A ‘GeminiPhone’ could boost Apple’s and Google’s AI game
Apple needs an AI bot while Google hopes to leap ahead of OpenAI
led by Amazon, Google, Facebook and the Tata Group, applied for NUE licences in partnership with companies such as Reliance and ICICI Bank as partners. Surprisingly, in 2023, RBI abandoned the whole idea, stating that it didn’t receive any proposal with value-additive solutions to India’s digital payments landscape. This reasoning sounds unconvincing. It is difficult to believe that such big players had no value-adding proposition. For argument’s sake, even if they had no new ideas but only wanted to invest, why rebuff them? Some media reports at the time suggested that NPCI-owning banks had lobbied against RBI’s NUE move.
There is also the matter of UPI being free of charge. The government’s mandate of a zero merchant discount rate (MDR) for UPI payments is not a sustainable approach. Again, to the credit of RBI, it came out with a public consultation paper in 2022 on “charges in payment systems,” suggesting different options for levying MDRs, including for UPI transactions. The RBI paper explained the issues and costs involved, clearly stating the roles of various UPI participants, ranging from the payer’s and payee’s payment service providers (PSPs) to the remitter’s and beneficiary’s banks, apart from NPCI and third-party application providers (TPAPS). However, even before RBI had received public comments on options suggested in the paper, the Union finance ministry dismissed the idea of levying a fee, stating that UPI is a digital public good of immense convenience for people and productivity gains for the economy.
To cover operating costs and reimburse banks and fintech firms for the MDR forgone, the government has a special scheme, but this budgetary provision is inadequate and the basis of apportioning the amount among UPI participants lacks clarity. As a result, not only do banks have little incentive to upgrade their digital payment infrastructure, the business models of PSPs and TPAPs that participate in the UPI system are opaque.
Foreign TPAPs with deep pockets dominate the UPI transaction business in India. At one time,
NPCI had envisaged a cap of 30% on the market share of TPAPs to prevent the system’s domination by a few, reduce systemic risk and encourage innovation. That has clearly not worked. At present, two foreign players, PhonePe and Google Pay, have a combined market share of over 80%. Once again, a clamour has arisen for encouraging domestic and smaller UPI players.
Fixing a UPI quota for each player may not be the right proposition; it would be a retrograde step, besides presenting implementation challenges. Instead, other options should be examined. Appropriate steps in this direction would include having a clear MDR policy and mandating a transparent business revenue model for participants.
NPCI, a not-for profit Section 8 company with income arising from fees, made a surplus of ₹809 crore in 2022-23. NPCI’s profits are likely to be even higher this year. Notably, except for NPCI, all other participants in the UPI system—that is, PSPs, banks and TPAPs—are profit-maximizing entities. There might have been some logic of setting up NPCI as a not-for-profit company, but there isn’t any for it to remain so. NPCI should be converted into a regular company. Not just that, going forward, it should aspire to get listed for public trading. That would give its working a new and welcome level of transparency, unlock its true value and encourage fresh investments.
To conclude, the UPI system and NPCI are in need of urgent reforms. To begin with, UPI transactions should bear an MDR fee. PSPs and TPAPs should be mandated to follow a transparent business revenue model to facilitate competition. This should be followed by licensing a few NUEs—perhaps two or three entities with interoperability among themselves and with UPI. As a monopoly in control of over 80% of total digital payments in the country may have systemic implications, it is important to reduce risks associated with market dominance. As for NPCI, it should be converted into a for-profit company and then listed on stock markets in due course.
There are many reasons Apple Inc’s stock hasn’t been so hot lately, but a big one is that investors at large feel the American company [which sparked off the personal computer revolution in the 1980s and then put smartphones in people’s hands two decades later to shift our relationship with hi-tech gadgetry once again] lacks a compelling ‘story’ on artificial intelligence (AI). By that, they mean Chief Executive Officer Tim Cook doesn’t seem as if he has much of a plan.
Google, to its credit, does have an AI story. Unfortunately, it’s a tragicomedy. Caught on the hop by upstart OpenAI, the search engine company’s Gemini model is best known not for its intelligence, but for its depiction of the first US president George Washington (1732—1799) as an African-American man, the Catholic Pope as an Asian woman and other assorted embarrassments.
This is all to say: the BFFs—that’s best frenemies forever—need each other’s support right now. They have long been rivals as mobile platforms, but news about talks of a tie-up between the two, which would bring Google’s Gemini AI to Apple’s iPhone, is a deal that can solve short-term headaches for both of these Big Tech companies.
The ‘GeminiPhone,’ a nickname I hope sticks purely because I know Apple will absolutely detest it, will give Apple devices a taste of the cutting-edge AI that all customers will soon expect as standard and developers are already demanding. Apple would have preferred, no doubt, to have built such a capability itself, but without the huge server farms on hand to train models, it has apparently been left behind (for now).
Google today provides enhanced AI capabilities, albeit in a still highly experimental form. Details of the partnership haven’t been filled out, and it doesn’t yet seem to have been fully agreed upon. But it has similarities with Apple and Google’s deal on search, in which Google pays handsomely to be the default search engine on iOS devices. As in that arrangement, Google gets that one thing it’s always hungry for: scale. It might at first seem curious that Google isn’t planning to keep its AI to itself, making it exclusive to Android phones, but that runs counter to its long-established North Star of simply having as many users as possible.
By delivering Gemini to the iPhone, it will not only get millions more potential users but affluent ones who might pay a premium to use advanced AI. More users means more data, and more data means a better, more valuable product.
Where this deal might differ from the one struck by Google and Apple on search, however, is that the BFFs are on a much more equal footing. Because of its laggard position, Apple needs Google’s AI in way it didn’t necessarily need Google Search. And so I would be surprised if Google ends up shovelling tens of billions of dollars into Apple’s hands for the privilege of being a default app on the iPhone—especially not when it’s being sued by the US Justice Department over that tactic for cementing its position in online search.
We also don’t know if this would be an exclusive deal, and, given those antitrust concerns, it may be wiser for it not to be— AI partnerships between tech companies are already on regulators’ radars.
According to Bloomberg’s Mark Gurman, Apple was also speaking to ChatGPT maker OpenAI, and it may well be interested in other chatbots and tools as well. Though, like search, I wonder if users will pick one general-purpose AI tool and mostly stick with it, making the firstmover’s advantage as important as ever.
Which brings me to another thought: What might Apple be giving up here? The company has shown in the past it has no qualms incorporating stuff made by third parties, even rivals, when they offer great features that Apple’s own users might want. The very first iPhone it launched, you may remember, came pre-installed with YouTube, for example.
But AI is surely not just a feature, but The Future—one which Apple won’t be satisfied with outsourcing for very long. The company is, of course, working on its own AI projects, as described in a recently published research paper. You could imagine it would seek to shut out Google, and anyone else, as soon as it can do a good enough job itself.
As of this week, investors in both companies seem delighted, sensing something of an answer to the Microsoft and OpenAI surge. [With Apple’s market cap having taken a hit], for months now, Cook has been steadily reassuring investors, hinting at significant breakthroughs to be (maybe, possibly) announced at its coming developers conference in June. Expectations are already sky high.