 180 days, 102 calls averaged +11.8%, against +23.6% for SMH.

**Caveat:** most of the 90-day lead over SMH rests on one call. On [March 25](https://x.com/TheValueist/status/2036829014900240629) the account published a research note on Sivers whose conclusion was guarded: it relayed an analyst's view that much of the upside appeared priced in. Buzzberg scored it as a long, and a reader buying at the next day's close would have made 672% in 90 days. Without that call, the 90-day average falls to +11.4%, behind SMH's +14.3%.

## 7. Temple 8 Research (@Temple_Eight): optics, macro and shorts

[Temple 8 Research](https://buzzberg.ai/speakers/temple-eight) mixes semiconductor and optical calls with macro trades on the Nasdaq-100, oil, the yen and Japanese stocks. Thirty percent of its scored calls are shorts, more than any other account here.

On [January 19](https://x.com/Temple_Eight/status/2013242544369672613), the account read Marvell's warrants to Amazon as evidence that it would supply Trainium 4 and recommended it as a 2026–2027 play. Marvell returned **+85.3%** over 90 days, against +18.8% for SMH.

The miss came on [March 10](https://x.com/temple_eight/status/2031415264370454747), when the account called AXT the best optical name to short, expecting China to restrict exports of indium phosphide. AXT nearly doubled over the next 90 days, and the short lost **91.7%**.

The 128 calls averaged **+7.1%** at 90 days, against +3.3% for SPY and +8.6% for SMH. At 180 days, 59 calls averaged +14.9%, against +24.6% for SMH.

**Caveat:** the 90-day record is about level with SMH, not ahead of it. In its favor, the 38 priced shorts made money on average: +8.1% over 90 days, while SPY rose 2.6% on the same dates.

## 8. Pep Invest (@Pep_Invest): small-cap AI hardware, high risk

[Pep Invest](https://buzzberg.ai/speakers/pepinveststocks), formerly @PepInvestStocks, writes detailed threads on small AI hardware suppliers: indium phosphide, test equipment, photonics and power chips.

On [February 11](https://x.com/PepInvestStocks/status/2021613540646301978), the account called AXT the owner of "the critical choke point for the AI data center photonics boom." A reader buying at the next day's close made **+395.8%** in 90 days, against +38.2% for SMH.

The miss came on [August 12](https://x.com/PepInvestStocks/status/2087652362190049571), in a reply saying the account owned [Boost Run](https://buzzberg.ai/trade-ideas/ticker/BRUN), Sivers and two other small caps. Boost Run fell **71.7%** over 90 days; SMH fell 17.9%.

The 81 calls averaged **+10.5%** at 90 days, against +3.5% for SPY and +18.7% for SMH. Only 39.5% finished positive, and the median call lost 9.8%. At 180 days, 51 calls averaged +22.4%, against +39.9% for SMH.

**Caveat:** a few huge winners carry the record. Without the two best 90-day calls, the average falls to **+2.5%**, below SPY's +3.4%.

## 9. KaizenInvestor (@Kaizen_Investor): space stocks and memory

[KaizenInvestor](https://buzzberg.ai/speakers/kaizen-investor) covers space companies, semiconductors and a broad set of other stocks, often with a full thesis and the price paid.

On [March 3](https://x.com/Kaizen_Investor/status/2028828493299978299), the account called [SK hynix](https://buzzberg.ai/trade-ideas/ticker/000660.KS) "a clear buying opportunity" at about 5.4 times forward earnings, crediting its hold on Nvidia's high-bandwidth memory supply. It returned **+178.3%** over 90 days. On [March 11](https://x.com/Kaizen_Investor/status/2031718305602400605), the account disclosed a Marvell position opened at $88.66; from the next day's close, it returned +204.4%. SMH gained 52.3% over both windows.

The miss came on [June 3](https://x.com/Kaizen_Investor/status/2062186211238670713). The post argued that space stocks, including [AST SpaceMobile](https://buzzberg.ai/trade-ideas/ticker/ASTS) and Intuitive Machines, had reached record institutional ownership and that the SpaceX IPO was "just the beginning" for other space stocks. The four companies Buzzberg scored from that post fell between **43% and 56%**.

The 56 calls averaged **+12.0%** at 90 days, against +3.3% for SPY and +16.3% for SMH; 39.3% finished positive. At 180 days, 38 calls averaged +11.9%, against +40.9% for SMH.

**Caveat:** one call carries the record. On [March 30](https://x.com/Kaizen_Investor/status/2038647612090278136) the account called Sivers "the only investable option out of the pack"; a reader buying at the next day's close made +512.0% in 90 days. Without it, the other 55 calls averaged +2.9%, against +3.2% for SPY. The 21 calls on the AI hardware chain averaged **+47.2%** against +24.9% for SMH, or +24.0% against +22.9% without Sivers. The 35 calls outside the chain, spread across space, power, software, healthcare and ETFs, averaged −9.1%.

## Where the returns came from

Across the nine accounts, 407 of the 877 calls at 90 days were on the AI hardware supply chain: chips, memory, optical networking and servers. Together they averaged **+28.3%**, ahead of SMH's +18.8% on the same days. But the median call made +9.6%, against a median of +14.0% for SMH, and only 42.5% of these calls beat SMH. Leave out the ten best and the average falls to +20.4%, close to SMH's +18.2%.

The groups behaved very differently:

- **Memory and storage** stood out: 39 calls averaged +46.1%, against +11.8% for SMH, and 69% of them beat it.
- **Optical networking** averaged +32.8% against +20.3% for SMH, but the median call made only +1.0%. A few multi-baggers did the work, led by two Sivers Semiconductors calls above +500%, then AXT and Applied Optoelectronics.
- **Chip makers, designers and equipment suppliers,** the part of the chain closest to what SMH holds, averaged +24.2% against +20.3%. The median call made +9.2%, against +15.4% for SMH, and 41% beat it.

The other 470 calls averaged **+0.5%**, below SPY's +4.3%, and only 39% beat SPY. At 180 days the split was sharper: +47.0% versus +31.1% for SMH on chain calls, and −0.5% versus +8.1% for SPY on the rest.

So the honest summary has three parts. These accounts were mostly right to be in the AI hardware trade. Inside it, memory stocks and a handful of optical winners lifted the average above SMH, while the typical chip call trailed the fund. Outside it, their calls lost to the S&P 500.

AXT shows how much timing mattered. Six of these accounts made ten calls on it. The early longs, from Serenity in December and Pep Invest in February, returned roughly four- and fivefold for a new reader over 90 days. Temple 8 Research's March short lost 91.7%, and Bull of Britain's April report, scored as a short, lost 99.3%. TheValueist's late-April post, scored as a long, lost 61.5%, and later longs in June and August lost too. The same company produced opposite results depending on when the call was made.

## How to use this list

If you are looking for fintwit accounts to follow on memory and the optical supply chain, most of these nine have picked better than the sector ETF there. Be more skeptical of their calls outside it, and of chip calls that simply track the sector.

Check the date and the run-up. Many of the biggest winners came from posts published between December and March. Some misses came after a stock had already moved: Jukan's Unitika post expected it to go up "even more," and TheValueist's late-April AXT post came after the stock had risen about fivefold since late December.

Read replies and lists differently from theses. A one-line "I own it" and a list of nearly 30 names both count as calls, but they tell you much less about the author's reasoning.

Pick the right benchmark. The best Twitter accounts for stocks in 2026 still had to beat SMH on their chip ideas, not only the S&P 500. Each account's profile lists its scored calls (the full history after a free sign-in), and with [Buzzberg MCP](https://buzzberg.ai/mcp) your AI assistant can pull an account's calls and the posts behind them.

## How we measured it

**Selection.** We used Buzzberg's Authors leaderboard on November 11, 2026, limited to X, all publication dates and authors with at least 30 scored calls, ordered by Buzzberg's Alpha score. We kept the accounts in the top 20 at both the 90-day and 180-day return horizons. The list order is the sum of each account's two positions; Jukan and Bull of Britain tie, and Jukan has the higher average Alpha score. Because the leaderboard ranks authors by how far their calls beat the S&P 500, the selection favors accounts with strong recent results; it does not show that those results will continue.

**Numbered calls.** Each scored call is an author's numbered LONG or SHORT entry on a ticker. Repeating a view while the call is open does not create another call, unless 90 days pass without one; then the old call ends and a new one starts. An exit post or a reversal also ends a call. One post that names several tickers creates one call per ticker.

**Entry and exit.** Profile pages measure from the price when a post appeared. A reader cannot always buy at that moment, so this article re-measures every call from the close of the first trading day after the day of the post, in New York time. The exit is the close of the first trading day after the post that ended the call (an exit, a reversal, or a repeat after 90 quiet days), or the leaderboard's 90- or 180-day exit, dated from the original post, if none came first. SPY and SMH are bought and sold on the same US trading days; for foreign listings and crypto, they use the first US close dated on or after the stock's own session, which for crypto can fall a few hours before the crypto close. Foreign stocks are measured in their local currency. Post dates in the article are UTC dates.

**Data checks.** Like the leaderboard, we set aside calls whose price moved more than tenfold or returned more than 900%, which usually signals a data error. We also set aside calls whose daily price history jumps more than threefold in a single day. One post time that Buzzberg had saved late was corrected from X before the entry was measured.

**Averages.** Each call carries equal weight. Returns are price changes only, before dividends, fees and taxes. A short gains when the stock falls.

**Repeats and coverage.** A ticker can appear more than once for the same author, and one list post can add many calls. Keeping only each author's first call on a ticker leaves 755 calls averaging +16.4% at 90 days, against +4.7% for SPY and +19.8% for SMH. Of the 900 scored 90-day calls in our November 11 extract, 877 could be measured this way. Ten were closed before a reader could buy, six had no matching benchmark session, five had no daily price, and two failed the data checks. At 180 days, 642 of 664 could be measured. The AI hardware chain uses Buzzberg's AI compute, custom chip, chip manufacturing and equipment, analog and power chip, substrate, memory and storage, networking and photonics, and AI hardware categories.

The figures are a snapshot. Leaderboard positions change daily as calls mature, so a later check can produce a different list.

## More research

[All research ↗](https://buzzberg.ai/blog)

- [12 min read
  9 Best Investing YouTube Channels to Follow in 2026
  Company research, AI hardware and crypto: SemiAnalysis, All-In, Forward Guidance, Delphi and more. Original episodes and four holding periods.](https://buzzberg.ai/blog/best-investing-youtube-channels)
- [9 min read
  Who Called the AI Infrastructure Winners Early in 2026?
  The dated arguments behind Sandisk, AXT, Dell, Backblaze and other AI infrastructure moves, with the misses checked alongside the winners.](https://buzzberg.ai/blog/ai-infrastructure-winners-2026)
- [8 min read
  QQQ vs SPY vs SMH: What Should AI Stock Picks Beat?
  A same-date 2026 test of AI-related stock views against the S&P 500, Nasdaq-100 and semiconductor ETFs.](https://buzzberg.ai/blog/qqq-vs-spy-ai-stock-picks-2026)

© 2026 Buzzberg[Markdown](https://buzzberg.ai/blog/best-stock-twitter-accounts-2026/index.md)[Disclaimer](https://buzzberg.ai/disclaimer)[Privacy](https://buzzberg.ai/privacy)[Back to top ↑](https://buzzberg.ai/blog/best-stock-twitter-accounts-2026#top)