BofA Sharply Raises Tencent Capex Forecast, AI Agent Strategy Underestimated
Alina Collins
BofA hiked Tencent's 2026–2028 capex forecasts by roughly 23%–25%, warning that free cash flow will shrink sharply — the core debate is how much Tencent's lead in China's AI-agent orchestration layer is actually worth.
How much did BofA raise, and where does the money go?
BofA now projects Tencent capex at RMB 185 bn / 225 bn / 250 bn for 2026/2027/2028, up 23%–25% from prior estimates.
This means → Tencent is converting profits into AI infrastructure at a pace far more aggressive than the Street had priced in.
BofA kept its Buy rating and HK$780 target unchanged — in other words, it views the spending as value-accretive.
With that much spending, how badly do profits get squeezed?
2026 free cash flow is forecast at roughly RMB 67.2 bn, down from RMB 215.6 bn in 2025 — a drop of nearly 70%.
In plain terms = Tencent flips from a cash machine to a cash furnace in the near term; the balance-sheet cushion tightens fast.
2027/2028 non-IFRS net profit — a profit measure that strips out one-off items — was cut by 3% and 6% to RMB 279.3 bn and 292.7 bn, respectively.
AI agents — what exactly is the market undervaluing?
Tencent's WorkBuddy currently ranks first in China by AI-agent user base, yet BofA argues the market has not fully priced this position.
This means → the slot is both the distribution gateway for AI models and a flywheel for user–model interaction data — whoever holds the gateway controls the starting point for the next wave of monetization.
In plain terms = think of it as the "WeChat moment" for AI agents — WorkBuddy is racing to lock in the same kind of platform chokepoint.
How will WeChat's AI agent actually make money?
BofA sees two core edges: first, Tencent's AI infrastructure plus a relatively lean WeChat-native model keep the token cost — the per-query computing bill — below rivals.
Second, Tencent can build a revenue-sharing framework with ecosystem partners. This reflects a platform playbook, not a winner-take-all grab.
Recent foundation-model upgrades and the WeChat AI-agent beta launch could, BofA argues, flip market sentiment from bearish to bullish.
Will Q2 earnings look good?
BofA forecasts Q2 revenue growth of 8.8% YoY, below the 10% consensus; adjusted net profit growth of 4.7%, below the 9% consensus.
This means → on a headline basis, the quarter will not excite — AI spending eats nearly half the profit-growth rate.
Strip out AI investment and adjusted operating profit grows 14% YoY — the underlying business remains healthy.
How are the individual business lines doing?
China gaming revenue is expected to rise 10% YoY, driven by incremental revenue recognition from *Delta Force* and steady performance from *Honor of Kings*; international gaming slows to 9% YoY on a high base.
Online advertising is forecast at 18% YoY growth, powered by continued Video Accounts ad ramp; cloud is expected to accelerate to 25% YoY.
BofA projects Q2 AI-related opex at RMB 10.5 bn, up from RMB 8.8 bn in Q1 — whether capex translates into quantifiable agent-driven revenue will be the key re-rating trigger for Tencent's AI valuation.
Content is for reference only, not financial advice.