Even by the hyperbolic standards of AI claims, the implications of Tencent being close to launching an embedded AI agent for WeChat are staggering. And, by virtue of its shareholding, the gains for Naspers are huge, too.
One China-based analyst wrote that if reports are accurate, the WeChat AI agent will be able to directly perform real-world tasks such as making appointments, ordering goods, purchasing tickets and paying bills – “completing the final link from intent to transaction with WeChat Pay”.
As Tencent president Liu Chi-ping indicated during the first-quarter earnings call earlier this year, the do-everything WeChat platform was designed for hosting AI agents. Its instant messaging, social media and mobile payment facilities are used by an estimated 1.4-billion people.
Zwelakhe Mnguni, chief investment officer at Benguela Global Fund Managers, describes the AI agent as a meaningful step for Tencent. “WeChat’s unparalleled daily engagement, payments infrastructure and commerce integration create a powerful distribution and data moat for an agent that can handle real-world tasks inside an app that users already live in,” he tells Currency, adding: “It’s a pragmatic way to deploy AI where Tencent already has the strongest advantages rather than competing head-on in standalone general-purpose models.”
Leaky ship no longer?
Tencent’s ongoing effort to develop a killer AI product has been no secret. But even Tencent’s founder-chair Pony Ma acknowledged it wasn’t doing too well. With unusual modesty for a tech mogul, he told a shareholders’ meeting in May: “A year ago, we thought he had boarded the [AI] ship, but later we realised it was leaking. Now we feel we’ve managed to stand on it, but we still can’t sit down comfortably – we still hope the ship can move faster.”
In the first quarter, Tencent poured $4.7bn into AI capex and reported a net loss of $1.3bn from new AI products. Management told shareholders they could expect AI capex to increase significantly in the second half of the year. They weren’t impressed.
The lack of AI progress and WeChat’s sluggish organic growth have been major contributors to the slump in Tencent’s share price since it reached a high of HK$677.50 last October. Ahead of the recent AI agent reports, it had slipped to a low of HK$427.
Mnguni reckons the selloff was overdone and disproportionate to any fundamental deterioration. The group delivered solid full-year results for 2025 and its performance was strong enough to buy at the weaker prices.
Mind you, Anchor Capital’s Mike Gresty doesn’t believe the share price slump was all down to Tencent’s operational issues, noting the flood of emerging-market funds that poured into South Korean micro-chip manufacturers. “This money was being pulled out of other investments, including Tencent,” he tells Currency.
Slow burn
While Gresty is disappointed by the “significant shift” around Tencent since October, and feels the share might have “gone off the boil”, he is always hesitant about taking a negative position on the group. “Over the years Tencent has gone through periods where the share hasn’t done well, then as a result of new investment, performance has been boosted and the share price surges,” he says.
After a slow AI start, he reckons there is now potential for substantial progress, but he acknowledges Tencent management are perfectionists, which might slow down the pace of that progress.
Still, it’s unsurprising that reports of an AI agent were so enthusiastically received. The news, unconfirmed by the company, sent the Tencent share price rocketing.
Realisation that any rollout could take some time – the compliance process could be lengthy, Tencent has limited access to high-quality chips and management is obsessed with perfection – may explain why the share price plateaued within a day of the report.
There’s also the fact the Chinese Communist Party will not easily allow a non-government entity to hold such sway over the country’s entire population. Then again, Pony Ma has been an astute political operator and Tencent would make an ideal national champion for President Xi Jinping.
The Naspers angle
So what about the smaller picture? Where does this potentially seismic development leave Naspers and Prosus’s shareholders? And for that matter, where does it leave CEO Fabricio Bloisi’s $100m moonshot payment?
For starters, considerably better off than if they were just relying on Bloisi’s increasingly food-dependent e-commerce rump.
Inevitably the recent Tencent surge dragged Naspers and Prosus with it; just as its previous seven-month decline has dragged them down.
The problem is Bloisi’s two major plays – the €4.1bn Just Eat Takeaway (JET) investment in Europe and additional investment in Brazil’s iFood – have been looking decidedly off-colour.
Gresty, for one, is disappointed. Especially as the market took an initial shine to Bloisi, the former iFood CEO. “He said he wasn’t going to fundamentally change things but would introduce a more entrepreneurial culture; he also shook-up the management team, which was necessary,” says Gresty. And he committed to keeping the share repurchase programme on track. What wasn’t to like?
Initially things did look good. Prosus seemed on course to build a profitable existence outside of Tencent.
But then Bloisi went on a bit of a food binge. He not only spent €4.1bn buying JET in Europe, he started pumping money into iFood in Brazil to secure its dominant position.
Prosus investors, as Protea Capital Management’s Jean Pierre Verster tells Currency, were disappointed by the capital allocation decisions made by Bloisi’s predecessor, Bob van Dijk. The worry is that they’re in for more of the same.
It’s not just that food delivery has lost its allure with investors; JET’s performance under the new owners has been uninspiring. And that’s not all, says Verster.
Food delivery own goal
Buying JET has resulted in the forced disposal by Prosus of its 26.3% stake in Delivery Hero – hardly surprising given the stance of EU competition regulators. This means Prosus is now a forced seller in a market without too many buyers. Except there is one – Uber Eats, the giant US player, which it turns out is keen to build up a strong position in the European market.
Prosus has already sold a 4.5% stake in Delivery Hero to Uber Eats and an additional 5% to Hong Kong fund Aspex. It’s been given a further 12 months by the EU Commission to complete of the sale, but unless the EU lifts the disposal obligation or blocks Uber Eats, Prosus could be in for a tough time in Europe.
Without the EU changing stance, Verster reckons Prosus could be about to hand Uber Eats a major foothold in the European market. The same market Prosus was intent on dominating.
Things aren’t looking too bright in Latin America either, where iFood had been enjoying a dominant market position. Despite pumping additional investment into iFood, it is losing out to Uber Eats and to China’s Meituan, a threat that Bloisi had initially dismissed. “Now iFood is just not looking so dominant,” says Gresty.
For Mnguni what’s happening to Prosus in Europe and Latin America has a chilling sense of déjà vu. “Both the European move with JET and the continued heavy investment in iFood strike me as classic examples of chasing large addressable markets that look attractive on population and urban density statistics but prove structurally difficult because of very low barriers to entry,” he says.
So, where does all of this leave Bloisi’s $100m long-term incentive?
Until the rerating last October, Prosus/Naspers did look on track to deliver a doubling in its aggregate market capitalisation between July 2024 and June 2028. Generating shareholder returns that were greater than half of a set peer group over the four years also looked a possibility. But the Prosus share is now just 20% ahead of its July 2024 level and profit growth is not looking strong.
And that means that at this stage, as ever, it will be down to Tencent. So everything, in other words, depends on a world-beating AI agent emerging in the reasonably near future.
ALSO READ:
Top image collage: supplied (Fabricio Bloisi); Visual China Group via Getty Images (Pony Ma); Currency.
Sign up to Currency’s weekly newsletters to receive your own bulletin of weekday news and weekend treats. Register here.
A prudent question would be: would SA and global EM focused fund managers and investors buy Naspers or Prosus for its ex-Tencent business case? Nope, at least not at the 5%-15% levels that many SA Equity and Balanced Funds hold the shares. As individual holds, you are reliant on the Chinese government not throwing a 0700 curveball.
This is not a sound investment, there is too much “Hail Mary” involved in allocating capital to this organisation (both shares) because it’s pretty much only valued on its Tencent stake, which they are selling heavily. Billions in takeaway businesses??