HLIT — Harmonic Inc. · 2026-09-05 · Verdict: WATCH · Conviction 3
Price $11.78 (screen row, universe_under2b.csv dated 2026-09-05; no live prices available) · Mkt cap $1.285B · EV $1.164B · EV/EBIT 82.4x (screen, on CY2025 GAAP EBIT) or 14.5x on FY26 guided GAAP operating profit · FCF yield 7.5% (screen, overstated) · Net cash $121.3M · ADV $21.5M Sources read: 10-K 2026-02-24 (Items 1, 1A, 7), 10-Q 2026-08-17 (Part II), DEF 14A 2026-04-24, 8-Ks 2026-04-02, 2026-05-11 (Q1 release), 2026-06-05, 2026-06-08, 2026-06-17 (divestiture close), 2026-08-12 (Q2 release), Form 4s (12 most recent of 56). No transcript in the pack.
1. What the business actually does
Harmonic sells broadband access equipment and software to cable and telco operators. The product is cOS, a virtualized platform running DOCSIS and fiber PON services on off-the-shelf servers, sold with Harmonic's own access nodes, PHY shelves and OLT modules, deployed with 161 customers serving 48.2 million CPE devices (10-K 2025, Item 1; 8-K 2026-08-12). Manufacturing is outsourced, mostly to Plexus in Malaysia. Revenue is 84% hardware and integration, 16% SaaS and service, and 89% Americas (10-K 2025, Item 7).
2. Why it is mispriced: the edge case, and its limits
The edge type is post-restructuring, and the screen's 82.4x EV/EBIT is a genuine artifact. On December 8, 2025 Harmonic agreed to sell its Video business to MediaKind and reclassified it as discontinued operations for all periods; the sale closed June 16, 2026 for $137.9M (10-K 2025, Item 7; 8-K 2026-08-12). What remains, Broadband, earned only $14.1M of GAAP operating profit in 2025 on $360.5M of revenue because DOCSIS 4.0 deployments slipped: Americas appliance revenue fell $148.0M on "customer deployment timing delays associated with DOCSIS 4.0 and network readiness" (10-K 2025, Item 7). Against management's FY26 operating profit guide of $74M to $86M, the same $1.164B enterprise value is 14.5x at the midpoint (8-K 2026-08-12).
The honest part is that the market has already made this correction. The stock is up 25.6% in six months and 17.7% over twelve, sits only 31.2% below its 52-week high, and trades $21.5M a day with BlackRock at 16.23% and Vanguard at 11.09% (DEF 14A 2026-04-24). There is no spin, no index deletion, no forced seller and no coverage gap. The screen is wrong; the price is not obviously wrong.
3. Unit economics and growth
The recovery is real. Q2 2026 revenue was $133.5M, up 53.6%, with GAAP operating profit of $23.6M against a $0.8M loss a year earlier; H1 revenue was $255.2M and H1 operating profit $44.1M (8-K 2026-08-12). Gross margin went from 48.5% in 2025 to 52.3% in H1 2026, guided at 50.9% to 51.8% for the year. Guidance has been raised twice: to $475M-$495M of revenue and $64M-$78M of operating profit on May 11, then to $505M-$525M and $74M-$86M on August 12.
The constraint is bookings. Backlog and deferred revenue rose from $332.3M at the end of 2024 to $573.8M at the end of 2025, implying roughly $602M of 2025 bookings against $360.5M of revenue, of which $346.9M landed in Q4 2025 alone (8-K 2026-05-11). All of 2026's growth is the conversion of that one quarter. Since then backlog has risen only $13.8M, to $587.6M, because H1 2026 bookings of $260.2M barely exceeded H1 revenue: book to bill of 1.02 (8-K 2026-08-12). H1 bookings are up 27.6% from $203.9M, which is good, but annualize to about $520M, below 2025's $602M. 2026 is covered; 2027 is a bookings question the filings do not answer. Only about 53% of backlog converts within a rolling year (10-K 2025, Item 1).
Concentration is severe. One customer was 54% of 2025 revenue (10-K 2025, Item 1A); the top two were 63% of Q2 2026 revenue, up from 58% in Q1. Rest-of-Market grew 44% and was roughly 60% of Q2 bookings, the encouraging diversification datapoint (8-K 2026-08-12).
4. Balance sheet and capital allocation
Cash of $231.9M against $110.6M of total debt ($2.9M current, $107.7M long term) gives $121.3M of net cash, $1.11 a share; the screen counted only the non-current portion, a $2.9M error that changes nothing. Shares fell from 111.2M at year end to 109.0M at July 3, 2026: Harmonic repurchased 4.2M shares for $43.0M in Q1 at roughly $10.24, then bought nothing in Q2 at an average price of $12.95, with $78.0M still authorized (8-K 2026-05-11; 10-Q 2026-08-17). Insiders own 1.74%, and across the twelve most recent Form 4s there is not one open-market purchase or sale. Shareholders approved 3,000,000 more plan shares in June, 2.8% of the count (8-K 2026-06-05). Executive PRSUs vest on Broadband bookings targets for 2025 to 2027 (DEF 14A 2026-04-24), the metric that has flattened.
5. Management: said versus did
Two raises in two quarters, and Q2 revenue of $133.5M beat the $115M-$125M guided in May. Against that, the FY26 guide implies H2 operating profit of $30M to $42M after $44.1M in H1, and Q3 revenue of $125M-$135M is below the $133.5M just delivered: the CEO's language is growth ("business momentum continued") while the arithmetic in the same release guides profit down sequentially. About $10.0M of FY26 stranded divestiture costs sit in that guide, and nothing in the bundle commits to removing them (8-K 2026-08-12).
6. Valuation
Reverse DCF: at $11.78 the $1.164B enterprise value against FY26 guided operating profit of $80M taxed at 23% is 18.9x NOPAT, implying at a 10% discount rate about 4.7% perpetual growth in after-tax operating profit from the guided level.
Base (50%): 2027 revenue flat at $525M on book to bill near 1.0, 16% operating margin, $84M EBIT at 15x plus net cash, $12.66. Bear (25%): the DOCSIS cycle pauses as it did in 2025, when this exact business fell from $488.2M to $360.5M of revenue in one year (10-K 2025, Item 7). Revenue back to $400M at 1.3x sales plus net cash, $6.00. Bull (25%): DOCSIS 4.0 broadens, Rest-of-Market compounds, 2027 revenue $600M at a 19% margin with stranded costs removed, $124M EBIT at 18x plus net cash, $21.57. Probability weighted $13.22, about 12% above price and inside the error bars of the assumptions.
7. Catalysts and timeline
Investor Day in New York on September 15, 2026, where the 2027 framework is what to listen for (8-K 2026-06-17). Q3 results in November. Most important, Q4 2026 bookings: whether the $346.9M pattern of Q4 2025 repeats decides 2027.
8. Kill criteria (pre-registered, observable)
- Bookings below reported revenue in two consecutive quarters, or backlog and deferred revenue below $560M at any quarter end.
- Initial FY2027 revenue guidance below $500M.
- GAAP operating margin below 12% in any quarter with revenue above $120M.
- Zero buyback for two further quarters with $78M authorized and the stock below $12.
9. Verdict
WATCH, conviction 3. Harmonic is a real post-divestiture cleanup where the screen is genuinely broken: a $1.164B enterprise value against $14.1M of 2025 trough EBIT reads as 82x, while the same enterprise value is 14.5x the $74M to $86M of operating profit management has now guided twice upward, on 52% gross margins with $121.3M of net cash and a 2% smaller share count. What it is not is a mispricing anyone is forced to hand you: the stock is up 25.6% in six months on $21.5M of daily volume with BlackRock and Vanguard owning 27% between them, so the artifact has already been paid for, and 14.5x forward operating profit for a supplier whose top two customers are 63% of revenue is fair rather than cheap. The decisive fact is that all of 2026's 40%-plus growth is the conversion of a single $346.9M booking quarter in Q4 2025, and since that quarter backlog has grown $13.8M on a book to bill of 1.02, which supports flat 2027 revenue rather than the 4.7% perpetual growth the price implies. Management stopped buying its own stock at $12.95 after buying $43.0M at $10.24, no insider has bought in the open market, and executive performance shares vest on the bookings line that has gone flat. Wait for the September 15 investor day and the Q4 bookings print.
Research for discussion, not investment advice. Positions and sizing are the reader's decision.