Anthony Wood's ROKU April Drip: FMR Still Holds 10.5%

The most recent Form 4 lines for Anthony Wood, Roku's founder and CEO, document the kind of cadence that retail investors keep mistaking for bearish insight: a few dozen thousand shares converted from super-voting Class B to Class A, sold the same day on a staggered execution schedule, and reported the next morning. April's filings totaled about $7.7 million across a handful of trading days. Held against Wood's 16.27 million remaining Class B-equivalent shares, that is roughly 0.5% of the founder's reported economic position.

The more interesting part of the tape sits one filing type over. Roku's 13G/A docket has been moving more decisively than the founder Form 4s — and it is moving in directions that don't line up with the simple "CEO is selling, others are selling too" reading.

What the April Form 4 actually shows

The April 16 line is the structural one. It's a Table II conversion of 25,000 Class B shares into Class A, paired with a same-day Class A sale of an identical 25,000 shares at $110.19. Class A shares held after the sale: 0. Class B-equivalent derivative holdings remaining: 16,268,111. That is the headline that gets reporting wrong — "CEO holds zero shares" is a Table I reading; the actual holding is in the Table II derivative line that almost no reporter looks at.

The pattern across April is a textbook 10b5-1 mechanical drip: same-day same-amount convert-and-sell, staggered execution prices that look algorithmic, regular cadence rather than block-deal optics. Whatever Wood's view of Roku, it was crystallized when the trading plan was set; April's executions are the calendar talking, not the CEO.

The 13G tape says something different from the Form 4

Strip out Anthony Wood entirely and look at who has been declaring positions in Roku above the 5% reporting threshold in 2026:

• Fidelity (FMR LLC) at 10.5% via a Schedule 13G/A filed February 5, 2026, with 13,679,650 shares. That is up from earlier 9%-handle filings in 2024 — a managed build, not an arrival.
• ARK Investment Management at 4.5% per its January 8, 2026 13G/A. The under-5% filing is itself the news: ARK is no longer required to file going forward, which means subsequent activity will only show up in 13Fs at quarterly cadence. The position dropped below the threshold sometime between mid-2025 and late 2025.
• Vanguard Capital Management at 5.26% per a fresh SC 13G filed April 30, 2026, with 6,876,921 shares — filed roughly five weeks after Vanguard Group's separate March 27 exit amendment. As with PayPal and other large-caps, this is almost certainly an internal Vanguard reporting-entity transfer rather than a new directional view.

Put together: the largest active-conviction institutional holder of Roku (Fidelity) has been adding while the founder is on his 10b5-1 calendar, and the second-most-recent declared holder (ARK) is the one quietly shrinking. Reading just the Form 4 tape gets the picture backwards.

The Class B math under the founder stake

The conversion mechanic deserves one paragraph of its own because every retail commentary thread we've seen on Roku gets it slightly wrong.

Roku's dual-class structure has Class A trading on Nasdaq at one vote per share, and Class B held privately at 10 votes per share. Class B converts one-for-one into Class A on sale by the holder — that's the "C" code on April 16's Table II line. Wood's 16,268,111 remaining derivative shares are effectively a privately-held Class B block convertible into Class A on demand. At roughly $90 a share for Class A, that block is worth on the order of $1.5 billion in economic exposure, before counting any unvested grants.

So when Form 4 Table I says "Class A shares held: 0" after April 16, the correct reading is: he just sold the 25,000 Class A shares he had on hand from converting Class B that morning. He still controls fifteen times that many shares via the Class B block. Anybody writing about a Roku founder "exit" without that framing is misleading their readers.

What to track from here

Two anchors for forward reading:

• The next 13G/A from Fidelity: FMR's stake going from 9% in 2024 to 10.5% in February 2026 was a multi-quarter build. The next amendment will either confirm continued accumulation (Roku's reweighting in their large-cap growth funds) or mark a reversal. Either is more informative than the founder's April calendar.
• Q2 2026 13F filings (due August 14): with ARK Investment Management now under the 5% threshold and ARK's quarterly disclosures the primary tracking surface for their active conviction, the Q2 13F will show whether they continued trimming or stabilized the position. Roku has been a meaningful weight in ARKK historically.

Also useful: anyone tracking founder-led tech names with Class B super-voting can compare Wood's cadence against Matthew Prince's April Cloudflare Class B drip. Different companies, near-identical mechanics — and both filings need to be read with the same multi-class framework or the Form 4 headline will fool you.

See Roku's full institutional ownership breakdown for the holder-level detail behind these 13G filings, and Anthony Wood's transaction history for the cumulative shape of the founder sell program.