Visible Alpha

Visible Alpha broker models via S&P Xpressfeed · 13 brokers · 382 line items · freshest revision 2026-06-05.

Cost Breaks Before Demand

The models describe a business with almost no top-line left: revenue grows under 3% in each forecast year, and the audience shrinks in every one. What actually moves is cost. Brokers lift cost of revenue 20.8% in FY-2026, cutting gross margin 3.6pt, and GAAP earnings halve on a non-operating swing they do not repeat. The cash pile keeps building regardless. Disagreement sits in cash flow and below-the-line items, not in revenue.

Source: derived from vendor data.

Cost, not demand, is what breaks in FY-2026

Source: derived from vendor data.

Gross margin is modeled down from 76.7% in FY-2025 to 73.0% in FY-2026 and stays there: 73.5% in both FY-2027 and FY-2028. The quarterly path shows this is not a single accounting quarter — margin falls in three consecutive quarters to 72.7% in 2QFY-2026 and only claws back to 74.7% by 2QFY-2027, still below where it started. Operating margin absorbs the same hit, 30.4% to 25.5%, with barely any recovery by FY-2028 at 26.6%.

Source: derived from vendor data.

The arithmetic is stark: revenue rises 2.9% in FY-2026 while cost of revenue rises 20.8%, so gross profit actually falls. Three years of modeled revenue growth — 2.9%, then 2.3%, then 1.9% — do not get gross profit back above the FY-2025 level until FY-2027. Whatever brokers are provisioning for on the cost line, they treat it as a permanent reset rather than a one-year charge.

GAAP earnings halve on a line brokers do not model as recurring

Source: derived from vendor data.

GAAP net income is modeled down 48.6% in FY-2026; the operating measure falls 17.3%. The gap is almost entirely below the operating line. Total non-operating income swings from $193.9M in FY-2025 to a $29.0M expense in FY-2026, with the company-specific interest and other income line collapsing from $172.1M to $0.9M. On the operating measure the FY-2026 setback is real but ordinary; on GAAP it looks like a halving.

Source: derived from vendor data.

GAAP diluted EPS is modeled at 1.02 in FY-2026 against 2.04 in FY-2025, and never regains the FY-2025 level inside the forecast horizon — 1.32 in FY-2028 is still a third below. The operating measure tells a flatter story: 1.73, then 1.42, 1.46, 1.51. Which number an investor anchors on decides whether this looks like an earnings collapse or a modest step down. The headline consensus itself is covered on the CapIQ tab.

The audience shrinks every year; monetization does all the work

Source: derived from vendor data.

Average MAU is modeled down 3.1% in FY-2026 and never recovers, sliding again in FY-2027 and FY-2028. Average DAU follows, down 2.6% in FY-2026 and flat thereafter. Net adds in MAU are negative in all three forecast years. This is a shrinking audience by consensus, not one bearish model dragging an average.

Source: derived from vendor data.

Every dollar of modeled growth comes from the user, not from more users. Ad revenue per DAU rises 7.9% in FY-2026 and keeps climbing to 6.38 by FY-2028; ad revenue per average MAU goes 2.51 to 2.86. Engagement is assumed stable — the active ratio sits at 0.44 in FY-2025 and 0.45 by FY-2028 — so the models are underwriting price and load, not usage. That is the fragile part of this forecast: if pricing does not hold, there is no volume to fall back on.

Advertising is the only line the models grow

Source: derived from vendor data.

Advertising and marketing is the large majority of revenue and supplies all of the growth: up 5.1% in FY-2026, then 2.0% and 1.8%. Value-added services is modeled flat to slightly down across the same three years, so the entire consolidated story is one line. Alibaba-related revenue is modeled to grow faster than the rest, up 6.1% in FY-2026, which quietly raises related-party concentration rather than reducing it.

The sub-splits inside value-added services carry so little agreement that they are not worth reading as forecasts. On game-related services one model carries more than twenty times another's FY-2025 figure, and data licensing spans a comparable range across five brokers. These are definitional disagreements about where revenue is classified, not views on the business.

Where the models actually disagree

Source: derived from vendor data.

The revenue debate is nearly settled: across 12 brokers the FY-2027 interquartile band is narrow against the size of the business, and both extremes are single models. The disagreement is all below the revenue line. On GAAP net income the lowest FY-2027 model sits at roughly half the median — a gap wider than the entire modeled growth of the business over three years. Free cash flow is worse.

Source: derived from vendor data.

Read the quartiles rather than the range. On revenue and operating income the interquartile band is tight and the extremes are single models. On free cash flow the whole distribution is wide: the lowest FY-2027 model carries under a fifth of the median's cash generation. The advertising revenue spread across 10 brokers is the one place where a genuine difference of top-line view is being expressed rather than a modelling convention.

The cash pile builds; the payout does not

Source: derived from vendor data.

Net debt is negative in every modeled year and gets more so, taking net debt to EBITDA from -1.8x in FY-2025 to -3.3x in FY-2028 — a net cash position deepening rather than being spent. Dividends are modeled down 6.8% in FY-2026 and roughly flat after that, so the models assume the cash simply accumulates rather than being returned or deployed. For a business with no modeled growth, that is the most consequential assumption on the page.

What this coverage does not support

Several of the most interesting company-specific lines rest on one or two models and should not be read as consensus. Quarterly advertising revenue per customer is a single-broker line, as are total ad customers and SME customer counts; key-account and SME revenue splits carry three to four brokers against 12 on total revenue. The KPL and Jiamian revenue lines are one model each. Forward quarterly coverage also thins sharply — 1QFY-2027 and 2QFY-2027 revenue carry 5 brokers against 12 for the equivalent annual figure — so the quarterly path beyond FY-2026 is a small-sample view. The freshest revisions in this set date from June 2026, ahead of any subsequent reporting.

Headline P&L consensus, momentum and beat/miss live in the CapIQ tab.